Understanding the Marshall Attack

Understanding the Marshall Attack David Vigorito

[8AI�IBIITI

First published in the UK by Gambit Publications Ltd 20 1 0 Copyright © David Vigorito 20 1 0 The right of David Vigorito to be identified as the author of this work has been asserted i n accordance with the Copyright, Designs and Patents Act 1 988. All rights reserved. No part of this publication may be reproduced, stored in or introduced into a retrieval system, or transmitted, in any form or by any means (electronic, mechanical, photo­ copying, recording or otherwise), without prior permission of the publisher. In particular, no part of this publication may be scanned, transmitted via the Internet or uploaded to a website without the publisher's permission. Any person who does any unauthorized act in relation to this publica­ tion may be liable to criminal prosecution and civil claims for damage. ISBN- 1 3 : 978- 1 -906454- 1 7-3 ISBN- lO: 1 -906454- 1 7-5 DISTRIBUTION:

Worldwide (except USA): Central Books Ltd, 99 Wallis Rd, London E9 5LN, England. Tel +44 (0)20 8986 4854 Fax +44 (0)20 8533 582 1 . E-mail : [email protected] Gambit Publications Ltd, 99 Wallis Rd, London E9 5LN, England. E-mail: [email protected] Website (regularly updated) : www.gambitbooks.com Edited by Graham Burgess Typeset by Petra Nunn Cover image by Wolff Morrow Printed in Great Britain by The Cromwell Press Group, Trowbridge, Wilts.

10 9 8 7 6 5 4 3 2 1

Gambit Publications Ltd

Managing Director: Murray Chandler GM Chess Director: Dr John Nunn GM Editorial Director: Graham Burgess FM German Editor: Petra Nunn WFM Webmaster: Dr Helen Milligan WFM

Contents

Symbols

4

Dedication

4

Acknowledgements

4

Bibliography

4

Introduction

5

Recommendations

10

Typical Ideas in the Marshall Attack

12

I e4 e5 2ltJf3ltJc63 �b5 a64 .i.a4ltJf65 0-0 .i.e7 6:tel b5 7 .i.b3 0-0

Part 1: Main Lines with 8 c3 dS 9 exdS tiJxdS 10 tiJxeS tiJxeS 11 ':xeS c6 12 d4 .id6 13 ':e1 'i¥h4 14 g3 'i¥h3 I

Spassky Variation

20

2

Old Main Line: 18 ...f5/18 ...bxa4

31

3

15 SLe3: Other Lines

51

4

Modem Variation: 15 ':e4

74

Part 2: Other Lines after 8 c3 dS 9 exdS tiJxdS 10 tiJxeS tLlxeS 11 llxeS c6 5

Refined Rook-Lift: 12 d3

6

Elite Equalizer: 12 d3 .td613 .:tel SLf5

108

7

Early Deviations

120

92

Part 3: Anti-Marshall 8

Anti-Marshall: 8 a4

137

9

Anti-Marshall: 8 h3

155

Other Anti-Marshall Lines

170

10

Index of Variations

190

Symbols x + ++ # !! !? ?! ? ?? (D)

capture check double check checkmate brilliant move good move interesting move dubious move bad move blunder see next diagram

+± ;!; =

+ =+= -+ 1 -0 112-112 0- 1 Ch

White is winning White is much better White is slightly better the position is equal Black is slightly better Black is much better Black is winning the game ends in a win for White the game ends in a draw the game ends in a win for Black championship

Ded ication For Heather

Acknowledgements Thanks to Jim Rizzitano for his support, Joe Fang for access to his library, and to Graham Burgess for his patience !

Bibl iogra phy Books

Anand: C89, S ahovski Informator, 2002 Gershon and Nor: San Luis 2005, Quality Chess, 2007 Johnsen and Johannessen: The Ruy Lopez: A Guide for Black, Gambit, 2007 Lalic: The Marshall Attack, Everyman, 2003 Nunn and Harding: The Marshall Attack, Batsford, 1 989 Nunn, Burgess, Emms and Gallagher: Nunn's Chess Openings, GambitJEveryman, 1 999 Pavlovic : Fighting the Ruy Lopez, Everyman, 2009

Periodicals Informator up to Volume 1 02 New in Chess Magazine up to 2009/2 New in Chess Yearbook up to 9 1

Electronic Resources ChessBase Mega Database 2009 ChessBase.com ChessCafe.com ChessLecture.com ChessPublishing.com Harding: Total Marshall CD The Week in Chess up to issue no. 770 Chess Today

Introd uction

It is a pleasure to introduce my second book for Gambit. The Marshall Attack in the Ruy Lopez is one of the most important openings in all of chess theory. It has been played by many of the world's top players over the last couple of de­ cades, including Anand, Kramnik, Ivanchuk, Leko, Adams, Kamsky, Bacrot, Svidler, Short, Grishchuk, Shirov, Ponomariov, Yakovenko, Beliavsky, Harikrishna, Kasimdzhanov, Ako­ pian, I . S okolov, Khalifman and especially Aronian . This book is designed to give Black a workable knowledge of the whole complex of the Marshall Attack arising after the following move-order: 1 e4 eS 2 �f3 �c6 3 .i.bS a6 4 .i.a4 �f6 S 0-0 .i.e7 6 l::te l bS 7 .i.b3 O-O! (D)

least hints ! ) that he is not willing to defend a middlegame with less space. S c3 White intends to build a broad pawn-centre with a quick d4. After 8 . . . d6 9 h3 (preventing the pin . . . .i.g4, which would put pressure on White's centre) White is ready to play 10 d4. However, by using the move-order 7 . . . 0-0, Black can throw a spanner in the works and at­ tempt to seize the initiative by opening the posi­ tion immediately. Many players prefer to avoid the Marshall Attack by playing various Anti­ Marshall systems. We shall look at these ideas in the second part of the Introduction. S dS! (D) •••

w

By playing this move (instead of 7 . . . d6) Black signals his intention to play actively. The main lines of the Ruy Lopez which arise after 7 . . . d6 8 c3 0-0 9 h3 allow White to follow up with 1 0 d4, which gains space in the centre. While Black has many respectable systems to choose from here (especially the Chigorin, Breyer and Zaitsev systems), they all allow White to seize space in the centre. Because of this advantage, White is able to play with a certain initiative, and many players would pre­ fer not to defend a solid, but cramped, posi­ tion. By playing 7 . . . 0-0 Black shows (or at

It is this move that constitutes the Marshall Attack. Black can also use the Marshall move­ order as a bit of a bluff and head back into Closed lines with 8 . . d6, but we have no interest in that approach. By offering a pawn, Black avoids the ' Spanish Torture' often associated with White's space advantage in the Closed lines of the Ruy Lopez. For the pawn, Black will get free development while White's move c3 hinders his bl -knight from quickly entering the game. Because White's kingside is sud­ denly lacking in defenders, Black hopes that his development lead translates into an attack against the white king. However, that is not the .

UNDERSTANDING THE MARSHALL A TTACK

6

only way that Black ends up with compensa­ tion for the pawn. There are many instances where Black seeks his fortune in the centre and on the queenside. gexdS White can still bail out with 9 d4. This is not terribly dangerous and it is covered in Chapter 1 0. 9 liJxdS The old Steiner Variation, 9 . . . e4, has been considered dubious for years and will not be covered here. 10 liJxeS liJxeS 1 1 ::'xeS (D) •••

by such players as Short and Kamsky, but de­ spite their relative successes it has not caught on. After 1 2 'ii'f3 (White can also play 1 2 d4, when Black has some compensation after ei­ ther 12 . . . iLf6 or 12 . . . 'ii'd 7) 12 . . . iLd6 13 iLxdS (worse is 1 3 ::' xdS ? ::' e8 ! , which favours Black after either 14 �f1 'ii'e7 ! I S 'ilkd l 'ii'h4 16 .l:.hS 'ii'e4 or 14 'ii'xf7+ �xf7 { 1 4 . . . �h8 ! looks inter­ esting as well } I S ::' xd6+ �f8 1 6 ::' xd8 l:1axd8 1 7 �f1 iLe4 ! ) 1 3 . . . c6 14 l:1e2 cxdS I S d4 'ifc7 1 6 g3 l:.ae8 and after either 1 7 iLe3 or 1 7 liJd2 Black clearly has some positional compensa­ tion. Whether it is enough is difficult to deter­ mine . . . Because this line does not appear to be trusted by the world's elite players and I do not have anything special to add to old theory, this book will only focus on 1 1 . . .c6 (D), to which we now return.

w

1 l c6! Black simply protects the dS-knight and in­ tends to play . . . iLd6 and . . . 'ii' h 4. Marshall him­ self originally played 1 1 . . .liJf6 ? ! (intending . . . liJg4) in the Marshall Attack's inaugural game against Capablanca. Marshall had been saving his idea for this particular occasion, but Capa­ blanca reacted with the remarkably cool 1 2 d4 (the actual move-order was 1 2 l:.e l iLd6 1 3 d4) 1 2 . . . iLd6 1 3 .l:.e l ( 1 3 ':'e2 i s also consid­ ered to be strong) 1 3 . . . liJg4 14 h3 'ifh4 IS 'ii'f3 liJxf2 (this was Marshall's original concept) 1 6 l:te2 ( 1 6 iLd2 has also scored well for White) 1 6 . . . iLg4 ( 1 6 . . . liJg4 can be met by 1 7 g3 ! 'ii'x h3 1 8 'ii'x a8 'ii'x g3+ 1 9 'ii'g 2) 17 hxg4 iLh2+ 1 8 �f1 iLg3 1 9 l:txf2 'ii'h l + 20 �e2 iLxf2 (after 20 . . . 'ii'xc l 2 1 'ii'x g3 'ii'x b2+ White has a pleas­ ant choice between 22 liJd2 'ii'x al 23 l:1xf7 and 22 �d3 'ii'x al 23 �c2) 2 1 iLd2 ! and White started to take over in Capablanca-Marshall, New York 1 9 1 8 . 1 1 . . .iLb7 is, on the other hand, rather re­ spectable. It has seen a few high-level outings ...

12 d4 This is the most natural move. White is not afraid and fights for space in the centre. This is the main line and is the most popular way of fighting the Marshall Attack. The quieter 1 2 d3 has been very popular too. This move gives White control of the e4-square which gives rise to some interesting tactical possibilities. After 1 2 . . . iLd6 1 3 .l:.e l , Black can play as he does in the main lines with 1 3 . . . 'ii'h4 (Chapter S) or take aim at the d3-pawn itself with 1 3 . . . iLfS (Chapter 6). Other moves such as 1 2 l:te l (and the related 1 2 g3) and the commit­ tal 1 2 iLxdS are covered in Chapter 7. 12 iL d6 (D) 13 l:.e1 This simple retreat is the most popular and natural continuation. The odd-looking 1 3 ':e2 •••

INTRODUCTION

7

15 .i.g4 16 'ifd3 White cannot play 1 6 f3 ? due to 1 6 . . . iLxg3 ! with a strong attack. 16 l:.ae8 Black activates his last piece. Note that 1 6 . . . iLf3 ? can be met by 1 7 'iff! ! . This idea is one of White's principal defensive resources in the main lines of the Marshall Attack. The somewhat rash pawn advance 1 6 . . . fS is cov­ ered in Chapter 3 . 1 7 lLld2 (D) •••

•••

covers the second rank and is considered in Chapter 7 . 13. . .'ifh4 14 g3 This weakens the light squares around the white king, but it gains time by attacking the black queen. It is necessary though, because af­ ter 1 4 h3? Black can simply play 1 4 . . . iLxh3 ! I S gxh3 'ifxh3 1 6 .l::teS ( 1 6 f4 lLlxf4 ! 1 7 iLxf4 iLxf4 1 8 .l::te 2 .l::ta e8 gives Black an overwhelming at­ tack) 1 6 . . . .i.xeS 1 7 dxeS and here Black has a pleasant choice between 1 7 . . . .l::tfe8, 1 7 . . . .l::tae8, 1 7 . . . l:.ad8 and 1 7 . . . 'iffS . 14 'ifh3 (D) •••

B

White has caught up in development and is ready to open the queenside with a4, so Black must play very actively. 1 7 .l::t e6 This is the main line. Black may double rooks on the e-file and he could possibly use the rook along his third rank. Black can also initiate a pawn advance with 17 . . . fS or play the flexible 17 . . . 'ifhS , both of which are covered in Chapter 3 . 18 a4 (D) •••

Black intends to play . . . .i.g4 and . . . l:.ae8 to complete his development and fuel his initiative. White has to decide how to develop his pieces. 15 iLe3 This develops the bishop to allow lLld2 while protecting f2 and blocking the e-file. The main alternative is the rook-lift I S .l::te4, which is cov­ ered in Chapter 4. Alternatives such as the com­ mittal I S .i.xdS are considered in Chapter 7.

8

UNDERSTANDING THE MARSHALL A TTACK

This advance is White's main resource. Not only does White want to open the a-file for his rook, but he introduces the concrete threat of 1 9 axb5 axb5 20 �xd5 cxd5 2 1 'ifxb5 winning a second pawn. White has a few alternatives, of which the most topical is Kramnik's 1 8 'iff! 'ifh5 1 9 f3 ! ?, which is discussed in Chapter 3 . After 1 8 a4 Black has two main ways t o play. He can ignore White's threat and play 1 8 . . . f5 , intending the further advance . . .f4 to destroy White's kingside. The move 1 8 . . . bxa4 looks like a different approach altogether, but in fact these two moves often reach the same position, although for very specific tactical reasons, it seems that 1 8 . . . bxa4 is the preferable route. The details of these tactical lines are covered in Chapter 2. Black's most popular move at elite level is 1 8 . . . 'i!f h5. This retreat protects the d5knight and so allows Black to meet �xd5 with . . . 'ifxd5 , thus avoiding the destruction of his pawn-chain. Although Black may get a direct attack in these lines, in general Black is playing for positional compensation by means of piece­ play. This move was introduced by Spassky and is covered in Chapter 1 .

8 �b7 Black develops a piece and sets a little move­ order trap of his own. Now if White plays 9 c3, then 9 ... d5 ! is an even better Marshall Attack for Black, because . . . �b7 will certainly be more useful than h3 if the position opens up. So White changes gears. 9 d3 (D) ••.

B

Anti- Marshall Lines 1 e4 eS 2 �f3 �c6 3 �bS a6 4 �a4 �f6 5 0-0 �e 7 6 l:.e1 bS 7 �b3 0-0 8 h3 (D) This unassuming pawn move is currently the most popular Anti-Marshall Line. Formerly, 8 a4 (Chapter 8) was White's main method of avoiding the Marshall. Black can then develop with 8 . . . �b7 or play the modem preference 8 . . . h4. White can also seek to open the position immediately with 8 d4 (Chapter 1 0), but that gives Black a pleasant choice between trans­ posing to a line of the Closed Ruy Lopez with 8 . . . d6 9 c3 �g4 ! or accepting the invitation to an open fight with 8 . . . �xd4. White does have alternatives such as 8 d3, but by playing this way White forgoes the option of gaining space with a quick d4, and Black should not experi­ ence any major difficulties. So what is the point of 8 h3 ? Well, first of all it gives Black the chance to fall back into the main lines of the Closed Ruy Lopez with 8 . . . d6 9 c3 ! By playing 8 h3, White makes a rather useful move while not quite giving up on the idea of playing d4.

Now this move is quite logical, because the e4-pawn is well protected and the b7-bishop may find itself passively placed. In fact, Black will often retreat the bishop back to c8 at some point in order to challenge White's light-squared bishop with . . . �e6. 9 d6 Now that White's d-pawn has taken a step forward, Black is willing to play a slower posi­ tion too because he will not be suffering for space. B y securing the e5-pawn, Black also threatens to grab the bishop-pair with . . . �a5 . Recently, playing in Marshall Attack style with 9 . . . d5 ! ? has been resurrected as well, and •••

INTRODUCTION

this modern gambit is also examined in Chap­ ter 9 . 10 a3 (D) White would like to keep his bishop on the nice a2-g8 diagonal, especially with Black 's bishop far away on b7 .

Because White's play has been rather slow, Black has a lot of choice in how to set up his pieces. There is the 'Breyer' retreat 1O . . . lbb8, intending to reposition the knight on d7 while allowing . . . c5 or . . . c6. Black can also play in 'Zaitsev' style with 1O . . . 'ifd7 . This connects the rooks and also allows the manoeuvre . . . lbd8-e6.

9

The classical plan, however, is to play in 'Chigorin' style with a quick . . . lba5 and . . . c5. Although playing like this decentralizes Black ' s knight, h e does gain a tempo b y attacking the b3-bishop, and the knight' s influence on the b3- and c4-squares could have some impor­ tance. After gaining space with . . . c5, the knight may also return to the fray by just coming back to c6 without worrying about being pushed again by White ' s d-pawn, as it is more pas­ sively placed at d3 than it would be on d4. In general, this is the approach that is advocated in this book. Not only is it the most popular ap­ proach, but it is the most in accordance with classical play in the Ruy Lopez. One must always remember that a Marshall player is first of all a Ruy Lopez player. There will always be lines that White can play to avoid the Marshall Attack even before move 8, such as the Exchange Variation, 5 d4, lines with a very early lbc3 or d3, etc. None of these lines are covered here, but the reader may want to re­ fer to The Ruy Lopez: A Guide for Black by Johnsen and Johannessen. Of course, White can avoid the Marshall even earlier with anything from the King's Gambit to the Four Knights Game. For these lines I refer the reader to Beat­ ing the Open Games by Emms.

Recommendations

The Marshall Attack is a rich opening with a long history. I think it can be employed by players of all levels, from club player up to world-championship level. Although it has a reputation as being drawish, this is only likely to be true at a very high level indeed. On the contrary, I believe that for most players the Marshall Attack would be an ideal weapon. It is dangerous and also has the benefit of being a logical, classical opening that can help develop one's chess overall. Besides, below mas­ ter level it is much easier to attack than it is to defend, and to get a chance to do so with Black right from the opening should be welcomed ! For the typical club player, I would start by going through Chapter 2, the Old Main Line. Al­ though this variation has fallen out of favour, it contains an abundance of typical Marshall Attack themes. In the Old Main Line, Black bums all his bridges and must play for the initiative. The value of every move is high and one can learn a lot about attack, defence and counterattack by going through the lines in this chapter. Similar play can be found in the Pawn Push variations of Chapter 3. Although one may not want to play these lines forever because of their rather dubious theoretical value, I think the creative black player could get some mileage out of these systems if he picks and chooses his lines carefully. Next I would study the lines of Chapters 4 and 5. The Modem Variation can lead to some very odd scenarios. Here it is common for White to offer material, usually the exchange, in order to wrest the initiative from Black. If Black is not careful, his position can soon become a positional wreck because of the pawn advance on the kingside with . . . g5 . Black must pay attention to the ini­ tiative and not just 'win' the exchange as soon as it is offered. Although the lines of the Modem Variation are not quite as cut-throat as the ones that are found in Chapter 2, the positions contained in this chapter could go a long way towards showing the balance between initiative and material. If White plays 12 d3, I think it is best to study the lines of Chapter 5 with 1 3 . . . 'ii'h4. These lines are not so popular nowadays but they have hardly been refuted and I think they well illustrate typi­ cal Marshall ideas. The themes are a little less obvious than the barbaric approaches of Chapter 2, but the positions are well suited to show how Black should fight for the initiative while a pawn down. Once you have investigated the main and most instructive systems it is time to move on to the more subtle variations and the sidelines. While most lines from Chapter 7 are not terribly danger­ ous theoretically, one may encounter them often in practice, whether as a surprise weapon or as a consequence of White simply not knowing the main lines. Black should also pay attention to some of the deviations on move 1 8 contained in Chapter 3, especially Kramnik's line with 1 8 'ii'f l . Finally, I would study the lines of Chapter 1 and Chapter 6. The Spas sky Variation is practically the universal choice of the world's best players and it has an excellent theoretical reputation. In these lines Black must not rely only on attacking skills, but also on 'feel' of the position. Black plays positionally here, and, being a pawn down, he must develop a good sense of what constitutes compensation. The same holds true in Chapter 6. In the main lines here Black can look to draw a pawn-down ending or play a middlegame where he is sometimes two pawns down. Both of these approaches are quite viable, and I would recommend that the Marshall student study both ap­ proaches because they will help develop an understanding of the Marshall Attack as well chess in general. You never know, if you are attacking and something goes wrong, you could find yourself defending a pawn-down ending, depending on the bishop-pair to provide enough activity to hold the game. Conversely, if positional play is not working, sometimes you may have to go all-out for the attack, a pawn or two deficit notwithstanding.

RECOMMENDA TIONS

11

Of course, there is the final part of the book as well : the Anti-Marshalls. A few of these lines are quite sharp, notably the 9 . . . d5 gambit of Chapter 9 and the central thrust 8 d4 of Chapter 1 0. Over­ all, however, the play in these chapters is of a more strategic nature. That's just chess; you cannot always force the play, especially with Black. The good news, however, is that these lines should hardly scare Black because he is generally fighting on even ground. Unlike most Closed Ruy Lopez lines, in the Anti-Marshalls Black usually has as much space as White does. Big novelties are rare in the Anti-Marshalls and a player who learns the ideas of the systems recommended in this book can look forward to the middlegame with confidence. The key to playing the Marshall Attack is absorbing typical ideas. These include typical tactical motifs for both sides, direct attacking ideas, and even defensive ideas. While many positions can become quite complicated and tactical, often the play is positional in many respects - Black is a pawn down but has some initiative. In order to prevent White from gradually neutralizing Black's threats and consolidating his position, Black must find a way to disturb White. Often this is with relatively quiet moves that simply improve the positioning of his pieces. There are pawn moves too, whether they are a typical . . . f5 pawn-rush, a probing . . . h5 or a central hit with . . . c5 . Throughout the book, there are plenty of diagrams to show Black's various methods of troubling White. The Mar­ shall student should worry less about memorizing variations and focus on learning ideas. By using this method there is a much better chance of success, because if one's memory fails, the problems can still be solved if one has a good understanding of the positions that arise.

Typical Ideas in the M arshall Attack

Here we look at twenty common themes that arise in the Marshall Attack. Anti-Marshall ideas are generally less striking so I have cho­ sen to let the respective Anti-Marshall chapters illustrate the ideas on their own. The chosen ex­ amples from the Marshall Attack vary from tac­ tical ideas to positional themes and include mating ideas as well as endgame patterns. Of course, there are many other ideas that will be shown throughout the book, but this is a little sampler to whet the reader's appetite.

Rook-Lift Renet - Nunn

European Team Ch, Haifa 1989 22 JWeS! 0-1 White cannot save both the a4-rook and the e3-bishop. ••

Pin along the Third Rank

Hellers - I. Sokolov

Haninge 1989 White always has to be wary of all destruc­ tive ideas near his king. 20 lbx e3! The game continued 20. . . �xd l 21 :axd l f5 22 lb n f4 23 �c 1 l:lef6, when Black had a strong attack, but the text-move is even better. 21 fxe3 .txg3! 22 hxg3 :h6 and White will soon be mated. •••

Cao Sang - Haza i

Hungarian Team Ch 199718

The .. :iVe8 Fork The next position has arisen several times. Here White has just wrongly captured a pawn on a4.

White does not look to be in any danger here, but Black has a typical idea using the fact that the white queen is undefended on d3 . There is a

TYPICAL IDEAS IN THE MARSHALL A TTACK

reason why White often hurries to evict the black queen with a quick 'iffl . 19 �f4! Sometimes a move like this works because Black has doubled rooks along the e-file. Black has ideas like . . . lLlxe3 followed by . . . �xg3 . 20 �xf4lLlxf4 21 'iVn lLle2+ 22 l:txe2 22 �h l 'ifh6 ! is very strong, because of . . . �f3 mating ideas. 22 l:txe2 23 lLlc4 Now Nunn's suggestion 23 . . . l:te6 would have given Black the better chances.

13

Kingside War

•••

•••

White Sacrifices a Piece to Stem Black's Attack R. Ju nge - Wegner

Bundesliga 1988/9

De Ol iveira - Maffei carr.

Here we have a typical situation. Black has a very active position but his queenside is crum­ bling - the a6-pawn is hanging and c6 and d5 will also come under pressure. Black has no choice other than to aim for total mayhem. 22...gS! 23 fxgS 23 ':xa6 is another option. 23...f4! Black invests two more pawns to rip open the position. There are even situations where a fourth pawn will be offered with . . . h6. 24 gxf4 �h3 (D)

1999

This position arises from the ' Internet Refu­ tation' covered in Chapter 2. White has cap­ tured on b5 and Black has answered with the decisive-looking advance . . . f4. 20 �xf4! White just gives up a bishop to eliminate the dangerous black f-pawn. 20 �xf4 Later we shall see that other moves are not much better. 21 ':xe6 �xe6 22 bxa6! This is better than taking on c6. Now White has four pawns for the piece and the passed a6pawn ensures that White will keep the initia­ tive. 22...�xd2 23 'ifxd2 and White went on to win. •••

Now White must be careful. 2S �hl This may well be the best move, but White's next is a blunder. 2S..:it'g4 26lLlf3? ':xe3!

UNDERSTANDING THE MARSHALL A TTACK

14

Black wins because the white queen is tied to the defence of the g2-square.

Exchange Sacrifice on e4

B

Chandler - Nunn

Hastings 198718

Anand - Topa lov

Las Palmas 1993 White has exchanged pieces on d5 and played �g2. Black does not want to defend with the passive 2 1 . . .'iff7, so he throws more wood on the fire. 21 J:t e4 This is an attractive move. If White captures the rook, Black's light-square control will be accentuated. White will be an exchange and a pawn up, however. 22 h4! ? Anand prevents any . . . �h3 ideas before tak­ ing the exchange, but this weakens g3. 22 h6 23 lLlxe4 fxe4 24 :0 11f6! Black intends . . . l:.g6 with counterplay on the kingside.

26 dxc5 .i.xc5 further activates Black' s pieces and leaves White very passively placed. Now the game continued 26 . . . cxd4 27 'ifxd4 l:.e4 28 'ifd3 �c5 and here I think the quiet 29 ..t>g2 ! would have enabled White to consolidate the extra pawn. Instead I suggest 26 lLlr6!?, when Black has good counterplay. •••

Two- Bishop Endgame

••

B

•••

The ...c5 Break In the following diagram, White intends to play .i.d2 and lLle3 to exchange another set of pieces, after which it will be difficult for Black to get any significant counterplay. 25 c5! Black is more active, so he opens the posi­ tion for his pieces. 26 b3 White prepares the c4 advance. After 26 .i.d2, Black can isolate White's d-pawn, which should give Black enough counterplay. Instead •••

I. Gu revich - Benjamin

Reshevsky Memorial, New York 1992 Here we have a typical Marshall endgame. White has difficulty creating a passed pawn, and Black's space on the kingside keeps White at bay as well. 28...� e2! 29 lLlh2 If White plays 29 gxf4 �xf4 his knight will only be able to move somewhere that will allow

TYPICAL IDEAS IN THE MARSHAll A rrA CK

Black t o head for a drawn opposite-coloured bishop ending (30 �d2 �xf1 ! =). 29...�f7 30 �g2 �f6 31 f3 h5! Taking away the g4-square. 32 g4 g6 33 lbn hxg4 34 hxg4 g5 35 lbd2 �dl Preventing lbb3 . 36 �2 �e6 37 �n �d7 38 �f2 iLc7 39 �e1 �c2 40 �e2 �a5 41 �el 1h.- lh.

15

B

Two- Bishop Endgame with Rooks

T imman - Nunn

Brussels 1988

del Rio - Hebden

Port Erin 2003 This ending has a similar structure but Black has to be careful because there are more pieces on the board. 24...b4 This looks active, but perhaps Black should wait with this and play 24 . . . �d3 . As we shall see, this advance exposes the d5-pawn to a lat­ eral attack by a white rook. 25 :eel bxc3 26 bxc3 l:r.c8 27 �d2 iLe4 28 l:.e3 h5 29 f3 �f5? After this the d5-pawn just drops. 29 . . . �c2 was necessary, but Black will still suffer. 30 l:r.35 and White went on to win.

White Gives Back Material to Reach a Favourable Endgame In the following position, Black has played the destructive . . . g5 and now makes a further sacri­ fice to break down the dark squares. 23...l:r.xe3 24 l:r.xe3 f4 25 gxf4 iLxf4 26 :g3

White takes advantage of the pin along the f-file. He will soon give back all of the material in order to simplify the position. 26 .. :ii'xg5 27 �hl �d6 28 'iWg2 �xg3 29 'ii'xg3 h5 30 :n :xfi+ 31 lbxfi 'iWc1 32 �gl 'ii'xb2 33 h3 White has reached a favourable 'iW+lb vs 'ii'+ � ending and with accurate play he can cre­ ate considerable problems for Black.

Black's Queenside is Weakened in the Endgame

Ya kovenko - Bacrot

Dortmund 2009 Despite the simplifications and opposite­ coloured bishops, White maintains an edge be­ cause Black's queenside pawns are vulnerable

UNDERSTANDING THE MARSHALL A TTA CK

16

to White ' s minor pieces. In this case, the e6pawn is weak as well. 34 JU6 35 ll)e5 35 ll)c5 is also tempting. 35 ii.c1 36 ll)g4 :f3+ 37 �g2 :f4 3S ii.b3 :xg4+ 39 �xfi White has won a pawn. Although it will not be easy to convert, Black is in for a long and miserable defence.

The 'Sveshnikov' Rook-Lift

••

•••

Fuelling the Initiative in the Endgame

Topalov - Adams

Sarajevo 2000 White has two pawns for the exchange and a sound pawn-structure. Adams finds a way to wrest the initiative. 21 f4! Black offers a pawn to wreck White's pawn­ structure and open lines for his own rooks. 22 ii.xf4 ii.xf4 23 gxf4 l:ta7! This is it ! This manoeuvre is often seen in the Sveshnikov Sicilian, where Black has an open second rank, as he does here. 24 �hl l:te7 25 ':'gl + 'il;JhS 26 ll)n 'iVe4 Black maintains the initiative into the end­ game and went on to win. •••

Ponomariov - Anand

Linares 2002 In this famous game, White is trying to use tactics to defend his kingside. If Black takes on e6, White will win both of Black's bishops for the rook. Although some correspondence games have shown the capture on e6 to have some via­ bility, from a theoretical standpoint Anand's brilliant sacrifice is sufficient. IS f4! Black gives up a whole piece to bring all of his remaining pieces into play ! 19 lbd6 ii.g4 20 'iVn 'iVxfi+! Remarkably, Black can even afford to ex­ change queens. 21 'il;Jxn l:taeS 22 ii.d2 ii.h3+ 23 �gl fxg3 24 hxg3 l:te2 25 ii.e3 ':'xe3 26 fxe3 :n + 27 'il;Jh2 g4! 28 ':'xd5 28 ll)d2 l:txal looks risky for White because of the activity of Black's pieces.

White's Exchange Sacrifice

•••

Ib.-Ib.

Black has perpetual check.

Ponomariov - Adams

Linares 2002

TYPICAL IDEAS IN THE MARSHALL A TTA CK

White's centralized rook looks a bit awk­ ward. White solves the problem of saving the rook by not saving it at all. This is justified be­ cause of the weaknesses of Black's kingside. 17 liJd2! iLfS After 1 7 ... liJxe4 1 8 liJxe4 the d6-bishop is attacked and the g5-pawn will fall, giving White two pawns for the exchange. 18 f3 Now if Black ever captures on e4, White will get a central pawn-roller. Sometimes Black can fight against these pawns effectively, but that is not the case here. 18 cS 19 'ii'f2 c4 20 iLc2 h6 21 b3 cxb3 22 axb3 :rc8 23 iLb2 White is ready to advance his c-pawn, whereupon he will dominate the centre. 23 iLb4? Black panics. 24 l:.eS iLxc2 25 cxb4 iLg6 26 l:.cS l:.e8 27 liJn l:.ad8 28 dS! White has a winning attack.

17

21 h6 22 g2 l:.d7 23 h4! gxh4 24 eS iLe7 25 gxh4 g7 If 25... iLxh4 then 26 liJf3 wins back the pawn. 26 axbS The immediate 26 liJe4 was also possible. 26... axbS 27 liJe4 The knight is heading to g3 and f5. White has a powerful initiative. •••

A Misplaced Rook

•••

•••

White's Exchange Sacrifice in the Endgame Ragger - Beliavsky

Graz 2008 Here Black has no real kingside attack and development is level. However, White's 'ac­ tive' rook is misplaced, and this allows Black to develop some initiative. 17 iLe7! 18 iLc2 'ii'g6 19 d4 iLfS Black is not opposed to this exchange be­ cause it helps his development. 20 iLxfS ..xes 21 .i.eS liJdS 22 g4 (D) •••

Yakovenko - Zhang Zhong

China-Russia Summit Match, Ergun 2006 Even with the queens exchanged, White's central roller combined with the weaknesses in Black's structure can give White good chances. 21 a4! Before starting operations on the kingside, White creates the possibility of opening the a­ file.

18

UNDERSTANDING THE MARSHAll A rrA CK

Frustrated, White lashes out, but this gives Black a clear target. 22 'ii'g6 23 l:lh3 f6 24 i.g3 fS! Black has a strong initiative. •••

White Gives Back the Pa wn to Exchange the Right Pieces

Shirov - Aron ian

MorelialLinares 2008

Fressinet - Hamdouchi

Belfort 2003 White's back rank is threatened and the d3pawn is under attack. White returns the pawn to c�mplete hi� development and reach a position WIth a supenor pawn-structure and minor piece. 17 li:)d2! i.xf4 18 'ii'xf4 i.xd3 19 li:)b3 'ii'c6 20 'ii'd2 .i.g6 (D)

Here we have a typical ending from the 12 d3 i.d6 13 l:le 1 i.f5 line. In this case,Black has been careless and White grabs space on the kingside,looking to create a 'second weakness' in the black position (this first is White's extra pawn on the queenside). 30 g4! .i.c6 31 l:ld3 l:le8 32 l:ld4 l:lb8 33 h4 i.e7 34 hS Further putting the squeeze on. 34 l:ld8 35 l:lc4 i.bS 36 l:lc7 l:ld7 37 lhd7 .i.xd7 (D) •••

38 h6! This breaks up Black's kingside and gives White a passed f-pawn. 38 gxh6 39 i.xf6 i.f8 40 .i.d4 'Yith an extra pawn and the better position, WhIte has a decisive advantage. •••

TYPICAL IDEAS IN THE MARSHALL A TTA CK

19

Black's Minority Attack on the Queenside Sometimes White exchanges on d5 early on to reduce Black's attacking potential. When it is difficult for Black to create play on the king­ side, he can look to the other side of the board.

Bacrot - Aronian

European Clubs Cup, Kallithea 2008 In this example, Black employs a typical method of active defence - he advances his kingside pawns. 29...h5! Not only does this prevent White from gain­ ing space on the kingside, it also will make him think twice before wandering over to the queen­ side, because White's own kingside pawns could become vulnerable to Black's strong bishops. 30 il.d4 il.c6 31 b3 f5 32 '>tf2 g5 33 il.e3 g4 34 f4 '>te6 35 il.d4 (D)

de Firmian - Sa rgisian

Politiken Cup, Copenhagen 2007 21.. JUcS Black prepares to play on the queenside. 22 a3 as 23 ltJd2 b4 24 axb4 axb4 25 l:haS l:haS 26 ltJb3 bxc3 27 bxc3 :a3 (D)

w

B

liz_liz

In view of the activity of Black's pieces, it is hard to see how White will ever be able to make any progress.

White's extra pawn is backward and Black has the bishop-pair. White played... 2S ltJc5 ...and offered a draw.

Part 1 : M a in Lines with 8 c3 d5 9 exd5 ltJxd 5 1 0 ltJxe5 ltJxe5 1 1 :xe5 c6 1 2 d4 �d 6 1 3 :e l iYh4 1 4 g3 iYh3 1 Spassky Variation 1 e4 eS 2 lLlf3 lLlc6 3 .i.bS a6 4 .i.a4 lLlf6 5 0-0 i.e7 6 %lei bS 7 i.b3 0-0 8 c3 dS 9 exdS lLlxdS 10 lLlxeS lLlxeS l l ':'xeS c6 12 d4 i.d6 13 .:tel 'ifh4 14 g3 'ifh3 1S i.e3 i.g4 16 'ii'dJ l:tae8 17 lLld2 l:.e6 18 a4 'ifhS (D)

19 axbS White opens the a-file for his rook. There is no reason to delay this move, and doing so can only harm White, because there could be some positions where White may not have time to transpose back to known lines. For example, if White decides later on that the time is right to play axb5 intending to meet ...axb5 with l:.a6, it is possible that Black will have a stronger move than ... axb5, especially as he will be playing against White's king. 19 axbS (D) •••

Black meets White's threat to destroy the queenside with 19 axb5 axb5 20 i. xd5 cxd5 21 'ifxb5 by protecting the d5-knight. This is less brutal than the pawn-pushing variations of the next chapter. Instead of forcing matters by launching the f-pawn and creating a kind of self-pin on the a2-g8 diagonal, Black relies on more positional means. Usually Black will fol­ low with ... %lfe8, but pushing the f-pawn is still a possibility in several lines, especially if White retreats his light-squared bishop with i. dl be­ cause there will be no pressure on the a2-g8 di­ agonal. This has become the main line of the Marshall with 15 i.e3 and is favoured by all of the elite players who employ the Marshall At­ tack with the black pieces.

Black is ready to increase the pressure by doubling rooks. White has four main options, of which the last two are the most important: 21 A: 20 i.xdS B: 20 lLle4 21 C: 20 lLln 22 D: 20 'ifn 26 Other moves are not dangerous at all:

SPASSKY VARIA TION

a) 20.i.dl ? allows a typical idea: 20 ...li)xe3 (20....i.xdl 2 1 l:taxdl f5 22 li)fl f4 23 .i.cl l:tef6 was also overwhelming in Hellers-I.Sok­ olov, Haninge 1989, but this is even stronger) 2 1 fxe3 .i.xg3! 22 hxg3 l:h6! is just winning for Black. b) 20 c4 and now: bl ) 20...li)xe3?! 2 1 l:xe3 (again 2 1 fxe3? loses to 2 1.. ..i.xg3 22 hxg3 l:h6) 2 1.. ..i.e2 22 'ii'c3 (22 'ii'c2?! l:xe3! 23 fxe3 'ii' h6 24 'ii'e4 .i.b4 gives Black good play) 22 ... b4 23 'ii'cl l:h6 24 h4 g5 25 'ii'el .i.g4 26 fJ .i.c8 27 c5 .i.b8 28 l:e8 gxh4 29 'ii'e7 l:xe8 30 'ii'xe8+ r3itg7 3 1 'ii'xc8 hxg3 32 'ii'g4+ was winning for White in Griinfeld-Pein, Tel Aviv 1990. b2) 20... bxc4 is the simplest. 2 1li)xc4.i.b4 22 l:ecl .i.e2 23 .i.dl .i.xd3 24.i.xh5.i.xc4 25 l:xc4 li)xe3 26 fxe3 .i.d2 112-112 Parma-Geller, Capablanca Memorial, Havana 197 1.

21

2 2. . .'ii' h5 i s almost always played, but after 23 l:a6 l:fe8 24 l:xc6 l:h6 25 h4 l:g6 26 'ii'a6! 'ii'dS 27 l:b6! White was much better in Piculjan­ Bendana, corr. 2003. Therefore I recommend the text-move, which looks like it should give Black sufficient play. 23 l:a6 .i.f8 24 b3 .i.h3 It is not going to be easy for White to under­ take anything active after either 25 'ii'xd5 cxd5 or 25 fJ .l:.be8.

B) 20 li)e4 (D)

A) 20 .i.xdS It does not seem terribly logical to make this exchange now that Black can recapture with the queen. 20 'ii'xdS (D) •••

White is not willing simply to defend on the kingside, so he tries to fight for the initiative immediately. It looks rather strange to walk into a pin, however. 20 .i.fS! This equalizes pretty cleanly. Matters are less clear after 20 ....i.c7?! 21 .i.d2 l:fe8: a) 22li)c5? l:xel + 23 l:xel l:xel + 24.i.xel li)f4! 25 gxf4.i.xf4 26 h4 was Anand-Kamsky, Amber Rapid, Monte Carlo 1994. Here Black should have played 26 ... 'ii'xh4! 27 'ii'e4 'ii' h2+ 28 r3itfl "h3+ 29 'ii'g2.i.e2+ 30 'itgl .i.h2+ 3 1 �hl (or 3 1 'ii'xh2 'ii'fl #) 3l....i.fJ 3 2 'ii'xfJ 'ii'xf3+ 33 r3itxh2 'ii'e2, winning. b) 22 .i.dl ! .i.xdl 23 l:texdl (if we com­ pare this to Line C it is clear that White's pieces are much more active) 23 ... f5 24 li)g5 l:te2 25 'ii'xf5! (worse is 25 li)fJ? l:8e3! with counterchances, Anand-Kamsky, FIDE Can­ didates match (game I), Sanghi Nagar 1994) and here neither 25 ... l:f8 26 'ii' h 3! 'ii'g6 27 'ii' g4 l:ee8 28 f3 nor 25...li)f6 26 'ii'fJ 'ii'xf3 27 •••

21 c4 21 'ii'fl l:tfe8 is considered under Line D I, while 21 b3 does not look too threatening. Both 2 1 ... 'ii' h5 22 c4.i.b4 and 2 1 ....i.h3 22 f3 .i.f5 23 'ii'fl .i.c2 24 c4 'ii'f5 offer Black enough play for the pawn. 21 bxc4 22 'ii'xc4 Not 22 li)xc4? .i.h3. 22 l:tb8 ! ? •••

.••

UNDERSTANDING THE MARSHALL A TTACK

22

liJxf3 liJe4 28 .i.c 1 is sufficient for Black. Therefore, 20 . . . .i.c7 should be avoided. 21 �d2 (D) My computer recommends the mysterious move 21 .i.c l ! ?, but this can be met in the same way.

This is a typical idea. White prepares to sim­ plify with �d2 and liJe3 , when Black will have nothing at all to show for his pawn. Black has a choice of how to try to disturb White: 22 C l : 20 J:tfe8 C2: 20 .i.f5 24 ••

•••

Black should avoid 20 . . . f5 ? 2 1 �f4 ! (this surprising resource comes up a lot when Black has self-pinned with . . . f5 ; also possible is the computer's suggestion 21 �xd5 cxd5 22 'ii'xb5 �f3 23 liJd2 f4 24 �xf4 ::txe l + 25 ':xe l �xf4 26 liJxf3 'ii'xf3 27 gxf4 'ii'g4+ 28 'iti>f1 but the text-move is simpler). Even though White's pawns on the kingside become damaged, the simplifications favour White. After 2 1 . . J:he l 22 ':xe l �xf4 23 gxf4 .i.f3 24 liJg3 'ii'g4 25 ':e3 �e4 26 f3 ! .i.xd3 27 .i.xd5+ cxd5 28 fxg4, as given by Nunn, White has a winning end­ ing.

B

21.. .:xe4! 22 ':xe4 liJf6 23 f3 'ii'g6 24 'ii'n 24 ':ae l i s met by 24 . . .�xg3 ! , while after 24 'ii'e 3 Black can play either 24 . . . liJxe4 25 fxe4 .i.xe4 = or 24 . . . �xg3 25 hxg3 'ii'x g3+ 26 �f1 (26 'iti>h l 'ii'h 3+ = ) 26 . . . 'ii'h 2 ! 27 �c 1 'ii'h l + 28 'iti>e2 liJxe4 29 fxe4 �g4+, when Black has at least a perpetual check. 24 liJxe4 25 fxe4 .i.xe4 26 .i.f4 �d3 27 'ii'f2 �xf4 28 'ii'xf4 h6 29 'ii'f2 ':'e8 30 .l:lel ':xel+ 31 'it'xel �e4 32 'it'f2 The position is equal and Svidler-Kamsky, Groningen 1 995 was drawn here. •

CI) 20 .:.fe8 (D) •••

W

•••

C) 20 liJn (D)

Black doubles on the e-file and in doing so prevents White's consolidation plans because 2 1 �d2? runs into 2 1 . . .�f5 ! , when White's queen suddenly has no squares. 21 i.dl This is a typical retreat. White wants to ex­ change pieces. However, because there is no pressure on the a2-g8 diagonal any more, both the d5-knight and e6-rook are free to move about and an advance of the f-pawn is also less risky. 21 �xdI 221Wxdl •••

SPASSKY VARIATION

Worse is 22 :axdl?! 'it'f3, when White is tangled up. 23 'it'e2 'it'e4 24 �d2 (or 24 'it'd2 'it'f5) 24...'ii'c2 gives Black good play. 22 'it'f5 (D) •••

Black still has pressure on the e-file and the possibility of pestering White by advancing the h-pawn. Nunn considers that although Black should be able to hold the balance,White is the only one with real winning chances if he plays carefully. However, it is not so easy for White to play carefully and make progress at the same time if Black plays with sufficient energy. 23 iLd2 After 23 l:ta6 Nunn suggested 23 ...h5! 24 l:txc6 h4,which has since been tried in practice. 25 iLd2 and here: a) 25...:Xel 26 iLxel h3 27 f3 iLffl 28 l:ta6 b4 29 :a5? (A.Sokolov-Hellers, Reykjavik 1990) and now Black could have played 29...�xc3!' b) With 25...h3!?, instead of trading rooks, Black increases the tension on the kingside by creating threats around the white king. After 26 l:txe6 l:txe6 27 �e3 (27 f3 'it'd3!) 27...�xe3 28 iLxe3 'it'e4 29 l:tc8+ �f8 30 'iitn :a6! (not 30... 'it'g2+? 3 1 �e2 'it'xh2 32 d5 and 33 iLc5 is coming) the threat of 3l ...:al ! forced White into 3 1 l:txf8+ �xffl 32 'iite2 in Metz-Oim,COIT. 1988. Clearly Black is the one playing for a win here. 23 ':'xe1 The immediate 23...h5! ? is also possible. 24 �xe1 h5 25 h4 c5 Again, Black must take steps to interfere with White's plan of iLd2 and �e3. The point of this advance is that after 26 �d2 White will get an isolated pawn on d4, and taking on c5 •••

23

may open the position more than White would like. Instead 25...g6 26 b3 b4 27 c4 �c3 28 iLxc3 bxc3 29 c5 c2 30 'it'd2 iLc7 3 1 l:tcl 'ii'f3 32 ':'xc2 (32 'it'xc2 l:te2) 32 ...'it'xb3 33 'it'c3 'it'd5 34 'ii'c4 'it'e4 35 l:td2 was Gashimov­ Shirov, Poikovsky 2008. Even here, it should not be easy for White to convert his extra pawn, although he managed to do so after some over­ optimistic play from his opponent. 26 b3 (D) White prepares the possibility of c4 to kick Black's knight. After 26 'iitg2 cxd4 27 'ii'xd4 (27 cxd4 'it'g6 28 'it'f3 �f4+ 29
B

26 ...�f6!? This looks like a good try. Other moves look less trustworthy: a) 26...g6 was suggested by Nunn. This is a typical positional move that controls the light squares,protects the h5-pawn and gives Black's king a little more breathing room. Nunn gives 27 c4 bxc4 28 bxc4 �f6,continuing 29 d5 �g4 intending ...�e5 with counterplay. However, White may improve with 29 l:ta6!?, which dis­ turbs Black's coordination and may cause some slight problems. b) 26 ...cxd4 27 'it'xd4 l:te4 28 'ifd3 iLc5 and now 29 :a8+
UNDERSTANDING THE MARSHAll A TTA CK

24

is impossible because of 30 ......xf2+. White can try 29 �g2! though, when he is ready to start pushing Black back. 27 :a6 i.b8 28lLle3 After 28 dxc5 "'xc5 White is very passive, so Black's compensation looks sufficient for the pawn. 28......e4 (D)

d2 or he can play 21 1i'd2 straightaway - or he can repeat the position, which may be seen as a practical drawback of this line for Black. We have: C2l : 21 "'dl 24 C22: 2l "'d2 25 Note that 21 'it'e2 i.g4 will either repeat the position or transpose to Line e21 after 22 "'d2.

C21) 21 'it'dl White still wants to play 'iVd2 but prefers to lure Black's bishop to g4 first. 2l i.g4 22 1i'd2 "'h3 23 i.dl White is a bit tangled up, so he seeks simpli­ fications. 23 f4 is well met by 23 ... l:lfe8. 23 i.xdl 24 :axdl Black has good compensation after 24 'ii'xdl f5! 25 i.d2 :g6 (or 25 ... :xel 26 i.xel f4) 26 'ii'f 3 f4 27 'ii'g2 'ii'd7. 24 f5 (D) •••

•••

..•

29 dxc5 29 lLlg2 can be answered rather simply with 29 ... cxd4 30 cxd4 :d8. 29 i.xg3! Black has good chances. •••

C2) 20 i.f5 (D) •••

When there is no pressure on the a2-g8 diago­ nal, this move should be the first one Black looks at. The threat of ... f4 induces White's reaction. 25 f4 Now Black secures good play on the light squares. 25 :fe8 This is a positional approach. The brutal 25... g5 is also good enough: a) 26 fxg5 f4 27 i.xf4 i.xf4 28 gxf4lLlxf4 29lLlg3 "'g4 30 ':'xe6lLlh3+ 31 Citg2lLlf4+ 32 �gl lLlh3+ 33 'itg2 112- 112 Ljubojevic-Nikolic, Belgrade 1991. •••

This move is the modem choice. Black is go­ ing to chase the white queen to d2 to interfere with White's plan of i.d2 andlLle3. White can lure Black's bishop back to g4 before going to

SPASSKY VARIATION

25

White leaves well enough alone and places b) 26 'iWg2 'iWxg2+ 27 'ili>xg2 l:tfeS 2S .i.d2 . his queen on a square where it cannot be ha­ :xe l 29 l:txe l l:txe l 30.i.xe l gxf4 3 1 �f3 fxg3 rassed. 32 hxg3 'ili>f7 = Karpov-Short, Tilburg 199 1 . 21 l:tCe8 26 'ilVC 2 Black can also try a couple of other moves: After 26 .i.f2 'iWg4 the f4-pawn will fall. a) 2 1 . . .i.h3 . ! ? is a recent try from Aronian. 26 ..g4 27 .i.d2 l:te2! (D) This worked surprisingly well in Shirov-Aro­ nian, Bilbao 2009 : 22.i.d l 'iWg6 23.i.f3 'ilVf5 24 .i.h l l:lfeS 25 l:le2 ? ! h5 26 'iWc2 'ilVg4 27 l:lee l h4 2S 'iWd l 'iWf5 29 'iWf3 ? 'iWg6 ! 0- 1 . An odd­ looking resignation perhaps, but 30 . . ..i.g4 3 1 'iWg2 h3 is threatened and White gets demolished after 30 'iWe2 hxg3 3 1 hxg3 �xe3 32 �xe3 .i.xg3 ! -+ or 30 'ilVd l hxg3 3 1 hxg3 �xe3 32 l:txe3 l:txe3 33 fxe3 .i.xfl 34 c;Pxfl l:txe3 -+. b) 2 1 . . .i.e4 . and here: bI) 22 .i.c2 and now 22 . . . f5 ! ? was sug­ gested by Pavlovic. After 23 .i.d l 'iWh3 24 f3 f4 25 fxe4 fxg3 ! Black has serious counterplay. A simpler idea is 22 . . . .i.f3 23 .i.d l l:lfeS 24.i.xf3 'ii'xf3 intending . . . h5 and . . .�f6 and Black looks fine. This is a common idea. Black inches into b2) 22.i.d l 'ii'f5 (22 . . .'ifh3 and 22 . . .'iWg6 are White's position. also possible) 23 'ii'e 2 �f6 24 �d2 .i.d5 25 b3 28 h3 'ilV hS 29 g4 'iWxh3 c5 26 dxc5 .i.xc5 27 'iWxb5 . Black drew with Also sufficient is 29 . . . l:txe l 30 l:txe l (not 30 very exact play after 27 . . .i.xe3 . 2S fxe3 �e4 29 gxh5 ? l:txd l +) 30 . . . fxg4 3 1 l:txeS+ "xeS 32 l:[fl 'ilVe5 30 �xe4 'ilVxe4 3 1 l:tf3 l:td6 32 .i.e2 hxg4 "e4 =. 'iWe7 33 e4 'ilVxe4 34 'ilVd3 'ilVe6 35 l:te3 'ilVh3 36 30 l:txe2 "xg4+ 31 �g3 l:txe2 32 'iWxe2 .i.fl 'ii'h6 37 'iWe2 .i.xb3 = in Eiben-Moreira, 'iWxg3+ 33 c;Phl 'iVh3+ 34 �gl corr. 2006. This may be alright in correspon­ White cannot play 34 Wh2? 'ii'f3+. dence play, but over the board Black would 34 JWg3+ have to make a lot of 'only' moves to hold the After 35 'ilVg2 'iWd3 ! Black is not worse, so balance. White has nothing better than to repeat. We now return to the position arising after 2 1 . . .l:tfeS (D): e22) ...

...

••

21 'ilVd2 (D)

22 .i.xdS

UNDERSTANDING THE MARSHALL A TTA CK

26

22 .ltc2 .lth3 23 'ii'd I i.g4 24 'ii'd 3 'ii'h 3 25 .ltd2 l::te 2 26 :xe2 i.xe2 27 lle l .ltxd3 28 Axe8+ .ltf8 29 .ltxd3 'ii'd7 30 ':e l was agreed drawn in Svidler-Yakovenko, Foros 2008, while 22 .ltd l 'ii'g 6 23 i.f3 has been tried a few times in correspondence games. 23 . . . i.e4, 23 . . . h5 and even 23 . . . h6 ! ? all offer Black rea­ sonable play. 22 ... cxdS 23 'ii'e2 White gets nowhere with 23 i.f4 :xe l 24 Axe l Axe l 25 'ii'xe l .lte4 26 �d2 i.xf4 27 �xe4 dxe4 28 gxf4 'ii'g4+ 29 �f1 1/2-1/2 Ivan­ chuk-I.Sokolov, Bie1 1989. 23 'ii'xe2 Black would like to play 23 . . . .ltg4 24 'ii'xb5 :b8, but this fails to 25 'ii'aS ! with the point 25 . . . :xb2? 26 'ii'd 8+ i.f8 27 Aa8. However, the simple 23 . . . 'ii'h 3 is also possible, with the idea 24 f3 b4 with counterplay. 24 Axe2 b4 I suspect Black should be able to hold this typical Marshall endgame, but achieving a draw is not automatic. After 25 :ee l bxc3 26 bxc3 ':c8 27 .ltd2 .lte4 28 :e3 h5 ? ! 29 f3 i.f5 ? 30 :as White won the d5-pawn and the game in del Rio-Hebden, Port Erin 2003 .

01) 20 :re8 The increasing pressure on the e3-bishop forces White's response. 21 i.xdS 'ii'xdS (D) •••

•••

0) 20 'ii'n (D)

This is another common way of bolstering the kingside. In recent years this has been the most popular move amongst the world's elite players. Black has a choice: 26 D l : 20 :re8 28 D2: 20 i.h3 •••

•••

22 h3 This is the choice of the top players, but it does not lead to much. 22 f3 i.h5 23 'ii'f2 i.g6 makes it difficult for White to come up with a plan, while 22 c4 gives Black a choice between 22 . . . 'ii'f5 and Nunn 's suggestion 22 . . . bxc4, which both look fine for Black. 22 'ii'g 2 is a common move. As Nunn points out, after 22 . . . 'ii'h 5 (D) Black's queen is well placed, while the same cannot be said of White's. Some examples:

a) 23 f3 .lth3 24 'ii'f2 f5 (or 24 ... i.f5 intend­ ing . . . .ltd3 with counterplay) 25 f4 g5 (25 . . . i.g4 stopping �f3 is another idea) 26 fxg5 f4 27 i.xf4 :xe l + 28 Axe l Axe l + 29 'ii'xe l i.xf4 30

SPASSKY VARIA TION

gxf4 1i'g4+ 3 1 � f2 'i¥h4+ 1/2- 1/2 Tiemann-EIburg, COIT. 2002. . b) 23 ':a6 f5 (23 . . . ii.h3 24 1i'f3 ii.g4 is also possible because 25 1i'h l f5 does not look very appetizing for White) 24 d5 ':h6 and here: b l ) 25 f3 ii.h3 26 1i'f2 f4 27 ii.d4 ':xe l + 28 1i'xe l ii.c8 ! 29 ':a8 1i'xh2+ 30 �n 1i'h3+ 3 1 �e2 cxd5 + Readers of Prisyv-Estrin, COIT. 1982. b2) 25 f4 ii.xf4 ! ? (or 25 ... ii.c5 26 liJn ii.f3 27 'ilVd2 ii.f8 28 ':xc6 ':xc6 29 dxc6 ii.xc6 with compensation) 26 gxf4 (not 26 liJn ii.f3 27 1i'd2 ii.xe3+ 28 ':xe3 ii.xd5 +) 26 . . . ii.h3 27 1i'g3 (27 1i'f2 ng6+ 28 �h l ii.g2+ 29 1i'xg2 ':xg2 30 �xg2 1i'g4+ 3 1 �f2 { 3 1 �h 1 ':xe3 } 3 l . . .�h4+ 32 �n 1i'h3+ 33 �f2 1i'xh2+ +) 27 . . . ':g6 28 ':xc6 ':xg3+ 29 hxg3 1i'f7 30 ':c5 1i'g6 3 1 �f2 ii.g4 gave Black the initiative in an unclear position in T .Olafsson-Harding, COIT. 200 1 . 22 ii.h5 Black can also play 22 . . . ii.f5 23 1i'g2 1i'xg2+ (or 23 . . . h5 ! ? 24 �xd5 cxd5 25 �g2 h4 26 g4 ii.c2 with compensation, Bergmann-Sakai, COIT. 2003) 24 �xg2 ':6e7 25 b3 f6 26 -Ua2 ii.e6 27 c4 ii.b4 28 ':c 1 ii.f5 29 g4 ii.d3 when he had some compensation in Leko-Kasimdzhanov, Linares 2005 . White can push for a while, but these types of endings are difficult to win. 23 1i'g2 (D)

27

in I.Gurevich-Benjarnin, Reshevsky Memorial, New York 1 992, is a typical ending that White will not ever win) 25 . . . ii.xg3 26 fxg3 ':xe3 27 ':xe3 ':xe3 28 liJn ':e2 29 g4 ! (29 ':xb5 ? ii.f3), with a big advantage for White, is given by Gershon and Nor. 24 �xg2 f5 (D)

•.•

23 1i'xg2+ This is a bit of a pre-emptive improvement. After 23 . . . f5 24 1i'xd5 cxd5 White can play 25 l:ta5 ! (25 liJn f4 26 ii.d2 ':xe l 27 ':xe l l::txe l 28 ii.xe l ii.e2 29 liJh2 �f7 30 �g2 �f6 3 1 f3 h5 32 g4 g6 33 liJn hxg4 34 hxg4 g5, as played

25 liJf3 25 l!a6 f4 26 gxf4 ii.xf4 27 ii.xf4 ':xe l 28 .l:Ixc6 is the only try for an advantage but it is not completely without risks. 25 ':gl ! ? is an odd-looking move. Pavlovic gives 25 . . . .:f6 ! (with ideas such as . . .f4 and . . . ii.f7 -d5+) 26 liJb3 ii.f7 27 liJc5 ii.d5+ 28 �h2 f4 29 gxf4 ii.xf4+ 30 ii.xf4 ':xf4 3 1 liJd3 ':ff8 32 ':a7 g6 with sufficient compensation for the pawn. 25 ...f4 26 ii.d2 fxg3 27 ':xe6 ':xe6 (D)

.•.

28 ':a8+ Not 28 fxg3 ? ':e2+. Now Black wins back the pawn and a sterile equality results.

UNDERSTANDING THE MARSHALL A TTACK

28

28 oltrs 29 .!Z)eS gxf2 30 'ottxf2 l:te8 31 l:ta6 oltd6 32 ii.r4 ii.xeS 33 ii.xeS ':e6 The position is equal and the game J.Polgar­ Adams, World Ch, San Luis 2005 was drawn here. •••

02) 20 ii.h3 (D) •••

ii.d5 26 ii.xf4 l:txe l + 27 'ifxe l oltxf4 28 oltxd5 cxd5 29 .!Z)f3 ± is given by Nunn. b) 22 . . . oltf4 ? ! gives White two good op­ tions: bl) 23 'iff3 R.xe3 (23 ... l:tfe8 24 gxf4 l:tg6+ 25 �hl R.g4 { or 25 . . . ii.g2+ 26 'ifxg2 l:txg2 27 �xg2 +- } 26 'fi'e4 ! +- is given by Nunn) 24 'fi'xf5 (24 fxe3 'it'd3 ! ?) 24 . . . ii.xf5 25 fxe3 '!z)xe3 26 ii.f3 and White still has pressure. b2) 23 .!Z)fl ii.xfl 24 l:txfl .!Z)xe3 25 fxe3 ii.xe3+ 26 �g2 'it'd5+ 27 'fi'f3 'fi'xf3+ 28 ii.xf3 b4 29 d5 cxd5 30 ii.xd5 lIe7 3 1 cxb4 ii.d4 was Fumero Sanchez-Hayden, COIT. 1 99 1 . White could now play 32 b5 ±. c) 22 . . . l:tfe8 ! ? (D).

This is the most obvious and direct move. 21 ii.dl This counterattack is a typical tactical re­ source to disturb Black's play. Instead 2 1 'ife2 ii.g4 just invites a repetition of moves. 21 'ifrs 22 'ife2 (D) •••

22 cS This is the choice of the world's top players. Alternatives: a) 22 ... .!Z)f4 ? ! 23 'iff3 'ifg6 24 ii.b3 (24 olte2 ? ! .!Z)xe2+ 25 'ifxe2 oltf5 with compensa­ tion) 24 . . . oltg2 (24 . . . oltg4 25 'it'xc6) 25 'ifd l •••

Harding says this "completely equalizes", but the elite all play 22 . . . c5 instead. . . This is still Black's best alternative to 22 . . . c5 and could use more tests. c 1 ) After 23 .!Z)f3 .!Z)f4 ! the knight will come to d3 , while 23 .!Z)fl c5 24 ii.c2 .!Z)xc3 25 bxc3 'fi'd5 26 f3 ii.xfl 27 'fi'xfl ':xe3 28 ii.e4 was drawn here in Herbrechtsmeier-Oim, COIT. 1988 because after 28 . . . lIxe l 29 lIxe l 'fi'b3 30 oltxh7+ �f8 3 1 ':xe8+ �xe8 Black will have no problems holding. c2) 23 'ii'f3 .!Z)xe3 ? ! (Black could instead try 23 . . . 'it'd3 ! ?) 24 ':xe3 ':xe3 25 fxe3 'ii'd 3 26 'ife2 'ifxe3+ 27 'ifxe3 ':xe3 was agreed drawn here in Hauptmann-Moros, COIT. 1976, but after 28 l:ta8+ ii.f8 29
SPASSKY VARIA TION

a) 23 �f1 presents Black with no problems: 23 . . . cxd4 24 cxd4 �b4 2S ':'a3 �c6 ! 26 l:ld3 .tb4 27 dS ':'d6 28 .td2 l:lxdS 29 .txb4 �xb4, Sax-I.Sokolov, Haninge 1 989. b) 23 �b3 is rare, but Black should be pre­ pared for it. 23 . . . �f4 ! (this will end up winning the white queen but Black should not relax just yet) 24 'iVf3 .tg2 2S 'iVg4 'iVdS 26 gxf4 ':'g6 27 'iVxg6 fxg6 28 �xcS .txcS 29 dxcS and now 29 . . . .th3 30 f3 gS? (30 . . . ':'e8 ! keeps the game in the balance) 3 1 .td4 l:.xf4 32 %leS 'iVc4 33 .te2 'iVb3 34 :e7 �f8 3S ':'b7 1 -0 was L.Mauro­ A.Coscia, corr. 1 996. Instead 29 . . . .th l 30 .tb3 (30 f3 .txf3) 30 . . ...xb3 3 1 �xh l was given by Van der Tak, continuing 3 l . . .'ilxb2 32 .td4 b4 33 ':'a7 %ld8 34 ':'xg7+ ct>f8 3S ':'xh7 ':'xd4 36 cxd4 'ii'xf2 =, and 3 l . . .hS ! ? is also possible if Black wants to try to win. We now return to 23 �f3 (D) :

29

w

This is a critical position in the Spassky Vari­ ation. Black has two bishops and an active posi­ tion, but White is a pawn up and he can fight in the centre with Black's bishops both on the h­ file. 27 .te2 White can also try 27 �eS 'ii'xf2+ 28 'ili>xf2 f6, and here: a) 29 .tg4 ? ! ':'xeS ! 30 .txh3 .txe3+ 3 1 �f3 cxd4 32 l:Iad l dxc3 33 bxc3 .tb6 34 ':'xeS ':'xeS 3S ':'dS ':'e3+ 36 �g2 ':'xc3 37 ':'xbS ':'c2+ 38 �h l .tc7 left Black a pawn up in Leko-Krarnni k, Amber Blindfold, Monte Carlo 2007, although White should be able to draw this. b) 29 �g4 (D) is much more challenging.

B 23 .tf4 This is the only good way to increase the pressure. 23 . . . �f4 ? ! is strongly met by 24 'ii'c 2 ! (better than 24 'ii'x bS �d3) 24 . . . �d3 2S dxcS �xe l 26 'iVxfS .txfS 27 �d4 ! and with four pieces hanging Black has problems, despite his temporary extra rook ! 24 'iVd2 After 24 .tc2 "g4 ! Black has no problems. For example, 2S 'it'd3 (also harmless are 2S �h4 'ii'xe2 26 l:lxe2 .txe3 27 fxe3 cxd4 28 cxd4 �xe3 = and 2S �eS 'iVxe2 26 ':'xe2 .txeS 27 dxeS ':'xeS =) 2S . . . g6 26 �eS .txeS 27 dxeS 'ilf3 28 'ii'e4 'ii'xe4 29 .txe4 �xe3 30 :xe3 ':'xeS = Svidler-Kramnik, Amber Rapid, Monte Carlo 2007 . 24 �xe3 25 fxe3 .th6 26 "f2 .:.re8 (D) •••

•••

Black must play very precisely here: b l ) 29 . . . .txg4? ! 30 .txg4 fS (30 . . . ':'xe3 3 1 dxcS and 30. . ..txe3+ 3 1 �g2 :6e7 32 dS look even worse) 3 1 .tf3 .txe3+ 32 'ili>g2 'ili>f8 33 dxcS .txcS 34 ':'xe6 ':'xe6 3S ':'a5 ':'b6 36 .te2 and Black is in for a difficult defence.

UNDERSTANDING THE MARSHALL A TTA CK

30

b2) 29 . . . .i.g5 ! 30 dxc5 (White is two pawns up, but the knight on g4 is trapped) 30 . . . 'ifi'f8 (30 . . . h5 is well met by 3 1 .i.b3 ! .i.xg4 32 l:ta5 , but another possibility is 30 . . . l:te4 3 1 .i.f3 l:txg4 32 .i.xg4 .i.xg4 33 l:ta5 l:td8) 3 1 'ifi'g l l:te4 32 tiJf2 l:txe3 33 .i.e2 (33 l:txe3 .i.xe3 is fine for Black after either 34 c6 11e5 or 34 b4 .i.xc5 ! ) 33 . . . l:txe2 3 4 ':xe2 l:txe2 35 tiJxh3 .i.e3+ 36 'ifi'h l .i.xc5 37 tiJf4 ':e5 and Black has finally equalized. 27 'ilkh5 28 e4 After 28 tiJe5 f6 29 .i.d l 'ilkg5 30 tiJd3 Black can play 30 . . . 'ilkxe3 ! . 28 l:tf6 2 9 .i.dl (D)

30 e5 l:tf5 (D)

W

•••

•••

31 'ilke2! Now 3 1 . . . .i.g4 can be met by 32 'ilt'xb5 . Black's position quickly collapses. 31 cxd4 32 cxd4 l:td8 33 'ilke4 'ilkg4 34 'ii'xg4 .i.xg4 35 tiJh4 l:tg5 36 .i.xg4 l:txg4 37 tiJf3 1 -0 Wang Hao-Grishchuk, Russian Team Ch, Dagomys 2008.

B

•••

Concl usions 29 g6? Grishchuk seems to have mixed up his prep­ aration. The position is unclear after 29 . . . .i.g4 30 ':fl .i.h3 (or 30 . . . g6 3 1 e5 ':f5 32 h4 g5 33 'ilkg2 cxd4 34 hxg5 .i.xg5 35 tiJxg5 'ilkxg5 36 .i.xg4 Wxg4 37 ':xf5 Wxf5 3 8 cxd4 Wd3 1/2- 1/2 B artsch-Utesch, COIT. 2007) 3 1 e5 (3 1 l:te l .i.g4 repeats) 3 l . . ..i.xfl 32 exf6 'ii'h 3 33 tiJh4 .i.d3 , as in Balabaev-Leenders, COIT. 2006. •••

The Spassky Variation is holding u p quite well in top-level play, which is a tribute to its sound­ ness. After 20 tiJfl both the older 20 . . .':fe8 and the more modem 20 . . . .i.f5 look fine for Black. 20 'ilkfl looks more challenging. Even though Black should hold the endgames that arise af­ ter 20 . . . :fe8, most players prefer the forcing 20 . . . .i.h3 . Both 27 .i.c2 and 27 tiJe5 are critical, but I believe Black can equalize against both moves with precise play.

2 Old M a in Line : 1 8 . . . fS/ 1 8 . . . bxa4

1 e4 e5 2 tLlf3 tLlc6 3 .i.b5 a6 4 .i.a4 tLlf6 5 0-0 .i.e7 6 :tel b5 7 .i.b3 0-0 8 c3 d5 9 exd5 tLlxd5 10 tLlxe5 tLlxe5 1 1 :Xe5 c6 12 d4 iLd6 13 l:tel fih4 14 g3 fih3 15 .i.e3 .i.g4 16 fid3 l:tae8 17 tLld2 ::te6 18 a4 (D)

B

Black is at an important crossroads. White threatens 1 9 axbS axbS 20 .i.xdS cxdS 2 1 fixbS . Black can play 1 8 . . . fihS to protect dS, which was covered in Chapter 1 . The alternative is to play for a direct attack on the kingside with . . . fS . Black can begin this plan by playing ei­ ther 1 8 . . . fS or 1 8 . . . bxa4, which often lead to the same position. The move-order that Black chooses should depend on the deviations for White that are possible after each move. The two ways to reach the Main Line are 1 8 . . . fS 1 9 fin fihS 20 f4 bxa4 2 1 ':'xa4 and 1 8 . . . bxa4 1 9 :'xa4 fS 20 fin fihS 2 1 f4 (D). In both cases we reach the position at the top of the next column. It is from here that we form the basis for most of this chapter. The move-order that Black chooses will allow White certain deviations, however. Some of these sidelines are danger­ ous, and some are not. Even though 1 8 . . . fS is the traditional way of reaching the Main Line, I think that 1 8 . . . bxa4 is the better way to head for the diagrammed position, so anyone want­ ing to play this variation would be well served learning this move-order. In any case, the two

methods to reach the Main Line should be studied together because there are many recur­ ring ideas that are good to know. The variations in this chapter are very tacti­ cal in nature and the play is a real slugfest, with both sides trying to get in the bigger and faster punch. Theoretically, Black has been consid­ ered to be on rather shaky ground, but there is still unexplored territory despite the line's his­ torical popUlarity in both over-the-board and correspondence play. In any case, this chapter shows a lot of thematic ideas in the Marshall Attack that will help Black find his way in other variations as well. Section 2.1 : The Old Road 18 f5? ! 31 Section 2.2: The New Road 18 bxa4 39 Section 2.3: The Main Line 42 •••

•••

Section 2 . 1 : The Old Road l S . . . fS? ! 1 e4 e5 2 tLlf3 tLlc6 3 iLb5 a6 4 .i.a4 tLlf6 5 0-0 iLe7 6 l:tel b5 7 iLb3 0-0 8 c3 d5 9 exd5 tLlxd5 10 tLlxe5 tLlxe5 1 1 l:txe5 c6 12 d4 iLd6 13 l:tel fih4 14 g3 fih3 15 iLe3 iLg4 16 fid3 l:tae8 17 tLld2 l:te6 18 a4 f5? ! (D) Black ignores White's demonstration on the queenside and threatens . . . f4 and . . . l:th6. If White meets a subsequent . . . 1Ih6 with tLln ,

32

UNDERSTANDING THE MARSHAll. A TTA CK

for Black. For example, 2 1 . . . .i.e2 22 l:xe2 :xe2 23 bxc6 ':xd2 24 .i.xdS+ �h8 25 'ii'n looks better for White. All is not lost, however, and 1 9 f4? ! probably i s a mistake i n view o f 1 9 . . . bxa4 ! with the idea 20 ':xa4? .i.xf4 ! , when 2 1 gxf4 is met by 2 1 . . .':g6 and B lack indeed has a strong attack. Thus we are left with two options for White. The first is the road White takes to head to the Main Line, while the second is the very danger­ ous 'Internet Refutation' . We have: A: 19 'iVn 32 37 B : 19 axbS! ? then . . . .i.f3 comes and White no longer has 'ii'n available to parry the mate threat on g2. Now, the Main Line is 1 9 'ii'n 'ii'h5 20 f4 bxa4 2 1 ':xa4, but there are a couple of other possibili­ ties, of which one is very important. Note that 1 9 .i.xd5 cxd5 20 Wn Wh5 2 1 f4 will also transpose into main lines. After the poorly-timed 19 f4? ! , the response 1 9 . . .l:.fe8 is always given as the refutation, but I am not sure about this move. After 20 axb5 (worse is 20 .i.f2? .i.e2 2 1 Wc2? .i.xf4 ! as given by Nunn, one possibility being 22 axb5 .i.xd2 23 'ii'xd2 .i.f3 24 .i.e3 axb5 ! -+) Black plays 20 . . . .i.xf4 (D).

A) 19 'ii'n (D)

This is the traditional move. White evicts the black queen and in doing so escapes potential pins along the third rank and prepares to physi­ cally block the advance of Black's f-pawn by playing f4 himself. 19 'iVhS 20 f4 The only real alternative is 20 axb5 (D). This was recommended by Shamkovich and although it is not bad, it never really caught on. Black has : a) 20 . . . axb5 ? ! should favour White because it will almost certainly help to have an open a­ file for the rook. 21 .i.xdS (2 1 f4 is also possible, leading to the note to Black's 20th move below, where Black avoids 20 . . . bxa4, thus allowing 2 1 axb5) 2 1 . . .cxd5 2 2 'ii'xb5 f4 2 3 .i.xf4 .i.xf4 24 ':xe6 .i.xe6 25 gxf4 'ii'g 6+ 26 �h l 'ii'c 2 gives Black some counterplay, but it is hard to believe that it is enough for the missing pawns. .•.

This has been considered to be virtually win­ ning for Black because, but after 2 1 gxf4? both 2 1 . . . ':g6 and 2 1 . . . ':h6 give Black a strong attack while 2 1 bxc6? .i.xg3 ! (better than 2 1 . . .':xe3? 22 .i.xd5+ 'it>h8 23 ':'xe3 .i.xe3+ 24 �h l .i.xd2 25 c7 ! ) 22 hxg3 'iVxg3+ leads to mate after ei­ ther 23 �h l .i.f3+ or 23 'it>n liJf4 ! . However, after 2 1 .i.f2 ! I cannot find a good continuation

Ow MAIN LINE:

18. . .f5118 . . bxa4 .

33

Black is pinned along the a2-g8 diagonal. It is an interesting situation where neither the d5knight nor the e6-rook is pinned, but if one piece moves, the other one will be. For this rea­ son B lack sometimes spends a tempo playing . . . �h8 to free up these pieces. 20 bxa4 Black avoids the threatened 2 1 axb5 axb5 22 .i.xd5 cxd5 23 'ii'xb5, opens the b-file, and lures White's rook off the first rank. The a6pawn is left to its fate, but Black hopes that he will be able to generate enough kingside play. To see why this move is supposed to be neces­ sary, let's look at Black's other sensible-look­ ing moves: a) 20 ... :'fe8? ! (D). •••

b) 20 . . . f4 is a more sensible attempt to take advantage of White's move-order and it is prob­ ably stronger as well. 2 1 .i.xf4 iLxf4 22 lhe6 JLxe6 23 We i ! ? (after 23 gxf4 axb5 Black has compensation according to Nunn, and this as­ sessment has held up in correspondence games) and now: b l ) 23 . . . JLh3 24 'iie4 (24 bxc6 ! ? �h8 25 iLxd5 'iixd5 26 'iie4 may well be a better op­ tion) 24 . . .'�h8 25 l£ln (25 bxc6 l£lxc3 26 bxc3 .i.xd2 27 'iid 3 ltxf2 ! 28 �xf2 JLf5 1eads to per­ petual check) 25 . . . lIe8 26 'ii'd3 axb5 27 gxf4 l£lxf4 28 l£lg3 'ii'h4 gave Black compensation in McKenna-C.Chandler, corr. 1 990- 1 . b2) 2 3. . .:'e8 24 bxc6 l£lc7 looks rather un­ clear. We now return to 20 f4 (D) :

This move is a bit ugly positionally, but something had to be done about the advance of Black's f-pawn. On the plus side it gives White some more space and controls the e5- and g5squares. Although the e3-bishop is a bit loose,

Black tries to counter White's attack on b5 and d5 by threatening the e3-bishop, but this natural move is probably just bad for tactical reasons. 2 1 axb5 ! axb5 (no better is 2 1 . . .:'xe3 22 :'xe3 l:he3 23 bxc6 ! lIe2 24 JLxd5+ and af­ ter either 24 . . . �f8 or 24 . . . �h8 White will play 25 h3 ! JLxh3 26 JLf3 leading to a winning end­ ing for White) 22 JLxd5 cxd5 23 'ii'xb5 'ii'f7 (23 . . . JLxf4 24 'ii'xd5 is simply winning for White, as pointed out by Nunn, while 23 . . . l:[xe3 24 11xe3 11xe3 25 'ii'x d5+ also wins for White) 24 JLf2 ! ± is given by Nunn. Black is just two pawns down. b) 20 . . . g5 ? ! (D) is a typical thrust that we shall see again and again. Black wants to break down White's pawn­ chain on the dark squares at all costs. However, I think this move does not show enough respect for White's resources and, although compli­ cated, I do not think this line holds up. 21 axb5

34

UNDERSTANDING THE MARSHALL A TTA CK

and forced - Black must play for destruction. 24 l:txe3 f4 25 l:tf3 ! �xf3 26 'ifxf3 "ifxf3 27 lZ)xf3 fxg3 was Fridel-Elent, COIT. 1 996. White can now play 28 �g2 gxh2 29 lZ)xh2 �xh2 30 �xh2 l:tf2+ 3 1 �g3 l:txb2 32 �f4 ±. The fire­ works are over and White has a big advantage in the endgame. We shall see a similar version of this forcing play in other positions. Here it just does not work, and this line looks like enough reason for Black to avoid 20 . . . g5 . c) 20 . . . �h8 (D).

axb5 22 �xd5 ! (22 fxg5 ? ! allows Black to es­ cape after 22 . . . f4 ! 23 �xf4 �xf4 24 gxf4 1::tx f4 25 'ifd3 �e2 ! , when White is forced to take perpetual check by 26 l:ta8+ �g7 27 l:ta7+ 'iti>f8 28 l:ta8+ 112- 112 Roelens-Elent, COIT. 1 996) 22 . . .cxd5 (D) and now:

w

w

b l ) 23 'ifxb5 ? ! l:th6 24 lZ)fl (after 24 'ifxd5+ �h8 25 'ifg2? gxf4 26 �xf4 �xf4 27 gxf4 �h3 + Black's attack is very strong) 24 . . . �f3 25 fxg5 �xg3 26 l:te2 f4 27 gxh6 �xe2 28 'ifc6 �h4 29 'ife6+ 'iti>h8 30 �xf4 �f2+ 3 1 �xf2 l:txf4+ 32 'iti>gl was drawn in Pietrocola-Elent, COIT. 1 999 because 32 . . . l:txfl + ! 3 3 l:txfl 'ilg5+ leads to stalemate after either 34 �h l �f3+ 35 l:txf3 'ifg2+ or 34 �f2 'iff4+ 35 �xe2 'ifxfl + 36 �xfl . b2) 23 'ifg2 ? ! is murky after 23 . . . l:tfe8 24 'ifxd5 (24 �f2 "iff7 25 l:txe6 l:txe6 is also un­ clear) 24 . . . "iff7 25 l:ta8 gxf4 26 l:txe8+ l:txe8 27 "ifxd6 fxe3 . b3) By playing 23 fxg5 ! , White basically calls Black's bluff. Now 23 . . . l:txe3 is thematic

This line is also supposed to be dubious, but maybe it is not so bad. By breaking the pins on the a2-g8 diagonal, Black threatens the e3bishop so White's reply is forced. Black still loses time and does nothing to address the queenside problems, but White has not found a convincing refutation yet. 2 1 �xd5 (2 1 �f2 l:th6 22 "ifg2 �h3 23 'iff3 �g4 is a draw) 2 1 . . .cxd5 22 axb5 and now: c 1 ) 22 . . . 'ife8? 23 bxa6 ! (Black must always be wary of sacrifices of this nature) 23 . . . l:txe3 24 a7 %:txel 25 'ifxe l "ifxe l + 26 l:txe l l:ta8 27 l:te6 �c7 (St.Collins-P.Barrett, COIT. 1 999) and now 28 �f2 ! (to stop . . . �e2) followed by 29 l:ta6 gives White a big advantage. c2) 22 . . . axb5 23 'ifxb5 (23 "ifg2 ! ? is also possible) 23 . . . l:th6 24 h4 ! (24 lZ)fl �f3 25 b3 g5 ! gave Black good play in Arias Duval-Gim­ enez, COIT. 2(03) 24 . . . g5 25 fxg5 �xg3 26 gxh6 �h2+ 27 �h l and now instead of 27 . . . 'ifxh4? 28 �g5 ! 'ifxg5 29 l:te8 +- Hage-Horak, COIT. 1 999, Black should play 27 . . . �b8 28 �g5 �f3+ 29 lZ)xf3 'ifxf3+ 30 �gl 'ifg3+ 3 1 �fl 'iff3+ with a draw. This is a line that could be ex­ plored further.

Ow MAIN LINE:

1 8. . .f5118. .

.

bxa4

35

21 i.xd5! ? 2 1 l:txa4 i s the Main Line, and will b e con­ sidered in Section 2.3. 21 cxd5 (D) •••

22 'ii'g2 (D) White attacks the d5-pawn, which is rather difficult to defend. 22 l:txa4?? is a surprisingly common blun­ der because of 22 . . . 'ii'e 8 ! . This is a typical tac­ tical idea by which Black attacks the a4-rook and e3-bishop. White resigned immediately in Renet-Nunn, European Team Ch, Haifa 1 989, while 23 i.f2 %he l 24 'ii'xe l 'iVxa4 25 'ii'e 6+ l:tf7 26 'ii'x d6 'ii'd l + 27 tl)fl h6 28 'ii'd 8+ �h7 29 'ii'x d5 i.h3 0- 1 was Kindermann-Lukacs, Budapest 1 987.

22 J:tce8 This is the most sensible move, simply offer­ ing the d5-pawn, but Black has also tried to shield the pawn with the exchange sacrifice 22 . . . :e4 (D). ••

This interesting idea will probably not hold up to modem (i.e., computer-assisted) scrutiny : a) 23 tl)xe4 ? ! fxe4 24 ':xa4 g5 25 %1xa6 gxf4 (not 25 . . . i.f3 ? 26 'ii'f2 i.xf4 27 l:.a5 ! +-) and here: a l ) 26 gxf4 and now 26 . . J:tf6 ? ! is line 'b3 ' below, but 26 . . . �h8 ! holds, since 27 l:txd6 (27 ':eal ':g8 28 l:ta8 i.f8) 27 . . . l:tg8 28 f5 i.f3 29 'ii'x g8+ Wxg8 30 l:td8+ leads to perpetual check. a2) 26 ':xd6 fxe3 27 l:txe3 i.h3 28 g4 wins a rook but not the game after 28 . . . 'ii'h4 29 'ii'x h3 'ii'f2+ 30 �h l l:ta8 3 1 ':d8+ l:hd8 32 g5 l:ta8 33 'ii'e6+ and White must give perpetual check, as has occurred in a few games. b) Therefore White tends to ignore the rook, at least for the time being: 23 ':xa4 g5 24 l:txa6 gxf4 and now: b l ) 25 tl)xe4?! fxe4 transposes to line 'a' above. b2) 25 ':xd6 ? ! fxe3 (25 . . . l:txe3 ? 26 'ii'x d5+ Wh8 27 ':xe3 fxe3 28 'iVe5+ is winning for White, as pointed out by Nunn) 26 tl)xe4 fxe4 transposes to line 'a2' . b3) 25 gxf4 is probably best. 25 . . . l:tf6 26 tl)xe4 fxe4 and now: b3 l ) 27 i.f2 l:tg6 28 ':xd6 (28 i.g3 ! looks good for White) 28 . . . %1xd6 29 i.g3 i.f3 30 'ii'f l l:tg6 3 1 �f2 is not so clear, as pointed out by Nunn. Despite the two extra pawns, the op­ posite-coloured bishops and White's draughty king give Black reasonable chances. b32) 27 h3 ! 'ii'x h3 (27 . . .l:tg6 28 l:txd6 ! i.e6 29 ':xe6 ':xg2+ 30 �xg2 'ii'f3+ 3 1 �h2 leaves White with way too much for the queen) 28 'ii'x h3 i.xh3 29 �f2 ±. If Black cannot improve here (and he probably cannot), then 22 . . . l:te4 has to be discarded.

36

UNDERSTANDING THE MARSHALL A TTA CK

Now back to 22 . . . ltfe8 (D).

29 . . . .i.xh3+ 30 �xh3 'ii'xd2 3 1 'iVxd6 'iVxf2 32 'ii'f8+ �h7 33 'ifxf5+ �h8 34 'iVd3 'ifxb2 35 d5 a3 36 d6 'ii'b6 37 d7 'ii'd8 was eventually drawn in Filipek-Miciak, COIT. 2003 . c) 28 h3 ! ? (again, this small move is a rather testing idea) 28 . . . .i.xh3 29 tt)c4 .i.c7 30 tt)e3 'ii' b 8 3 1 'ifa2 'ifb7 32 'it'xa4 'ii'x b2 33 'ii'xa6 'ii'x c3 34 'iith 2 JLg4 35 tt)xg4 fxg4 36 'it'c8+ �h7 37 'iff5+ �g8 38 'ii'xg4 ± Nebel-Polevoy, COIT. 2oo3 . We now return to 24 . . . iLe2 (D):

w

23 'it'xdS �h8 24 .i.f2 .i.e2 Black maintains the tension. The alternative is 24 . . . .:.xe l + 25 ':'xe l ':'xe l + 26 .i.xe l 'ii'e 8 27 iLf2 h6 ! . Black takes care of his back-rank problem in preparation for a . . . 'iVe2 invasion. White has tried a few moves here: a) 28 'ifxd6 (28 tt)c4 'iVe2 29 tt)e3 iLh3 30 'iVxd6 is the same) does not lead anywhere. Af­ ter 28 . . . 'ife2 White is a piece up but he cannot win. In fact, he can even lose if he is not careful: 29 tt)fl .i.h3 30 'it'b8+? (30 tt)e3 'it'xb2 31 .i.e l a3 32 'it'f8+ �h7 33 g4 ! and White will deliver perpetual check himself) 30 .. .'i!th7 3 1 �3 'ifd3 32 tt)g2 'it'bl + 33 tt)e l (D) (or 33 .i.e l 'it'e4).

B

2S �g2 White wants to play tt)f3-e5 . The alternative is 25 ltxa4? ! , and then: a) 25 . . . .i.c4? 26 'it'xc4 ! (worse is 26 l:txc4 ':'xe l + 27 .i.xe l l:he l + 28 tt)fl lf2_ 1f2 Popovic­ Pinter, Hastings 1 9801 1 ) 26 .. .l:be l + 27 .i.xe l ':'xe l + 28 tt)fl 'ii'e 8 and now : a 1 ) 29 �f2 l:tbl 30 ':'a2 'ii'e l + 3 1 <;t>g2 'ii'e4+ 32 �h3 (32 �gl 'ii'e 3+ repeats) 32 . . . g5 ! gives Black enough counterplay. a2) 29 ':'xa6 ! looks convincing as 29 . . . 'ife3+ 30 �g2 ':'e2+ 3 1 �h3 +- just wins for White. b) 25 . . . iLxf4 (this idea of Nunn's is best) 26 gxf4 iLc4 27 ltxc4 ltxe l + 28 tt)fl (or 28 .i.xe l ltxe l + 29 tt)fl ':'xfl + 30 �xfl 'ii'd l + - Nunn) 28 . . .'it'g4+ 29 iLg3 h5 is unclear. 2S h6 26 tt)f3 26 ltxa4 ? ! .i.xf4 27 gxf4 'iVg4+ 28 .i.g3 ':'e3 29 :aa l ! (not 29 �f2? .i.d3 ! 30 ':'xe3 ':'xe3 with decisive threats, Lang-Ramon, COIT. 1 985) 29 ... l:.d3 ! gives Black sufficient counterplay. 26 .i.xf4 27 tt) gl ! This is forced. Nunn gave 27 gxf4? 'iVg4+ 28 iLg3 iLxf3+ 29 'ii'xf3 lte2+ 30 ltxe2 ltxe2+ 3 1 'iff2 as likely to lead to a draw, but in fact Black •••

33 . . . a3 ! 0- 1 Ernst-Sammalvuo, Osterskars 1 995 . b) 28 �g2 'ii'e2 29 h3 (29 tt)c4 .i.xf4 30 h3 { not 30 gxf4? .i.h3 + ! mating } 30 . . . .i.xh3+ 3 1 �xh3 'ifxf2 32 gxf4 'iffl + is a draw because the white queen must protect the c4-knight)

•••

Ow MAIN LINE:

i s winning after 3 1 . . .g5 ! 3 2 h 3 :xf2+ 3 3 xf2 'ii'x h3 34 d5 gxf4 35 i.xf4 'ii'd 3 . 27 i.d2 28 1lxe2 :xe2 29 �xe2 'ifxe2 30 'ii'xf5 i.e3 (D)

18. . .f5118. . . bxa4

37

B) 19 axb5! ? (D)

•••

B w

Black is hoping to capture on b2 and push the a-pawn. 31 l:.e1 ! Not convincing is 3 1 'iff7 i.xf2 32 'ifxf2 'ifb5 33 'iff7 'ifxb2+ 112_ 112 Matsukevich-Filip­ pov, COIT. 1 983. 31 'ifxb2 3 2 'iff7 :g8 3 3 l:txe3 a3 (D) •••

w

34 d5! This allows Black to get a second queen, but it is a clear-cut win for White. Too hasty is 34 .l:.e6? a2 35 l:.xa6 a 1 'if 36 .l:.xa1 'ifxa1 =, while 34 .l:.e 1 a2 35 d5 'ifd2 36 c4 'ifxe 1 37 i.xe 1 a1 'if 38 i.f2 somewhat favours White. The text­ move is much more convincing. 34 a2 35 d6 al'if 36 l:.e8 .l:.xe8 37 'ifxe8+ �h7 38 'ife4+ and 39 d7 wins. •••

If there is a fatal problem with the 1 8 . . . f5 move-order, this is it. This line, known as the 'Internet Refutation' , was brought into the lime­ light by the American Daniel Quigley back in 1 998. Neither Nunn's 1 990 work nor Anand's 1 993 monograph had reason to consider 1 9 axb5, which at first looks like a blunder. How­ ever, in NCO (2nd impression), Nunn considers it to lead to a clear advantage for White. Oddly, Lalit's 2003 work does not even mention it, even though it is extremely dangerous and may well constitute a refutation of Black's move-order. In any case, before Black ventures 1 8 . . .f5 he had better at least be aware of this possibility. Why was this move ignored for so long? It had always been assumed that the threat of 1 9 . . . f4 with an attack against the white king was too strong to allow White to spend time capturing a pawn on the queenside. Nowadays it would be easy to see with the help of a strong engine that this idea deserves very serious at­ tention, but back when the move was discov­ ered it was quite a find. 19 f4 The alternative is 1 9 . . . axb5, but this allows White simply to carry out his threat that began with 1 8 a4 by playing 20 i.xd5 cxd5 2 1 'ii'xb5 (D). White is two pawns up and d5 is hanging. The only chance is 2 1 . . .f4 but this is inadequate after 22 i.xf4: a) 22 . . . .l:.xf4 just loses: 23 .l:.xe6 (23 gxf4 may also win) 23 . . .l:hf2 24 .l:.e8+ (simplifying •••

UNDERSTANDING THE MARSHALL A TTACK

38

with 24 ':a8+ �fS 25 l:txf8+ ':xfS 26 ':e8 is another way to win) 24 .. .'�f7 25 �xd5+ 'ii? xe8 26 %:ta8+ 'ii?e7 27 :a7+ �d7 28 'ii? xf2 �xh2+ 29 'ii?e 3 �xg3+ 30 li:)f3 +- . Black is too much material down. b) 22 . . . �xf4 23 ':xe6 �xe6 (after 23 . . . .i.xd2 24 l:te8 ! White has a winning attack with his major pieces) 24 gxf4 �g4+ 25 'ii?h l �xf4 26 �e2 ±. White must still take some care, but ul­ timately the two extra pawns should win. These lines provide a good illustration of White's counterplay being stronger than Black's attack, even though Black is playing for mate. 20 �xf4! (D) White is willing to sacrifice a piece in order to break through on the queenside. Not only is bxa6 a possibility, giving White a strong passed pawn, but when White plays bxc6, the defence of the d5-knight will disappear. It is also impor­ tant that White maintains pressure on the a2-g8 diagonal - if Black moves the d5-knight or the e6-rook then the other piece will remain pinned. This move is not only strong; it is forced be­ cause other moves are just bad: a) 20 li:)f1 ?? .i.f3 mates. b) 20 �f1 ? fxe3 21 ':xe3 axb5 =t . c) 20 �xd5 ? cxd5 21 �xf4 ':xe l + 22 l:txe l .i.xf4 =t . d) 20 bxc6? fxg3 21 fxg3 (2 1 hxg3 �xg3 ! leads to mate) 2 l . . .�xg3 22 %:te2 �f5 ! wins for Black. After the text-move Black has several moves to choose from, but only one deserves serious consideration. 20 �xf4 This is the best try because others are refuted easily; for example: •••

a) 20 . . . l:txf4? 2 1 �f1 ! . A typical defensive idea in this line: White shores up the kingside and offers an exchange of queens while all of Black's pieces are hanging in the air. 2 1 . . .�h5 22 bxc6 :ff6 23 ':a5 ! is overwhelming. b) 20 . . . li:)xf4 is refuted in the same manner: 2 1 �f1 ! �h5 22 gxf4 �xf4 23 �g2 �xd2 24 :e5 ! (Quigley) forces Black into a lost ending after 24 .. :it'g6 25 l:txe6 �xe6 26 �xg6 hxg6 27 �xe6+. c) 20 . . . l:th6 (D) is very creative, but does not quite work:

w

c l ) 2 1 �xh6? ! l:txf2 22 l:te8+ c;i?f7 23 �xd5+ cxd5 24 l:te7+ ! 'ittfS (both 24 . . . 'ittxe7? 25 'ii'e 3+ and 24 . . . �xe7? 25 'itt xf2 'ii'x h2+ 26 'itte l lose quickly) 25 �xg7+ 'iitg 8 26 'iitxf2 �xh2+ 27 'itte 3 (27 �e l .i.xg3+ 28 'it'xg3 'it'xg3+ 29 'ii? f 1 �h3+ drawing - Harding) and Harding myste­ riously stops here, but after 27 . . . �xg3+ 28 li:)f3 'ii'x f3+ 29 'iitd2 'it'f2+ 30 ':e2 �f4+ 3 1 'ii? d l �xg7 3 2 bxa6 �f1 + 33 'ii?c 2 �xal 3 4 l:tg2 h5 35 �f3 �g5 36 l:txg4 a draw is likely.

OLD MAIN LINE:

c2) 21 �fl ! is the strongest move. 2 1 . . .l:lxf4 (2 1 . . . .Jtxf4 22 bxc6 +-) 22 bxc6 �f3 23 'ii'xf3 ':xf3 24 .Jtxd5+ ':f7 25 l:le8+ �f8 26 c7 l:lc6 27 .Jtxc6 ':xc7 28 �d5+ ':f7 leaves White with many ways to win, the simple 29 �e3 being one possibility. 21 ':xe6 �xe6 (D)

18. . .f5118

. . .

bxa4

39

24 a7 �f7 25 'it'e2 is the same. 24 .Jtf7 25 a7 'it'd7 26 'ii'a6 White has four pawns for a piece and the passed a-pawn will keep Black all tied up. Although there are some complicated tacti­ cal lines, clearly Black is struggling against 1 9 axb5, while h e has also failed to come up with a clearly satisfactory reply to 1 9 'ii'f l 'ii' h5 20 f4 bxa4 2 1 �xd5 . •••

Section 2 . 2 : The New Road 1 8 . . . bxa4 1 e4 e5 2 �f3 �c6 3 .Jtb5 a6 4 .Jta4 �f6 5 0-0 �e7 6 l:lel b5 7 .Jtb3 0-0 8 c3 d5 9 exd5 �xd5 10 �xe5 �xe5 11 ':xe5 c6 12 d4 �d6 13 llel 'ii' h4 14 g3 'it'h3 15 .Jte3 �g4 16 'ii'd3 l:lae8 17 �d2 ':e6 18 a4 bxa4 (D)

22 bxa6! The passed a-pawn will keep Black tied up. 22 bxc6 �xd2 23 'ii'xd2 �f4 24 .Jtxe6+ 'ii'xe6 25 f3 is also possible, but it allows Black better counterchances than the text-move. 22 �xd2 23 'ii'xd2 White can also play 23 a7 �c7 24 .Jtxe6+ 'ii'xe6 25 'ii'xd2 l:la8, but I think the text-move is more flexible. 23 ':a8 (D) 23 . . . �c7 24 'it'c2 ':a8 25 a7 'ii'h 6 26 �xe6+ 'it'xe6 27 c4 'it'd6 28 'it'e4 ± De Oliveira-Maffei, COIT. 1 999. •••

•••

24 'it'e2

This move is very strange-looking at first. Black destroys his pawn-structure and leaves his queenside to its fate. Also, White now has access to the c4-square for his pieces as well as the pos­ sibility of playing c4. However, there are some tactical points to this move - whichever way White recaptures, one of his pieces will be de­ flected. Taking with the bishop gives up pressure on the a2-g8 diagonal, while recapturing with the rook leaves White's back rank more vulnera­ ble. Another point to this move is that sometimes Black may switch over to the queenside and use his king's rook along the b-file. 19 ':xa4 1 9 �xa4 is also possible, but Black is OK after 1 9 . . . .Jtf4 ! (exploiting the pin along the

40

UNDERSTANDING THE MARSHALL A TTACK

third rank) 20 .txf4 lLIxf4 2 1 'ii'f l lLIe2+ 22 l:he2 l:he2 23 lLIc4 (23 'it'xh3 .txh3 24 lLIc4 l:tfe8 25 lLIe3 l:txb2 26 .txc6 l:.eb8 27 .te4 a5 is a little better for Black) 23 . . . l:.e6 (after 23 . . . 'iVh5 24 .td l l:te6 both 25 .txg4 'it'xg4 26 lLIe3 'ii'g 6 27 l:txa6 l:tfe8 { Tosic-Hazai, Vrnj acka B anj a 1 984 } and 25 l:ta5 'iVh3 26 'it'xh3 .txh3 27 lLIe3 %:ta8 { Cao S ang-Hazai, Hungary 1 998 } are equal) 24 lLIe5 'it'xfl + 25 ':'xfl .th3 and now 26 l:td l ! (26 l:ta l ? c5 ! gives B lack the initiative) rules out 26 . . . c5 due to 27 .td7 ':'xe5 28 dxe5 , but 26 . . . l:tc8 should be tenable for B lack. 19 f5 20 'it'n Not 20 f4? .txf4 ! 2 1 gxf4 (2 1 .txd5? cxd5 22 gxf4 10ses to 22 . . . l:th6 ! , one point being 23 l:.e2 l:.g6 ! 24 l:tg2 'it'xg2+ 25 �xg2 .te2+) 2 l . . .l:.g6 22 .i.xd5+ cxd5 (also good is 22 . . . �h8 ! ? 23 �h 1 l:.h6 24 l:.e2 cxd5 25 l:tg2 l:tg6 26 l:tf2 l:te8 27 l:tal l:.ge6 =+=) 23 .:r.xa6 .i.e2+ 24 l:txg6 .txd3 25 ':'g3 'ifh6 =+=. 20 'it'h5 (D)

21. f4 Heading into a tactical firestorm. If Black wants to avoid the mess that follows, he can play 2 l . . .lLIxe3 22 fxe3 (not 22 %:txe3? f4 23 l:.xe6 fxg3 ! 24 hxg3 .i.xg3 -+) 22 . . . c5 with some compensation. Objectively, Black should not fear the main line though. 22 cxd5 (D) ••

•••

•••

22 l:txe3 Black should avoid 22 . . . fxe3 ? 23 dxe6 l:.xf2 24 e7+ (this is the human continuation, but 24 h3 ! is even stronger according to the machines) 24 . . . �h8 25 e8'it'+ 'it'xe8 26 l:txe3 l:txfl + 27 lLIxfl , when White has too much for the queen. However, Black does have an alternative in 22 . . . fxg3 ! ? Then 23 hxg3 .txg3 24 'it'g2 ':'xe3 ! 25 %:txe3 (25 dxc6+ %:txb3 26 fxg3 l:txb2 is good for Black) 25 . . . .txf2+ 26 'it'xf2 ':'xf2 27 dxc6+ �h8 28 xf2 'ii'h 2+ 29 �e l 'it'h4+ 30 �fl 'iVh 1 + is a draw. This line alone is enough to show the theoretical soundness of Black's posi­ tion after 2 1 . . .f4. 23 fxe3 fxg3 24 dxc6+ �h8 25 hxg3 l:txn + 26 l:txfi 'iVg5 27 l:tf4 Worse is 27 lLIc4 .te2 ! 28 �f2 .txfl 29 lLIxd6 'it'f6+ 30 �g 1 'iVxd6 3 1 �xfl 'ii'x g3 and Black is much better. 27 .txf4 28 exf4 'it'e7 29 lLIn (D) White is doing well from a material stand­ point and his passed pawns look dangerous, but Black has a resource. 29 'ii'e4! This nice move throws White off balance. 30 d5 'ii'd3 31 l:.b4 g5! 32 c7 �g7 This was Jo.Fernandez-Claridge, corr. 1 998. White's uncoordinated pieces leave him unable •••

Now 2 1 f4 leads to the Main Line (Section 2.3), but Black has managed to avoid the dan­ gerous 1 9 axb5 from Section 2. 1 , Line B. White has two other moves to consider: 40 A: 21 c4 41 B: 21 l:txa6

A) 21 c4 This had always been considered the critical test of 1 8 . . . bxa4, but the latest evidence sug­ gests that this move is actually more dangerous for White than it is for Black.

•••

•••

Ow MAIN LINE:

1 8. . .f5118. . . bxa4

41

Memorial, Riga 1 995. Here 28 . . . l:Ixf4 would induce White to take a draw with 29 ':'c8+ rt;f7 30 l:c7+ �f8. However, 26 gxf4 :xf4 27 ':'e8+ rj;f7 28 iLxd5+ 'ii'xd5 29 l:cc8, as given in Informator, looks stronger. 23 'ifxe1 iLxf4 24 l:xc6 (D) 24 gxf4 ':'e8 25 'iVaI iLh3 26 iLxd5+ cxd5 27 f3 h6 + is given by Pavlovic.

B

to hold everything together, despite the passed pawns.

B) 21 l:xa6 (D) White starts chopping. This is much less risky for White than 2 1 c4, and it is also more of a challenge to Black's move-order. 24 l:e8 24 . . . iLxd2 25 'iVxd2 l:e8 26 l:c5 (the pin looks annoying, but White has difficulty capi­ talizing on it) 26 . . . iLf3 27 iLc4 ? ! (after 27 iLxd5+ iLxd5 28 c4 'ii' h 6 ! 29 'ii'c 3 iLc6 30 d5 iLd7 3 1 lta5 'iWg6 32 ltal 'ii'e4 Black is OK Pavlovic; 27 'it'c l is also possible) and now: a) 27 . . . rt;h8 and now Pavlovic focuses on 28 b4?, but 28 lLxd5 iLxd5 29 c4 ! 'iVh6 ! 30 'iVd l looks good for White. b) 27 . . . 'iVf5 28 'iVc 1 (Pavlovic again offers a rather ' cooperative' line, 28 iLfl ? 'iVbl ) 28 . . .iLe4 29 'iVd l gives White four pawns for the piece, though Black still has counterplay. c) After 27 . . . 'ii'h6 ! 28 'iVxh6 (28 iLxd5+? rj;f8 ! -+) 28 ... l:Ie l + 29 iLfl gxh6, only White is in danger. 30 c4 �f6 ! ? (30 . . . iLe2 3 1 cxd5 leads to an ending where White's pawns aren' t s o great a threat) 3 1 l:e5 �e4 32 l:e8+ �f7 33 l:Ie5 sees White clinging on by his fingertips. 25 iLxd5+ rt;rs 26 iLe6 Worse are 26 'iVxe8+? 'iVxe8 27 gxf4 'iVe2 + 1.Fernandez-Bennedik, COIT. 1 998 and 26 'iVaI 'iVxd5 27 l:Ic5 'iVe6 28 gxf4? iLh3 29 �fl 'iVg4+ 30 �g3 'it'f3 -+. 26 lLxd2 27 'iVxd2 lLxe6 (D) Black retains counterchances, but it is not easy to get the bishop into an attacking position. •••

-

21 f4 With his queenside crumbling, Black must push forward with his kingside attack. 22 iLxf4 As in the 1 8 . . . f5 ? ! 1 9 axb5 ! ? line, White is willing to shed a piece to end Black's attacking chances. Not 22 1bc6? fxe3 23 1be3 �xe3 24 iLxe6+ iLxe6 25 'it'e l iLd5 26 l:Ixd6 as 26 . . . 'ii'h 3 ! mates. 22 l:Ixe1 22 . . . iLh3 23 ':'xe6 lLxfl 24 �xfl lLxf4 25 l:Iaxc6 'iVf3 and now 26 lLxd5 ? ! 'iVxd5 27 gxf4 'iVf3 28 d5 was played in Ivanchuk-Short, Tal •••

•••

•••

42

UNDERSTANDING THE MARSHAll A rrACK

on the e-file. Thus there are three main continua­ tions: 42 A : 21 gS? ! 44 B : 21. l:b8?! C : 21 l:fe8 46 •••

••

•••

In heading for the Main Line, Black has to decide which deviations are more dangerous to deal with - those after 1 8 . . . f5 or those after 1 8 . . . bxa4. Although White has some compli­ cated tries after 1 8 . . . bxa4, both of White's devi­ ations after 1 8 . . .f5 are clearly more dangerous.

Section 2 . 3 : The M a i n Line

The only real alternative to these three moves is 2 1 . . . �h8 ? ! , but then 22 .i.xd5 ! (better than 22 lLlc4 ? ! .i.xf4 ! 23 .i.xf4 lLlxf4 24 l:xe6 lLlh3 + ! 25 �g2 lLlg5 ! 2 6 'iif2 .i.f3+ 2 7 � f l f4 ! , when Black is pressing) 22 . . . cxd5 23 l:xa6 leaves Black with nothing better than 23 . . . .:fe8, trans­ posing to the note to Black's 22nd move in Line C 1 , because 23 . . . 'ii'e 8 24 ..tf2 ! ..te2 25 l:xd6 is winning for White (Nunn) and 23 . . . g5 24 fxg5 l:tfe8 25 l:txd6 ':xd6 26 .i.f4 ± is clearly not satisfactory either. How�ver, Black is struggling there too, so 2 1 . . .�h8 should probably be discarded.

A) 21 gS? ! (D) •••

1 e 4 e S 2 lLlf3 lLlc6 3 .i.bS a6 4 .i.a4 lLlf6 5 0-0 .i.e7 6 ':el bS 7 .i.b3 0-0 8 c3 dS 9 exdS lLlxdS 10 lLlxeS lLlxeS 11 ':xeS c6 12 d4 .i.d6 13 l:el 'iih4 14 g3 'iih3 1S .i.e3 .i.g4 16 'ii'd3 llae8 17 lLld2 ':e6 18 a4 bxa4 (or 18 ... f5 ? ! 19 'iif l 'ii'h5 20 f4 bxa4 2 1 ':xa4) 19 ':xa4 fS 20 'iin 'iihS 21 f4 (D)

This is the starting position for the Main Line. Black has to decide how to fight for the initia­ tive. He can try to rip open the kingside, utilize the freshly opened b-file, or simply double rooks

This used to be the main line and it has prob­ ably been refuted, but it is tricky, with some lines still open to dispute. Because this is the most brutal and direct way to play, 2 1 . . .g5 is certainly critical. Black will try to attack at all costs. 22 l:xa6 gxf4 Essentially the position is a race. Black's at­ tack may look strong but his own position is rather rickety. The play will be very tactical from this point forward. White has : 43 A I : 23 ':xc6 43 A2: 23 .i.xf4!

OLD MAIN LINE:

18. . .f5118. .

.

bxa4

43

AI) 23 l::txc6 (D)

Black seems to hold his own after this. 23 fxe3 This looks simplest, but other moves are possible: a) 23 . . . l::txe3 ? is bad due to 24 .i.xd5+ 'iii>h 8 25 l:txd6 fxg3 26 h3 ! .i.xh3 27 l::txe3 .i.xfl 28 tt'lxfl f4 29 ':'e7 ! and White's attack is faster, Matsukevich-Kucherov, COIT. 1 983. b) 23 ... .:.h6 is risky. 24 .i.xd5+ �h8 25 'ii'g 2 fxe3 26 tt'lc4 f4 27 ':'xd6 ':'xd6 28 tt'lxd6 f3 29 .i.xf3 ':'xf3 30 h4 ! (better than 30 ':'fl 'ii'd 5 ! or 30 b4 .i.h3 3 1 'ii'e 2 { 3 1 Va2? ':'f8 ! -+ } 3 1 . . .l::t f l + 32 'it'xfl .i.xfl 33 l::t x fl e2 34 l::te l 'it'f3 +) 30 . . . .i.e6 3 1 g4 ! (3 1 l::t f l ? ! e2 ! 32 l::te l .i.d5 33 :xe2 'it'g6 34 'iii> h 2 l::te 3 ! with counterplay) 3 1 . . . .i.xg4 (3 1 . . . 'it'xh4 32 'it'xf3 'it'xe l + 33 'iii> g 2 'iii> g 8 34 tt'le4 e2 35 tt'lg3 .i.c4 36 'it'a8+ +-) 32 tt'lc4 is good for White. c) 23 . . . fxg3 also looks fine for Black, but it is complicated: c l ) 24 hxg3 f4 25 ':'xd6 fxg3 26 "ilig2 ':'xd6 27 'iIfh I ? (better is 27 .i.xd5+ ':'xd5 28 'ii'x g3 with unclear play) 27 . . . .i.h3 ! 28 "ilie4 was soon drawn by repetition in Elburg-K.D.Miiller, COIT. 1 984. However, after 28 . . . .i.e6 29 "ilih l Black does not have to repeat because 29 . . . 'it'xh l + 30 'iii>xh 1 tt'lxe3 is clearly better for him. c2) 24 'it'g2 f4 (also good enough is 24 . . . 'iii>h 8 25 .i.xd5 .i.h3 ! { not 25 . . . l::tg 6? 26 l::txd6 l::txd6 27 hxg3 +- Nunn } 26 'it'e2 .i.g4 27 'it'g2 { White must repeat because 27 .i.f3 ? l::txe3 -+ and 27 tt'lf3 ? f4 -+ cannot be considered } 27 . . . .i.h3 with a draw) 25 .i.xd5 f3 26 .i.xe6+ �h8 (D). •••

w

What a position ! White is a rook and a piece up but his kingside has been cut to ribbons. Of course the position is probably just equal . . . 27 h3 ! ! (an amazing move) 27 . . . f2+ ! 28 .i.xf2 gxf2+ 29 'it'xf2 .i.xe6 ! 30 'it'xf8+ .i.xf8 3 1 l::texe6 'it'g5+ 3 2 'iii>h l 'it'xd2 leads to a likely draw. 24 .i.xd5 'iii>h 8! 25 .i.xe6 exd2 26 l::ta l .i.e2 26 . . . .i.xg3 27 hxg3 .i.f3 is not so simple after 28 'it'b5 ! , giving the king an escape-route and threatening 'ii'e5+. 27 'ii'f2 .i.b5 28 l:.xd6 dl'it'+ 29 l:.xdl 'ii'xdl + 30 �g2 .i.d3 31 .i.d5 ':'e8 32 .i.f3 (D)

B

32 .i.e2 32 . . . .i.e4 also draws easily. 33 .i.c6 .i.d3! This is simplest. White has nothing better than repeating the position with 34 .i.f3 . •••

A2) 23 .i.xf4! (D)

UNDERSTANDING THE MARSHALL A TTACK

44

B) 21. lXb8?! (D) ••

This should be winning for White, but it will require some care. 23 .ixf4 24 llxe6! .ixd2 25 .ixd5 After 25 lle5 'itth 8 (not 25 . . J:lb8 ? 26 l:hc6 ! lXxb3 27 'ikc4 and White wins, Karklins-Chow, Midwest Masters, Chicago 1 982) 26 'ikf2 .ig5 (26 . . . 'ii'g 5? loses to 27 .ixd5 cxd5 28 h4 ! ) 27 lXxc6 lLlf6 White is a lot of material up, but Black still has some chances on the kingside. 25 cxd5 26 'ii'g2 f4 27 'ii'xd2 .ixe6 27 . . . fxg3? 28 llf6 ! is just winning for White. 28 llxe6 fxgJ 29 hxg3 (D) •••

•••

B

29 ll0+ 29 . . . 'it'g4(?) 30 l:Ie3 :a8 3 1 'it'e2 +- was given by Nunn. 30 'ittxn 'it'hl + 31 'itte2 'ifh5+ Strangely enough, White can only avoid per­ petual check by shedding both his g-pawn and his rook. Still, after 32 g4 'it'xg4+ 33 'ittf2 'ifxe6 34 'ifg5+ 'ittf7 35 'ii'f4+ 'ittg 6 36 'ii'g 3+ 'ittf7 37 b3 White soon obtains two connected passed pawns, which should be enough to win (Nunn). •••

For a long time, this move took over as the main line. It looks very logical because Black threatens 22 . . . llxb3 (destroying the piece that is pinning everything on the a2-g8 diagonal) followed by capturing on e3. 22 .ixd5 This is by far the most common move, but there are several others : a) 22 'ii'xa6 llxb3 23 lLlxb3 and at this point 23 . . . llxe3(?) was given as winning by Nunn, but 24 llxe3 lLlxe3 25 'ii'c 8+ .if8 26 l:la8 'iff7 27 lLlc5 leaves Black defenceless. However, 23 . . . 'ife8 ! wins back the material. After 24 'ii'd 3 lLlxe3 25 1L1c5 .ixc5 26 dxc5 h6 27 lXa7 1L1d5 28 lXxe6 'ikxe6 Black stands well. b) 22 'ii'f2 lXxb3 23 lLlxb3 .i.d l ! is another one of Black's tricks. After 24 .i.d2 (not 24 l:Ixa6 .i.xb3 25 lXxc6 ? ! lLlxe3 26 lXxe3 'ii'd l + 2 7 'ii'f l 'ii'x fl + 2 8 �xfl lXxe3 2 9 lXxd6 .i.c4+ +) 24 .. Jhe l + 25 'ii'xe l .i.xb3 26 l:txa6 lLle7 matters are not so clear, but Black should have good chances with White's king position so loose. c) 22 lha6 l:lxb3 23 lLlxb3 lLlxe3 24 l:Ia8+ �f7 25 'ii'f2 (25 'ii'd3 .i.xf4 has the point that 26 gxf4? fails to 26 . . . .if3 -+) and now 25 . . . .ih3 ? ! 2 6 'ii'e2 (even better i s 2 6 llaal ! , when Black is in trouble) 26 . . .'ifxe2 27 llxe2 lLlc4 is given as + by Nunn, but after 28 llxe6 'itt xe6 29 lLla5 lLlxa5 30 lIxa5 White is certainly not worse. Fortunately for Black, he has the odd-looking 25 . . . 1L1d l ! 26 'ii'd 2 (26 'ii'f l lLle3) 26 . . . .i.xf4 ! , when White should deliver perpetual check with

Ow MAIN LINE:

27 l:ta7+ 'iii'g 8 28 l:Ia8+ because 27 gxf4? loses to 27 . . . .i.e2 ! . d ) 2 2 R.f2 ! ? R.e2 ! (not 2 2. . . l:Ixe 1 ? 2 3 'ii'xe 1 l:Ie8 24 'ii'f l l:Ie2 25 l:Ixa6 .i.h3 26 'ii'd 1 .i.g4 27 'ii'f l R.h3 28 l:Ixc6 ! .i.xfl 29 R.xd5+

f8 30 1i.f3 'ii'e 8 3 1 lLlxfl ± Rohde-Beliavsky, Ali­ cante 1 978) and now: d 1 ) 23 'ii'g 2? and now 23 . . . ltxb3 ! 24 lLlxb3 1i.f3 25 'ii'f l lLlxf4 ! (instead 25 . . . R.e2 repeats) is an idea of Shamkovich's that modem analy­ sis shows to be winning for Black: 26 gxf4 1i.xf4 27 1i.g3 1i.xg3 28 hxg3 f4 ! 29 ltxe6 'ii'h 1 + 30 f2 fxg3+ and Black wins. d2) 23 .i.xd5 cxd5 24 'ii'g 2 is a better try for White: 24 . . . 'ii'f7 25 lLlf3 (25 'ii'x d5 ? ! .i.xf4 26 lLlf3 l:txb2 27 ltb4 %hb4 28 cxb4 lte4 29 'ii'xf7+ 'iii' xf7 30 l:Ixe2 l:txe2 3 1 gxf4 = Muratovic­ Vitomskis, COTTo 1 985) 25 . . . 'ii'e 8 26 lLle5 ! (26 l:Iaa 1 ? ! 'ii'b 5 27 lLle5 1i.xe5 28 dxe5 1i.d3 29 b4 1i.e4 30 'ii'f l 1i.d3 112- 112 Usachy-Gabrans, COTTo 1 975) 26 . . . 'ii'xa4 27 'ii'xd5 l:Ie8 28 l:Ixe2 gives White more than enough for the exchange. 22 cxdS (D) •••

Black has enticed White to give up his light­ squared bishop, but his position will prove to be rather loose. White has a complicated, yet pleas­ ant choice: B l : 23 'ii'g2 45 46 B2: 23 l:Ixa6

B1) 23 'ii'g2 White activates his queen and targets the d5-pawn. 23 'ii'e8 •••

18. . .f5118.

. .

bxa4

45

Black cannot defend the pawn, so he tries to create pressure on the e-file, which has been weakened by two of White's major pieces leav­ ing his first rank. 24 'ji'xdS

h8 (D)

w

2S �f2! This move holds everything together and leaves Black struggling. Instead 25 lLlc4 ? ! runs into 25 . . . 1i.xf4 ! . After 26 gxf4 l:Ig6 27 lLld6 (27 'ili>h 1 ? loses to 27 . . . 'ii'd7 28 'ii'g 2 1i.h3) Black can head for a draw or play for a win with some risk: a) 27 . . . 1i.f3+ leaves White with little choice but to allow a draw with 28 �fl 1i.xd5 29 lLlxe8 ':xb2 30 lte2 ltb 1 + 3 1 lte 1 ltb2 32 lte2 l:Ib l + 3 3 %te l 112- 112 Kamsky-J.Polgar, Linares 1 994. b) 27 . . . 1i.h3+ 28 �f2 l:Ixb2+ 29 l:Ie2 l:Ig2+ 30 'ii'x g2 l:Ixe2+ 3 1

xe2 'ii' h 5+ 32 'ii'f3 .i.g4 33 'ii'xg4 fxg4 (Black needs this pawn to create counterplay, so 33 . . . 'ii'x g4+? 34 'iii'd2 +- is out of the question) 34 �d3 'ii'xh2 35 l:Ixa6 'ii'h 1 36 ltb6 h6 was unclear in Canfell-Blatny, Alten­ steig 1 988. 2S gS Also difficult is 25 . . . ltxb2 26 lta2 ltxa2 27 'ii'xa2 g5 28 fxg5 ! f4 29 gxf4 1i.xf4 30 1i.xf4 l:Ixe 1 3 1 'ii'xa6 ± City of London-City of Ruurlo, cOTT. 1 987. 26 l:Ixa6 l:Ixb2 27 'ii'a 8! 27 l:Ia2 ? ! gxf4 28 gxf4 l:Ixa2 29 'ii'xa2 .i.xf4 has been shown to be sufficient for Black. 27 l:Ib8 28 'ii'c6 'ii'hS 29 h3! This little move comes up time and time again as an unusual tactical resource. White gains time to throw Black off balance. •••

•••

46

UNDERSTANDING THE MARSHALL A TTA CK

29 .'ili'xh3 30 'ii'g2 gxf4 31 'ii'xh3 iLxh3 32 gxf4 (D) ••

This moves refutes Black's play, so we shall not bother looking at the alternatives 24 'ili'f2 and 24 'ili'g2.

B

± Popolitov-Trushakov,

COIT.

1 980- 1 . 24 'ii'f7 Everything is bad: a) 24 .. .l:he3 25 ':xe3 ':xe3 26 ':xd6 ':e l + 2 7 lLl n + - Grigoriev-Guzek, COIT. 1 990. b) 24 . . . iLf3 25 ':xd6 ':xe3 26 ':xe3 ':xe3 27 1L1xf3 'ili'xf3 28 'ii'x d5+ +-. c) 24 . . . �h8 25 'ili'xd5 ':'h6 26 'ili'g2 iLh3 27 'ii'f2 +-. 25 h3! Again this clever move. White gains time to direct his knight to the excellent e5-square. White must avoid the greedy 25 'ili'xd5 ? ':xe3 26 'ii'xf7+ �xf7 27 l:txe3 ltxe3 28 ltxd6 be­ cause 28 . . . ':e l + 29 lLln iLh3 suddenly turns the tables. 25 iLxh3 25 . . . iLh5 26 'ii'xd5 iLxf4 27 ':xe6 ':xe6 28 lLln +- Short-Pinter, European Clubs Cup, Rot­ terdam 1 988. 26 1L1fJ h6 26 . . . ':'xe3 27 ':'xe3 ':xe3 28 ':xd6 h6 29 �f2 +-. 27 lLle5 iLxeS 28 ':'xe6 'ili'xe6 29 dxe5 ± Or even 29 iLf2 ! ? ±. It looks like 2 l . . .':'b8 will not hold up, as both B l and B2 cause major problems for Black. •••

82) 23 ':xa6 (D)

•••

23 ':be8 Other moves are possible but Black still has problems: a) 23 . . .l:txb2 24 'ii'g 2 ! ':b5 25 c4 dxc4 26 'ii'c 6 'ii'e 8 27 'ii'xe8+ ! (even stronger than 27 lLlxc4) 27 . . .l:txe8 28 ':xd6 c3 29 lLlc4 c2 30 lLle5 +- P.Coleman-W.Schiller, COIT. 1 996. b) 23 . . . 'ii'e 8 24 iLf2 'ii'd7 25 ':xe6 'ii'xe6 and after 26 c4 dxc4 27 'ili'xc4 'ili'xc4 28 lLlxc4 iLe2 29 ':'c6 iLxc4 30 ':'xc4 ':xb2 White was unable to win in Chandler-P.Littlewood, Lloyds Bank Masters, London 1 987. 26 b4 ! ? is one possible improvement for White. 24 'ili'b5! (D) •••

C) 21 ':fe8 (D) Black simply doubles rooks on the e-file. This is considered Black's best chance to keep the Main Line afloat. White can decimate Black's •••

OLD MAIN LINE:

J8. . J5/J 8. . . bxa4

47

b3) 25 h3 ! ? looks promising for White after 25 . . . .ixh3 26 �f2. This has scored very well for White in correspondence play. The main point is that 26 . . . .ig4 27 :h l ! wins for White. 23 .l:.xe3 ':'xe3 24 ':'xc6 (D)

queenside b y sacrificing his e3-bishop or he can defend it. Note that 22 iLt2? fails to 22 . . . ':'xe l 23 .ixe l ':'e2 -+. We examine: 47 C l : 22 .l:.xa6 C2: 22 'ii'f2 48

el) 22 ':'xa6 This should probably lead to a draw. 22 .:.xe3 This is Black's main continuation and looks good enough. Instead Black can play 22 . . . �h8? ! 23 iLxd5 cxd5 (as mentioned before, this line can also arise from the move-order 2 1 . . .'it>h8 ? ! 2 2 .ixd5 cxd5 2 3 ':'xa6, but i t all looks very sketchy for Black), when White has: a) 24 'ii'f2 g5 25 ':'aal ! (25 ':'xd6 ! ? ':'xd6 26 fxg5 ':'de6 27 h4 is also possible) 25 . . . gxf4 26 .ixf4 ':'e2 27 ':'xe2 ':'xe2 28 'ii'xe2 ! .ixe2 29 .ixd6 .id3 30 ':'e l .ie4 3 1 b3 ;j; He.Mtiller­ Huerga, corr. 200 1 . b ) 24 'ifb5 ':'h6 (24. . .g5 25 'ii'x d5 gxf4 26 ':'xd6 ':'xe3 27 ':'xe3 fxe3 28 ':'d8 exd2 29 'fi'e5+ �g8 30 ':'xe8+ 'ii'x e8 3 1 'ii'x e8+ �g7 32 'fi'a4 dl 'fi'+ 33 'fi'xd l .ixd l is given by Nunn and is probably a draw) when White has: b I ) 25 h4 iLxf4 26 ':'xh6 (26 gxf4 ltxa6 27 'ii'x a6 'ii'xh4 can only be better for Black) 26 . . . .ixh6 27 lDfl f4 28 .id2 .ie2 leads to a mess . b2) 2 5 lDfl .if3 2 6 .i c 1 ':' f8 2 7 ':'e3 .ie4 28 ':'c6 (not 28 'ii'e 2? 'ii'xe2 29 ':'xe2 .id3 30 ':'xd6 ':'xd6 3 1 ':'e3 .ie4 =+= Unzicker-Wahls, Bundesliga 1 989190) 28 ... ':'g6 29 'ii'a6 'ii'h 6 30 ':'e2 looked a little sketchy for Black in Un­ zicker-Nunn, Bundesliga 1 990/ 1 . •••

This is the point of White's play. With one move White attacks both the d6-bishop and the d5-knight. 24 'ii'e8 Black cannot hold on to his extra piece and he is three pawns down so he must counterat­ tack. 25 .i.xd5+ (D) 25 ':'xd6 is an attempt to keep some play in the game. After 25 . . . .:.e l 26 l:hd5 �f8 27 ':'e5 ':'xfl + 28 lDxfl 'fi'b8 29 lDd2 'fi'a8 Black has counterplay. Pavlovic continues 30 lDc4 .ih3 3 1 lDe3 'ii'a l + 32 �f2 'ii'xb2+ 33 .ic2 g6 34 ':'c5 iLg4 35 ':'c8+ �g7 36 ':'c7+ �f8, with a likely draw. •••

25 �8 26 lDe4 ...

UNDERSTANDING THE MARSHALL A TTA CK

48

26 l:txd6 lIel 27 lIe6 (or 27 .i.e6 lIxfl + 28 tDxfl 'ii'b 8 ! ) 27 . . . ltxfl + 28 tDxfl 'ii'b 5 29 .i.g2 'ii'x b2 gives Black good counterplay, as shown in several games. 26 fxe4 27 lIxd6 lId3! (D) •••

24 c4 ! , when White will get his knight to e5 or play c5 . Therefore Black must react violently. 22 g5! Once again a full-board war has erupted. White has the typical choice between blasting away the queenside and confronting Black on the kingside: C21 : 23 lIxa6 48 C22: 23 fxg5 49 •••

C21) 23 l1xa6 (D)

This position had been considered to be good for Black, despite the 3 ( ! )-pawn deficit. Black threatens both 28 . . . l:td l and 28 . . . e3 . Neverthe­ less, White can draw with 28 l:ta6 (28 .i.b3 e3 is risky for White after 29 'ii'x d3 e2 30 l:te6 'ii'xe6 3 1 .txe6 e l 'ii'+ 32 �g2 'ii'xe6 or 29 lId8 ? ! 'ii'x d8 3 0 'ii'x d3 e2) 2 8 . . . 'ii'b 5 2 9 lta8+ rJite7 30 l:ta7+ �d6 3 1 c4 'ii'x b2 32 l:a6+.

C2) 22 'ii'f2 (D)

White mows down the queenside, confident that he can weather the storm on the other side of the board. Allowing the g-file to open up is risky, but White is counting on the manoeuvre tDfl -g3 to hold things together. 23 gxf4 24 gxf4 �h8 25 .i.xd5 cxd5 26 tDn (D) •••

B

White defends the e3-bishop and is ready to start picking apart the queenside. Now 22 .. .'�h8 can be met by 23 .i.xd5 (but not 23 tDc4? tDxf4 ! 24 gxf4 .i.f3 with a winning attack) 2 3 . . . cxd5

26 .th3! •••

Ow MAIN LINE:

Strangely enough, Black does not necessar­ ily want to bring a rook to the g-file. Indeed, 26 . . . ':g8 27 �g3 �f3 28 ':a8 ! ':eg6 29 ':xg8+ .l:.xg8 looks a little shaky for Black after 30 l:.al ! (30 :fl �e4 3 1 ':al 'it'h3 32 ':a6 �f8 33 :a8 h5 34 'it'e2 �h7 was He. Muller-Idler, COIT. 1 994 and here White should probably just re­ peat with 35 :a7+ �h8 36 :a8) 30 . . . 'it'g4 3 1 ':a6 (3 1 �fl h5 32 .l:.a6 l:.g6 3 3 �e l h4 gave Black counterplay in laloszynski-Biedunkie­ wicz, COIT. 1 998) 3 1 . . .j.xf4 32 j.xf4 'ii'xf4 33 ':al ! (intending ':f1 ) 33 ... 'ii'g4 34 'ii'e 3 �e4 35 'it'h6 ± Glaser-Morgado, COIT. 1 987. 27 �g3 "g4 With ideas like . . . h5 and . . . j.xf4 in the air, Black has good counterplay.

18. . .f5118 . . bxa4

49

.

in White' s court. White has a few other moves, of which only one is worth considering: a) 25 j.d l ? ':xe3 26 j.xh5 .l:.xe l + 27 �fl �xf4 ! -+ . b) 25 �fl ? j.xfl 26 �xfl j.xf4 ! 27 �xd5 cxd5 28 'it'xf4 ':f8 +. c) 25 .l:.xa6 has been seen a lot in COITe­ spondence play. After 25 . . . j.xf4 (D) White has :

C22) 23 fxg5 (D)

c 1 ) 26 .JtxdS 'ifxg5+ (26 . . . 1Wg4+ 27 �h l cxdS 28 .l:.xe6 ':xe6 29 �fl �xg5 30 �d2 �h4 3 1 �g3 �xg3 32 'ii'x g3 'it'xg3 33 bxg3 .l:.xe l + 3 4 j.xe l 1eaves Black two pawns down although the game was eventually drawn in Fleck-Weg­ ner, Bundesliga 1 989/90; however, there is no need to go into this, even if it is objectively drawn) 27 'it'g3 ! (there is nothing else, but Black is still doing well) 27 . . . cxdS 28 ':xe6 :Xe6 29 �xf4 :Xe l + 30 �f2 'iWxg3+ 3 1 �xg3 j.e6 (or 3 1 . . .j.f5). The endgame should be drawn, but if anyone is better it is Black. c2) The queen sacrifice 26 1Wxf4 ! is White's best try, but 26 ... �xf4 27 �xf4 'ii'g4+ 28 �g3 'iti>f8 (also possible is 28 ... 'ifxg5 29 �e4 'ii'f5 30 �f2 { or 30 j.e5 :f8 3 1 j.xe6+ 'ii'xe6 32 .l:.a7 'it'g6+ 33 �g3 :f7 34 .l:.a8+ :f8 35 :a7 1/2- 1/2 Loc-Potrata, COIT. 1 997 } 30 . . . 'it'b5 3 1 �xe6+ �xe6 32 ':a7 'ii'f5 33 �e4 :f8 34 �e5 'ii'g 6+ 35 �h l :fl + 36 ':xfl 1/2- 1/2 Sanner-Pansier, COIT. 2002) 29 �xe6 ':xe6 30 ':xe6 (30 ':a8+ �f7 3 1 :a7+ �g6 32 ':xe6+ 'iWxe6 33 �f2 'it'f5+ 34 �e l does not help White much after 34 . . . 'iWb5 ! ) 30 . . . 'iWxe6 3 1 �f2 'ii'f5+ 32 �e l �g4 gave Black good counterplay in Kloep­ ping-Schwetlick, COIT. 2002. We return to 25 'iti>h l (D): °

This is more critical. White is willing to go into a hand-to-hand fight in the kingside and centre. 23 f4 Black must go forward and tear up the white kingside. 24 gxf4 j.h3 Black is a few pawns down, but there are ideas such as . . . �xf4, . . ... g4, . . . :f8, and even . . . h6 to open more lines. White must be very careful. 25 �hl This is probably best. In 1 990 Nunn wrote "A good line on move 25 for White is hard to find" and 20 years later after a lot of back and forth, the ball is still (or should I say 'again' ?) •..

50

UNDERSTANDING THE MARSHAll A TTACK

B

either 28 . . . .i.xe5 29 dxe5 'ii'xe5 30 ':xa6 or 28 . . . :f8 29 'ii'g 3 (29 .i.c2 is also good) 29 . . . 'ii'f5 30 ':aa 1 ! . b ) Following 2 5 . . . .i.xf4. Pavlovic gives the forcing line 26 .i.xd5 cxd5 27 'ii'xf4 ':xe3 28 %:tg1 %:te2 29 g6 hxg6 30 l:txa6 .i.g2+ 3 1 ':xg2 %:txg2 32 �xg2 'ife2+ 33 �g3 'ii'xa6 34 liJf3 'ii'e6 35 liJe5 �g7 with approximate equality.

Conclusions

Now : a) After 25 . . . 'ii'g4? ! White has: a l ) 26 liJf3 ?? :xe3 -+ R.Junge-Wegner. Bundesliga 1 988/9. a2) 26 .i.xd5? cxd5 27 %:tg1 'ii'f5 28 l:tg3 .i.xf4 (Hasche-Stock. corr. 1 99 1 ) 29 .i.xf4 %:tel + 30 ':g l ':8e2 -+. a3) 26 l:tg 1 ? is met by 26 . . . 'iff5 ! with a strong initiative. For example. 27 %:txa6 %:txe3 28 ':xc6 ':e2 =+= or 27 .i.xd5 ? 'ii'xd5+ -+. a4) 26 ttJc4 ! (the knight heads for e5 and covers the e3-bishop at the same time) 26 . . . 'ii'xf4 (the only chance) 27 liJe5 ! 'ife4+ 28 'iti>gl . It looks like Black still has problems here after

Black has two ways to reach the Old Main Line. Currently 18 . . . f5? ! looks a bit suspicious be­ cause of 19 axb5 ! ? Perhaps Black can hold here somehow. but it will certainly be an unpleasant task. 1 8 . . . bxa4 is a safer move-order. because 2 1 c4 looks dangerous for White and 2 1 ':xa6 does not look fatal. Once we are in the Main Line. both players must walk a tightrope. Even though this line is supposed to favour White the­ oretically. we can see that matters are not clear at all. There are many pitfalls for both sides and the lines are well worth studying to learn the plethora of typical Marshall Attack ideas. The rare 2 1 . . .g5 could be worth a roll of the dice. In­ stead 2 1 . . Jlb8 ? ! is looking rather suspicious. but 2 1 . . ..:fe8 is holding up pretty well.

3 1 5 il.e3 : Other Li nes

1 e4 eS 2 ltJf3 ltJc6 3 iLbS a6 4 iLa4 ltJf6 S 0-0 iLe7 6 :leI bS 7 iLb3 0-0 8 c3 dS 9 exdS ltJxdS 10 ltJxeS ltJxeS 1 1 l:txeS c6 12 d4 iLd6 13 l:tel 'iWh4 14 g3 'iYh3 IS iLe3 iLg4 16 'iWd3 (D)

Section 3.2: The Pawn Push 16 l:lae8 17 ltJd2 fS Section 3.3: The Adams Variation 16 :ae8 17 ltJd2 'iWhS Section 3.4: White Avoids 18 a4 .••

•••

54 60 65

Section 3 . 1 : The Accelerated Pawn Push 1 6 . . . f5 1 e4 eS 2 ltJf3 ltJc6 3 iLbS a6 4 iLa4 ltJf6 S 0-0 iLe7 6 l:.el bS 7 iLb3 0-0 8 c3 dS 9 exdS ltJxdS 10 ltJxeS ltJxeS 11 l:lxeS c6 12 d4 iLd6 13 l:tel 'iWh4 14 g3 'iWh3 IS iLe3 iLg4 16 'iWd3 fS (D)

In this chapter we look at deviations for both sides in the 1 5 iLe3 line. The first two of these look at 'pawn push' variations where Black throws his f-pawn for­ ward without playing his rook to e6 first. The first of these lines, the Accelerated Pawn Push, even does without . . . l:.ae8, although if Black plays this move later it will often transpose to the Pawn Push of Section 3 . 2 . Both of these variations are quite risky for Black but could be very dangerous against an ill-prepared op­ ponent. There is a sea of transpositions and hopefully I have successfully navigated through them. The third line covered here is the Adams Variation. Essentially this is an 'Accelerated' Spas sky Variation, as Black has omitted the move . . . l:te6. This variation is quite popular, and has been played frequently by Marshall experts Leko, Grishchuk and Gustafsson. The fourth part of this chapter covers the main move 1 7 . . . l:te6 where White plays something other than the most common move, 1 8 a4. Section 3.1 : The Accelerated Pawn Push 16 fS 51 •••

w

This risky advance could be quite troubling for an unprepared player. Black bets everything on a direct attack. It will often transpose to Sec­ tion 3 . 2 so here we shall only cover independ­ ent variations. 17 f4 After 1 7 'iWfl Black can head for the main lines with 1 7 . . . 'iWh5 or go into an ending with 1 7 . . . 'iWxfl +. Black gets good compensation af­ ter either 1 8 l:.xfl f4 1 9 gxf4 iLxf4 20 ltJd2 iLc7 or 1 8 'iti>xfl f4 1 9 gxf4 iLxf4 20 �g l iLc7 . After the further 2 1 ltJd2 :f6 22 ltJe4 l:tg6 23 ltJg3 l:.f8 24 iLc2 iLf5 ! 25 iLxf5 l:txf5 26 iLd2 h5 27 l:te4 iLf4 28 iLxf4 ltJxf4 29 'iti>fl l:.f8 30

UNDERSTANDING THE MARSHALL A rrA CK

52

':'d l ltJh3 3 1 ':'d2 ltJg5 Black was clearly better in Kindennann-I.Sokolov, Bie1 1 988. 17 g5 Continuing the brute-force approach. Instead 17 . . .':'ae8 1 8 ltJd2 heads for Section 3.2, while 1 7 ... �h8 18 i.xd5 ( 1 8 ltJd2 runs into 1 8 . . . ltJxf4 ! 1 9 gxf4? { 1 9 i.xf4 i.xf4 20 i.d l = } 1 9 . . . i.xf4 20 ltJn i.f3 2 1 'ifd2 1i'g4+, when White can resign) 1 8 . . . cxd5 1 9 ltJd2 g5 20 'ifn 'ifh5 transposes to the main line. 18 'iWn Not, of course, 1 8 fxg5? i.xg3 ! , when Black wins. 18 'iWh5 19 ltJd2 �h8 The pressure on e3 and f4 induces White to take on d5 . 20 i.xd5 cxd5 (D) •••

'ifxf4 28 gxf4 ':'b4 gave Black good compensa­ tion in the endgame in Short-Ehlvest, World Cup, Skelleftea 1 989. 22 f4 23 .i.xf4 (D) •••

B

•••

23 .l:lxf4? ! Back in 1 989 Nunn recommended 23 . . . i.xf4 24 gxf4 .l:.ab8, "when it is not clear if White can prove any advantage". This has been tried since then and it seems that Nunn was right and this is a better choice. White has tried: a) 25 ':'a2 .i.h3 26 'ii'e 2 'ifxe2 27 ':'xe2 ':'xf4 28 ':'e3 ':'g4+ 29 ':'g3 ':'xg3+ 30 hxg3 .i.g4 = looks sufficient for Black. b) 25 'ifxa6 .i.f3 26 ltJxf3 and now not 26 . . . 'ifxf3 ? 27 'ife2 ± Berzins-Weegenaar, COIT. 1 994. Instead of recapturing on f3, Black has the surprising 26 . . ...g4+ ! (D). ••

w

We have a position similar to those in Chap­ ter 2, but here Black has started pawn-play with his queen's rook still on a8. Now White faces a common choice. He can immediately begin his counterplay on the queenside or he can try to make Black prove his case on the kingside. 52 A: 21 a4 53 B : 21 fxg5! ?

W

A) 21 a4 bxa4 2 1 . . .':'ae8? ! 22 axb5 axb5 transposes to Line B l in Section 3 . 2, which is dubious for Black. 22 fxg5 22 ':'xa4 :tae8 transposes to Line B 2 in Sec­ tion 3 . 2, while 22 c4 l:Iab8 23 .l:.abl .i.h3 24 1i'f2 'ii'g4 25 cxd5 gxf4 26 .i.xf4 .i.xf4 27 'ifxf4

After 27 �f2 1i'xf4 White must play with great care: 28 'ife6 ! (the only move) 28 . . . 'ii'xf3+ 29 ..ti>gl ':'xb2 30 'ii'e 5+ ..ti>g8 3 1 'ifg3 'ifxg3+ 32 hxg3 .l:.a8 should lead to a draw.

15 i.e3:

OTHER LINES

24 gxf4 l:tf8 25 l:teS �xe5 26 dxeS (D)

B

53

�h3 24 'iie 2 'iix e2 25 ':'xe2 ':'xf4 26 tDn l:tg4+ 27 tDg3 l:txg5 28 l:.ae 1 ':'ag8 29 ':'e5 was just winning for White in Neishtadt-Antosz, COIT . 1 959. 23 gxf4 l:tf8 24 l:.e5! �xeS 25 dxeS h6 (D)

w

Now: a) 26 . . . h6? ! 27 .xa6 ! bxg5 28 'ii'd6 ! ':'xf4 29 l:tn ! �f5 30 'ii'xd5 �h3 3 1 l:txf4 gxf4 32 'ii'f3 ± Sax-Ehlvest, World Cup, Skellefteii 1 989. The attack is spent. b) 26 . . . �e2 ! ? 27 'ii'g 2 l:txf4 and here 28 'ii'x d5 ? ! allows 28 . . . .:.g4+ 29 �h l 'ii'x g5, forc­ ing a draw, while after 28 �h l , both 28 . . . l:tf5 and 28 . . . �b5 retain drawing chances.

B) 21 fxgS! ? (D)

B

White is not afraid of a direct fight, espe­ cially considering the lack of pressure on the e-file. 21. f4 22 �xf4 l:txf4 Black has to go in for this to fuel his attack. Black gives up more material but his a8-rook will quickly join the fray. 22 . . . �xf4? 23 gxf4 ••

Black must play this to open the position. This may still be insufficient, but White must take some care. 26 ':'e1 This is the main continuation, but it is not clear if it is best. a) 26 gxh6? ':'g8 27 �f2 'ii'x h2+ 28 �e3 d4+ ! (28 . . . 'ifg3+? 29 �d4 ':'f8 chases the king, but fails to mate it) 29 cxd4 (29 �d3 dxc3) 29 ... 'ii'g 3+ 30 �e4 �f5+! 3 1 ..ti>d5 (3 1 ..ti>xf5 'ii'g6#) 3 1 . . ..g6 with a winning attack. b) 26 a4 hxg5 27 axb5 and now 27 . . . �e2 ? ! 28 'ii'g 2 ! gxf4 2 9 �h l is clearly better for White, but the immediate 27 . . . gxf4 ! ? looks like a superior try. c) 26 'ii'f2 hxg5 27 'ii'c 5 ':'xf4 28 ':'n �f3 29 'ifc8+ c:l;g7 30 tDxf3 ':'xf3 3 1 'it'd7 + �h6 32 'ife6+ ..ti>g7 33 'ii'e 7+ ..ti>h6 34 'ii'd6+ ..ti>g7 35 'ii'c 7+ c:l;h6 36 'ii'c 6+ ..ti>g7 37 'ii'b 7+ ..ti>h8 38 'ii'a 8+ ..ti>g7 39 'ii'a 7+ and now Black can im­ prove over 39 . . . ..ti>h8 ? 40 ':'xf3 'ii'xf3 41 'ifd4 ± Aseev-Frolov, St Petersburg 1 995 with 39 . . . ..ti>g8 40 ':'xf3 'ifxf3 4 1 'ifd4 ..ti>f7, when the king will come to e6 and Black should hold without much trouble. d) 26 e6 ! looks quite strong. 26 . . . hxg5 (or 26 . . . �xe6 27 'ii'e l �h3 28 'ii'e 5+ �h7 29 �h l ±) 27 e7 ':'xf4 28 'ii'd 3 iLd7 29 ':'e l �c6 30 tDn �g7 3 1 tDg3 'iff7 32 ':'e5 �f6 33 'ii'e 3 �e8 34 h4 1 -0 Christoffel-Gaertig, COIT. 1 989. 26 hxgS 27 fS! ':'xfS (D) •••

UNDERSTANDING THE MARSHAll A rrACK

54

1li'xg2+ 36 �xg2 ':'xd2+ 37 �f3 �f8 with a slight advantage for Black) 30 . . . .i.f3 3 1 'fi'd8+ g7 32 'fi'f6+ with perpetual check. Note that 32 1li'd7+? is worse because of 32 . . . 'fi'f7 ! , when only Black has winning chances. White has to exchange queens because of the . . . l:.g2 + threat. 29 'iIi'h4 30 :n i.f5 31 'ii'e3 l:.xfi+ 32 lDxn i.xh3 33 e6 'ili'g4+ '/2- '/2 A.Sokolov-Ehlvest, World Cup, Rot­ terdam 1 989.

w

•••

28 1li'd3 White has also tried 28 1li'g2 which looks even better. After 28 . . . i.h3 29 1li'e2 1li'h4 (29 . . . i.g4 30 1li'e3 +-) 30 e6 l::tf4 3 1 �h 1 l::tf2 32 1li'e5+ 'ito>h7 33 e7 .i.d7 White has the cute 34 lDf3 ! ltxf3 35 e81li' i.xe8 36 'fi'xe8 l:.f6 37 l:.e7+ 'it>h6 38 'fi'h8+ 'it>g6 39 'ili'g7+ 'ito>f5 40 l:.e2 ±. 28 .:If2! (D) 28 . . . i.h3 is worse, but not for the reasons given in most books. After 29 e6 ':'f4 White should play 30 e7 ! (instead of 30 'fi'e2 ':'g4+ 3 1 �h l i.g2+ 3 2 1li'xg2 ':'xg2 3 3 �xg2 1li'g4+ { not 33 . . . 1li'e8 ? 34 e7 +- } 34 �h l 1li'h4 ! =, when White's rook does not have a secure square on the e-file) 30 . . . .:.g4+ 3 1 �f2 ':'g2+ 32 �e3 and White's king escapes, leaving him with a decisive advantage.

Section 3 . 2 : The Pawn Push 1 6 . . . l:iae8 1 7 liJd 2 f5 1 e4 e 5 2 lDf3 lDc6 3 i.b5 a 6 4 i.a4 lDf6 5 0-0 i.e7 6 l:.el b5 7 .i.b3 0-0 8 c3 d5 9 exd5 lDxd5 10 lDxe5 lDxe5 1 1 ':'xe5 c6 12 d4 i.d6 13 l:.el 'fi'h4 14 g3 'ili'h3 15 i.e3 i.g4 16 'ili'd3 l:.ae8 17 lDd2 f5 (D)

•••

w

29 h3 Taking the rook loses, while 29 lDfl should also lead to a draw after 29 . . . l:.xb2 30 'ili'xd5 (worse is 30 e6? ! i.f3 3 1 'ii'd4+ g8 32 lDd2 'ili'h3 33 'ili'f2 'fi'g4+ 34 �fl i.g2+ 35 'ii'x g2

Here Black prefers to develop his rook be­ fore launching a pawn attack. This line is a little more controlled than that in Section 3 . 1 , but it is still risky. 18 1li'n We shall take this as the main line because it has been the preferred move-order amongst the elite. 1 8 i.xd5+ cxd5 1 9 f4 g5 20 1li'fl 'ili'h5 2 1 'fi'g2 �h8 transposes to the main line, while the most common move 1 8 f4 will also transpose after 1 8 . . . �h8 (or 1 8 . . . g5 1 9 1li'fl 1li'h5) 1 9 .i.xd5 cxd5, i f play continues 2 0 'fi'fl 'it'h5 . In­ stead 20 a4 bxa4 (20 . . . g5 2 1 'ili'fl 'fi'h5 again is a transposition to the main lines) 2 1 'ili'xa6 l:.xe3 ! 22 l:txe3 i.xf4 23 lDfl i.xe3+ 24 lDxe3 i.f3 25

15 i.e3:

OTHER LINES

'iVd6 ':e8 26 l:txa4 'iVh5 gives Black enough play for the pawn. 18 'iVh5 19 f4 (D) •••

19 'ith8 Black can also play 1 9 . . . g5 ? ! , but this does not seem to work. White has : a) 20 j,xd5+ cxd5 will just lead to the main lines: 2 1 'iVg2 'ith8 is Line A I . b) 20 a4 bxa4 (20 . . . �h8 2 1 j,xd5 cxd5 transposes to Line B 1) 21 .i.xdS+ cxdS 22 l:txa4? (22 'iVg2 is better, but unexplored) 22 . . . gxf4 23 gxf4 (23 .ltxf4 .ltxf4 24 ':xe8 'iVxe8 ! hits the a4-rook) 23 . . . :tf6 (23 . . . �h8, intending . . J:tg8, is also very strong) 24 'it'f2 :tg6 25 'ith 1 .lth3 is winning for Black, Vazquez-Granda, San Fer­ nando 1 99 1 . c) 20 fxg5 l:he3 2 1 l:he3 f4 22 :tf3 ! (we shall see this idea again) 22 . . . .i.xf3 23 'it'xf3 'it'xf3 24 tbxf3 fxg3 25 'itg2 gxh2 26 tbxh2 .ltxh2 27 'itxh2 .l:.f2+ 28 'itg3 l:hb2 29 .i.xd5+ cxd5 30 'itf4 l::te 2 3 1 a4 bxa4 32 ':xa4 :te6 is given by de Jongh as only marginally favouring White, but after 33 :taS he is much better. d) 20 'iVg2 gxf4 (20 . . . 'ith8 2 1 .ltxd5 cxd5 is Line A l below) 2 1 .ltxd5+ 'ith8 (worse is 2 1 . . .cxd5 22 'iVxd5+) 22 .ltxf4 .ltxf4 23 j,xc6 .lte3+ 24 'iith l :te6 (not 24 . . . .lth3 25 'iVe2 +-) 25 .i.f3 f4 (or 25 . . . .ltxd2 26 .ltxg4 l::txe l + 27 :txe l fxg4 28 'it'xd2 'iVd5+ 29 'itgl 'iVxa2 30 d5 +-) 26 .ltxg4 'iVxg4 27 'iVf3 'iVg7 was Dolma­ tov-Blatny, Pardubice 1 993. Now best is 28 gxf4 :txf4 29 'iVxf4 .ltxf4 30 l::txe6 'iVd7 3 1 :te2 with a winning ending - Blatny. 20 .ltxdS cxd5 (D) Compared to the main lines of the previous chapter, Black's rook is still on e8 because he •••

55

w

has spent a tempo on . . . �h8 to entice .ltxd5 . Even though White has given up his bishop, he must often do this anyway and Black' s inability to double rooks quickly gives White time to counterattack. Here he has a choice: A : 21 'iVg2 55 57 B: 21 a4

A) 21 'iVg2 (D)

B

The d5-pawn cannot be conveniently pro­ tected, so Black must leave it to its fate or try a speculative exchange sacrifice that we have seen before. We examine two possibilities for Black: 55 A I : 21 g5 56 A2: 21. l::te4 •••

••

AI) 21 gS •••

UNDERSTANDING THE MARSHALL A rrACK

56

Black plays for maximum confusion. How­ ever if White stays cool, he is certainly not without resources himself. 22 'ii'xdS White could also try 22 fxg5, and here: a) 22 . . . l:txe3 23 l:txe3 f4 24 1:1f3 ! �xf3 25 'ii'xf3 'ii'xf3 26 lLlxf3 fxg3 27 �g2 gxh2 28 lLlxh2 �xh2 29 'iltxh2 l:tf2+ 30 �g3 l:txb2 3 1 'iltf4 is the same as the ending as in note 'c' to Black's 1 9th move above, but Black's king is even worse on h8 than g8. b) 22 ... l:te4 23 1L1xe4 (23 h4 l:tee8 24 "f2 f4 25 gxf4 is given as +- by Bennedik, but maybe Black can make 25 . . . l:txe3 work after 26 l:txe3 1:Ixf4 or 26 'ii'xe3 .i.xf4) 23 . . . fxe4 24 l:tn .i.f3 and now B ruckel-Bennedik, COTTo 1 998 con­ cluded 25 1:Ixf3 (25 'ii'd 2? 'ii'h 3 ! is good for Black: 26 'ii'f2? h5 ! 27 gxh6 l:tg8 -+ or 26 g6? .i.xg3 ! 27 g7+ 'iltxg7 28 .i.h6+ 'ilth8 29 �xf8 �f4 ! ) 25 . . . l:txf3 26 .i.d2 e3 27 �e 1 'ii'g4 28 a4 bxa4 29 'ii'c 2 e2 30 �f2 l:txg3+ 112- 112. 22 l:tdS 22 . . . gxf4? 23 'ii'xd6 fxe3 24 l:txe3 wins for White because Black's rook is loose on f8 . 2 3 'ii'c6 gxf4 2 4 .i.xf4 Not 24 gxf4? l:tg8. 24 .i.xf4 25 gxf4 .i.e2 26 'iti>hl (D)

l:txe3 l:txe3 3 1 "a8+ l:te8 32 'ii'c 6 l:te2 33 l:tg2 (33 'ii'f6+ �g8 34 1L1f3 'ii'h 3 35 'ii'd 8+ 'iti>g7 36 l:tg3 'ii'n + 37 1L1g 1 l:txb2 is also OK for B lack) 33 . . . l:te 1 + 112- 112 Popovic-Nunn, World Team Ch, Lucerne 1 989. 27 .....h4 2S 1L1f3! Instead 28 'ii'd6 .i.g4 29 l:tae 1 (29 l:taf1 �e2) 29 . . .'ilf2 30 l:txe8 .i.f3+ is a perpetual, while 28 'ii'g 2 l:tg8 29 'ii'c 6 l:tgf8 30 'ii'g 2 l:tg8 3 1 'ii'c 6 was drawn in A.Sokolov-Nunn, World Cup, Rotterdam 1 989. 2s :ifxf4 29 lLlgS ! 'ii'e3 Not 29 . . . �g4? 30 1Llf7+ cJ;;g 7 3 1 l:taf1 +­ (A.Sokolov) . After 29 . . . 'ii'e 3, both 30 l:tg2 and 30 l:.g3 , intending t o double on the g-file and perhaps advance the d-pawn, leave Black with prob­ lems . .•

A2) 21 1:Ie4 (D) •••

•••

w

•••

B

26 l:.deS After 26 . . . l:.d6? Nunn gives 27 'ii'g 2, but White has even better in 27 'ii'xd6 ! .i.f3+ 28 'iti>g l l:tg8+ 29 'itf2 'ii'xh2+ 30 'iti>e3 l:te8+ 3 1 'iti>d3 .i.e4+ 32 1L1xe4 fxe4+ 3 3 'iti>e3 'ii'g 3+ 34 'itd2, winning as there is no perpetual check. 27 l:tgl 27 'ii'g 2 l:te3 gives Black enough activity. For example, 28 l:tg 1 .i.g4 29 l:tae 1 l:tfe8 30 •••

Black offers an exchange to obtain light­ square domination. This idea is very attractive, and even though it is rather speculative, Black really burned his bridges anyway when playing the Pawn Push. In any case, this seems like the best try. 22 1L1xe4 22 h4 h6 23 1L1xe4 fxe4 will transpose, while 22 a4 is also possible. Black should reply 22 . . . bxa4, transposing to Line B2 below, be­ cause after 22 . . . g5 23 fxg5 (or 23 1L1xe4 fxe4 24 fxg5) 23 . . . �h3 24 "f3 l:tg4 25 axb5 f4 26 �xf4 �xf4 27 bxa6 "xg5 28 a7 Black has problems:

15 .i.e3:

OTHER LINES

a ) 28 . . . 'fig8 is given b y Harding but after Nunn' s recommendation 29 'iit h l White is better: 29 ... .i.xd2 30 'fixfS 'fixfS 31 a8'ii' :g8 32 'it'xfS I1xfS 33 l1e2 ±. b) 28 . . . l1a8 and now 29 l1e8+(?) I1xe8 30 a8'fi l:lxa8 3 1 ':'xa8+ 'iitg 7 32 l1a7+ is a draw Nunn. However, 29 �e4 ! dxe4 (otherwise 30 �f2 +-) 30 l:lxe4 looks winning for White. 22 fxe4 23 h4 h6 24 lin I1f6 Insufficient is 24 . . . g5? ! 25 'it'h2 gxf4 26 .i.xf4 l:lxf4 27 l:lxf4 .i.xf4 28 gxf4 e3 29 "g3 e2 30 �f2 'ii'f5 3 1 h5 +- Klover- Kant, Ostrava 1 98 1 . 25 a4 lIg6 26 axb5 axb5 (D)

57

B

•••

White prefers to activate his rook rather than spending time with his queen hunting the d5-pawn. Black has a decision to make be­ tween two thematic ideas : 57 B l : 21. g5? ! 58 B2: 21,..bxa4

W

••

B1)

Now Topalov suggested 27 l1a7, while after 27 f5 .i.xf5 28 .i.f4 .i.e7 29 .i.e5 .i.xh4, White has a choice: a) 30 l1a7 .i.xg 3 ? ! (30 . . . .i.f6 3 1 .i.xf6 ':'xf6 32 I1f4 I1g6 with compensation, Ganin-Arias Duval, corr. 2003) 3 1 .i.xg3 e3 32 lIe7 .i.e4 33 l1e8+ ? ! (33 I1fS+ �h7 34 :xe4 100ks better, as then . . . e2 won't hit the rook) 33 . . . �h7 34 l:lxe4 dxe4 35 'ii'xe4 e2 ! (not 35 . . . 'ii' h 3 ? 36 11f3 e2 37 'ii'b 1 ! h5 38 �f2 h4 39 .i.e5 'iWg2+ 40 'iite 3 1 -0 Anand-Topalov, Las Palmas 1 993) 36 lIe l (not 36 l:lf3 ? e 1 'iW+) 36 . . . 'iWh3 and Black wins back the piece with no problems. b) 30 'iWh2 ! ? e3 31 l:la7 .i.d3 and now 32 l:lxg7 l:lxg7 33 .i.xg7+ �xg7 34 'iWxh4 'ii'xh4 35 gxh4 .i.xfl 36 �xfl only liquidates to a drawn pawn ending, but 32 :e 1 seeks more: 32 . . .'iWf3 (32 . . . 'iWxe5 33 dxe5 .i.xg3 34 'ii'x g3 l:lxg3+ 35 �h2) 33 'fig2 'fixg2+ 34 'iitxg2 .i.e4+ 35 �h3.

B) 21 a4 (D)

21...g5?! This is the most direct, but it is probably dubious. 22 axb5 axb5 23 fxg5 Other moves are less testing: a) 23 l1a6 ? ! gxf4 24 .i.xf4 gives Black a pleasant choice: al ) 24 . . . .i.h3 25 I1xe8 (25 'ii'f2? ':'xe 1 + 26 'iWxe l .i.xf4 27 gxf4 and now 27 . . . l1g8+ or 27 . . . 'iWg4+ ) 25 . . . .i.xfl 26 ':'xfS+ .i.xfS 27 �xfl gives White sufficient compensation for the queen, but no more than that. a2) 24 . . . l:lxe 1 25 'iWxe 1 .i.xf4 26 gxf4 .i.e2 27 �fl 'iWf3 2 8 'iWf2 'iWd3 29 �g3 'ii'd 1 + 30 'iitg 2 .i.d3 gives Black good attacking chances according to Nunn. b) 23 'fif2 is interesting. After 23 . . . h6 White has: b 1 ) 24 �fl :e4 ! - Nunn. b2) 24 b3 I1g8 25 c4 .i.h3 26 fxg5 hxg5 and now 27 l:la6(?) .i.b4 intending . . .f4 is given by Nunn, but White may improve here, perhaps with 27 c5 . b3) 24 l:la6 l:le6 25 �b3 gxf4 26 I1xd6 (26 .i.xf4 l:lxe 1 + 27 'ii'xe 1 .i.xf4 28 gxf4 .i.e2 ! 29 h3 b4 +) 26 . . . l:lxd6 27 .i.xf4 gives White com­ pensation for the exchange. 23,..lIxe3 (D)

UNDERSTANDING THE MARSHAU A TTACK

58

28 "g2 �xg3 29 'ifxg3 h5 Or 29 . . . 'ifxd2 30 'ifxg4 ':'g8 3 1 'it'f3 "g5 32 :fl ± Nunn. 30 l:.n .l:.xfi+ Nunn gives 30 . . . l::tg 8 3 1 h3 'ifxd2 32 "e5+ .l:.g7 33 hxg4 hxg4 34 "h5+ �g8 35 'iVxd5+ cj;h7 36 "f5+ intending ':f2 ±. 31 ttJxfi 'iVc1 32 �gl 'ifxb2 (D)

w

w

Black breaks down the defence of the dark squares. While the attack may be strong enough to win back material, Black will still be strug­ gling. 24 ':'xe3 f4 25 gxf4 .i.xf4 26 l:.g3 26 'ii'f2 is another possibility, although after 26 . . . 'ii'x g5 27 h l .i.f5 28 .l:.ae 1 .i.xe3 29 1be3 it may not be so easy for White to win. 26.....xg5 (D) This endgame is critical for the Pawn Push. White has excellent winning chances. 33 h3 iLdl 34 ttJe3 'iVbl 35 �g2 After 35 'ife5+ ? ! cj;h7 36 'ii'f5+ "xf5 37 ttJxf5 cj;g6 ! Black's active king allowed him to hold the draw in Timman-Nunn, B russels 1 988. 35 'ifb2+ Even worse is 35 . . . b4? 36 "e5+ �h7 37 ttJf5 +- Haba-Dobrovolsky, Czech Ch, Trinec 1 988. 36 "f2 "c1 Or 36 . . . "xf2+ 37 �xf2 .i.b3 38 ttJg2 <J;; g 7 39 ttJf4 �h6 40 h4 ±. 37 'iVf6+ �h7 38 'ife7+ �g6 39 ttJxd5 ± Nunn.

w

•••

27 cj;hl After 27 ttJf3 ? Black has the amazing reply 27 . . . 'ii'g 8 ! ! -+ (instead of 27 . . . 'ii'g 7? 28 :a7 ! 'ifxa7 29 ':'xg4) , which wins after 28 ':'xg4 'ifxg4+ 29 'ifg2 .i.e3+ 30 'itth l 'ii'xf3 . The text-move was investigated in detail in the late 1 980s and the analysis by Nunn and others has held up very well. 27 .i.d6 27 . . . .:.g8 28 ttJf3 'iff5 and now 29 ttJe5 was given by Nunn, while 29 ttJh4 may be even better after 29 .. :it'g5 30 ':'g2 or 29 . . . 'ii'e4+ 30 ttJg2. •••

82) 21 bxa4 (D) This typical deflection is probably best. 22 .l:.xa4 After 22 'ii'g 2 it is a good time for 22 . . . .:.e4 ! (22 . . . g5 23 'ii'xd5 gxf4? fails to 24 'ii'xd6 fxe3 25 ':'xe3 because the fS-rook is loose) 23 ttJxe4 (23 l::txa4 g5 24 ttJxe4 fxe4 25 ':'xa6 gxf4 26 ':'xd6 fxe3 27 ':'xe3? { 27 ':fl may survive } 27 . . . .i.h3 28 g4 'ifxg4 ! 0- 1 Timman-Nunn, Wijk aan Zee Radio 1 983) 23 . . . fxe4 24 'iffl (after 24 .l:.xa4 •••

15 .i.e3:

OTHER LINES

59

23 .:e6 (D) After 23 . . . h6 24 'iif 2 (24 'iix a6 is also pos­ sible) 24 . . . .:e6 25 c4 ':c8 (Haba-Vajda, 2nd B undesliga 2005/6), Haba gives 26 cxd5 ':g6 27 ':ac 1 +- with the idea lDc4-e5 . •••

w

w

both 24 . . . g5 and 24 . . . �f3 are unclear) 24 . . . �f3 25 'iixa6 'iVh3 26 :'e2 (26 'iif l 'iig4 intending . . . h5-b4 - Nuon) 26 . . . �xe2 27 'iixe2 g5 28 lha4 gxf4 29 �xf4 �xf4 30 gxf4 e3 ! gave Black sufficient play in Hindle-P.Littlewood, England 1 987. 22 g5 (D) •••

w

23 ':aal ! This is really the only move, as White needs to defend his back rank. Other moves are bad: a) 23 :'xa6? gxf4 24 �xf4 (Black loses af­ ter 24 gxf4 ':g8 25 �h l �h3 -+, 24 l:lxd6 fxe3 25 lDb3 f4 -+ or 24 �f2 �h3 ! 25 'iid 3 ':xe l + 2 6 �xe l 'ii'd l -+) 24 . . . ':xe l 2 5 'fIxe l �xf4 26 gxf4 .i.e2 =t. b) 23 fxg5 ? ! ':xe3 24 ':xe3 f4 25 gxf4 (25 ':f3 �xf3 26 'fIxf3 'iie 8 ! ) 25 ... �xf4 26 ':g3 (White may be able to hold with 26 'iif2 but he has nothing to be happy about after 26 . . . 'iix g5 27 l:lg3 ':f6) 26 . . .'fIe8 27 ':xg4 �e3+ ! 28 �g2 ':xfl 29 lDxfl .i.c I ! =t Ulmanis-Van der Heij­ den, COIT. 1 986.

24 fxg5 Nunn pointed out back in 1 989 that this move may be critical. Now it is the main line, but other moves are also possible: a) 24 :'xa6? is a mistake. After 24 . . . gxf4 25 ':xd6 (25 �xf4 ':xe l 26 'iixe l �xf4 27 gxf4 �e2 0- 1 Janosevic-Blatny, B ad Worishofen 1 989) 25 . . . .:xe3 ! -+ (25 . . . fxe3 26 ':xe6 exd2 27 ': 1 e2 f4 is good for Black according to Nunn but after 28 ':6e5 White may defend) White has no good moves; for example, 26 gxf4 �e2, 26 :'e6 fxg3 27 'iig 2 lhe6 28 ':xe6 �h3 or 26 ':xe3 fxe3 27 lDb3 �h3 28 'iie l 'iif3 . b) 24 'fIf2 ! ? is also critical. 24 . . ..:fe8 (24 . . . h6 is Haba-Vajda in the note to Black's 23rd move above) 25 fxg5 h6 26 gxh6 �h3 (after 26 . . . f4 27 �xf4 ':e2 White has 28 .i.e5+! �xe5 29 ':xe2 �xe2 30 dxe5 +-) 27 lDf3 f4 28 �xf4 'iixf3 29 �e5+ (White can also try 29 'fIxf3 ':xe l + 30 ':xe l ':xe l + 3 1 �f2 ':fl + 32 �e3 ':xf3+ 33 �xf3 but the text-move looks stron­ ger) 29 . . . .:xe5 30 'iixf3 ':xe l + 3 1 �f2 l::t I e6 32 'iih5 ':f8+ 33 �gl �f5 34 l:tfl ± Shen Yang-Hebden, Gibraltar 2008. 24 ':fe8 24 . . .f4 ? ! 25 �xf4 �e2 26 'iif2 �xf4 27 gxf4 'iig4+ 28 'fIg3 ':xf4 29 'ii'xg4 ':xg4+ 30 �h l or 24 . . . ':xe3 25 ':xe3 f4 26 ':f3 �xf3 27 'iixf3 'iixf3 28 lDxf3 fxg3 29 'iii>g 2 is an even worse version of the ending we have seen already. 25 'ii'f2 f4 26 gxf4 (D) •••

UNDERSTANDING THE MARSHALL A rrA CK

60

B

B

26 .i.h3 Black has also tried 26 . . . h6. Then: a) 27 h4 hxg5 28 hxg5 c:itg7 (28 .. Jlh6? is creative but unsound: 29 'ii'g 2 ! ':'h7 30 c:itf2 ! .i.xf4 3 1 .i.xf4 ':'f8 32 ':'e5 .i.h3 33 'ii'g 3 ':'xf4+ 34 'it>e l J:g4 35 'ii'e 3 1 -0 Yanes-Cortinas, COIT. 1 992) 29 ttJfl .i.f3 30 ttJh2 ':'h8 3 1 'ii'g 3 .i.e2 32 .i.d2 J:e4 is unclear. White is three pawns up but it is not easy to consolidate. b) 27 ttJfl hxg5 28 ttJg3 'ii' h 3 29 fxg5 ':'f8 (29 . . . .i.xg3 30 hxg3 .i.f5 has been suggested but it looks insufficient) 30 'ii'g 2 'ii'h4 3 1 .i.f2 ':'xe l + 32 ':'xe l .i.f3 33 'ii'f l .i.e4 34 ':'xe4 (worse is 34 'ii'e 2 .:tf3 35 .:td l 'ii'g 4 with the initiative, but 34 ':'e3 is possible since after 34 . . . .i.f4 35 'ii'xa6 .i.xe3 36 .i.xe3 'iWg4 Vladi­ mirov's '=t' is over-optimistic because 37 'ii'h6+ 'ii;> g 8 38 'ii'h 5 'ii'e6 39 .i.d2 .i.f3 40 'ii'h 6 'ii'x h6 41 gxh6 leaves White no worse) 34 . . . dxe4 35 'ii'g 2 11f3 (also plausible is 35 . . . 'ii'x g5 36 'ii'h 3+ 'ii;> g 8 37 'ii'e 6+ ':'f7 38 'ii'xd6 'ii'd 2) 36 g6 (36 ttJfl ! ?) was Dolmatov-Vladimirov, European Clubs Cup, Moscow 1 989. Now Black should play 36 . . . .i.xg3 ! 37 .i.xg3 'ii'h 6 38 .i.e5+ 'ii;> g 8 39 'ii'c 2 'ii'xg6+ 40 .i.g3 'ii'g4 +. 27 'ii;> h 1 Instead 27 ':'xa6 1eads nowhere for White af­ ter 27 . . . ':'xe3 28 ':'xe3 'ii'g4+ 29 'ii'g 3 'iWd l + 30 'ii'e l 'ii'g4+ 3 1 'ii'g 3 'ii'd l + 32 'iWe l lh- lh Liang linrong-Blatny, Novi Sad Olympiad 1 990. 27 h6 28 ttJf3 (D) After 28 gxh6 .i.c7 ! Black seems to have enough play. 29 ttJf3 .i.g4 30 ttJg5 .:tg8 3 1 'ii;> g l .i.f3 32 .i.d2 (not 32 b4? .:txg5+ 33 fxg5 .:te4 34 ':'e2 'ii'h 3 0- 1 Schulz-Monner Sans, COIT. 1 994) 32 . . . J:xg5+ 33 fxg5 'ii'g4+ 34 �fl 'ii'h 3+ 112- 112 Lovakovic-Arias Duval, COIT. 2000. •••

•••

28 .i.f5 Black should avoid 28 . . . .i.xf4? 29 ttJe5 .ixe5 30 dxe5 hxg5 3 1 .:tg l ! g4 32 .i.d4 'ii'g 6 33 J:ae l c:ith7 34 .:te3 'ii'h5 35 b4 ± Kostal-Vaindl, COIT. 200 1 . 29 ttJe5 hxg5 30 fxg5 .i.e4+ 3 1 'ii;> g l :rs 32 .if4 'ii;> g8 33 .:txe4 Otherwise 33 . . . .i.xe5 34 dxe5 'ii'g 4+. 33 dxe4 34 'ii'g3 .i.xe5 35 dxe5 .:tee8 36 .:tel .:td8 37 .:txe4 .:td1 + 38 'ii;> g2 .:tb1 39 e6 l:xb2+ 40 'it>gl J:b1+ 112-112 Brookes-Olofson, COIT. 1 998. .••

•••

Section 3 . 3 : The Ada ms Va riation 1 6 . . . �ae8 1 7 liJd 2 'ti'h 5 1 e4 e5 2 ttJf3 ttJc6 3 .i.b5 a6 4 .i.a4 ttJf6 5 0-0 .i.e7 6 .:tel b5 7 .i.b3 0-0 8 c3 d5 9 exd5 ttJxd5 10 ttJxe5 ttJxe5 1 1 .:txe5 c6 12 d4 .i.d6 13 .:tel 'ii'h4 14 g3 'ii'h3 15 .i.e3 .i.g4 16 'ii'dJ .:tae8 17 ttJd2 'ii'hS (D) This kind of 'Accelerated Spassky' variation was invented by Adams by accident( ! ) against Ivanchuk when he simply 'left out' 17 . . . J:e6 1 8 'ii'f l . Nevertheless, 1 7 . . .'ii'h5 has proved to be a viable move-order, especially if Black is will­ ing to transpose back into the main lines of Chapter 1 . This line also has the virtue of avoid­ ing Krarnni k's line, which is covered in Section 3 .4. By leaving out . . . .:te6, Black ensures that 'ii'f l from White will not gain a tempo by chas­ ing Black's queen away, and Black also avoids pinning himself on the a2-g8 diagonal. Black sometimes will still play . . . J:e6, but there is also

15 i.e3:

OTHER LINES

the option to double rooks on the e-file starting with . . . l:Ie7. White has these main tries: 61 A : 18 �c2 63 B : 1 8 'ii'n C : 18 a4 64 I S Jtd l �xd l 1 9 l:axd l f5 ! gives Black ex­ cellent play, and 20 f4? loses immediately to 20 . . . �xe3 because the d l -rook is hanging. White's only other real possibility is I S �f1 .l:te6, when 19 Jtd I ? ! ( 1 9 a4 Jtf5 is like 17 . . . l:Ie6 I S a4 'it'h5 1 9 axb5 axb5 20 �f1 of Chapter 2, but White has delayed axb5 here, which is likely to transpose but may favour Black if White does not get an opportune time to play axb5) 1 9 . . . f5 20 �xg4 'it'xg4 2 1 .i.d2 .l:tg6 22 ..tg2 ? ! (both 22 c4? ! bxc4 23 'it'xc4 ..ti>hS + and 22 f3 'ili'h3 23 'ili'e2 f4 24 'ii'g 2 'it'd7 25 g4 h5 26 h3 hxg4 27 hxg4 c5, with compensation, are given by Adams) 22 . . .f4 23 f3 'ili'h5 24 g4 'it'h4 ! (not 24 . . . l:xg4+? 25 fxg4 'ili'xg4+ 26 �h 1 f3 27 �g3 Jtxg3 2S :g 1 +-) 25 l:e2 l:Ixg4+ 26 fxg4 f3+ 27 ..ti>h l fxe2 2S 'ii'xe2 'ii?h S gave Black a strong initiative in Ivanchuk-Adams, Terrassa 1 99 1 .

A) 18 �c2 (D) White switches to the bl -h7 diagonal, creat­ ing pressure against h7. 18 (5 This is logical because White has removed his bishop from the a2-gS diagonal. Ponomar­ iov analysed some other ideas in NIC Maga­ zine: ...

61

a) IS . . . g6 1 9 f3 ( 1 9 �e4 l:xe4 20 "xe4 �f5 2 1 Jtd l 'iWxd l 22 l:axd l �xe4 23 �h6 f5 24 JtxfS 'ii? xfS is unclear) 1 9 . . . �f5 20 �e4 'ili'xf3 2 1 Jth6 Jtxe4 22 ':'xe4 'ii'x d3 23 �xd3 l:xe4 24 .i.xe4 .l:teS 25 Jtf3 is a little better for White because of the bishop-pair. b) I S . . . c5 ? ! (D) and here:

b l ) 1 9 dxc5 �xe3 20 fxe3 Jtxc5 2 1 �b3 Jtb6 22 �d4 l:e5 23 �d 1 l:feS 24 �xg4 'ii'xg4 25 .l:te2 h5 26 l:If1 l:.e4 gives Black compensa­ tion for the pawn. b2) 19 Jtd l ! ? �f6 (better than 19 . . . �xd l 20 l:.axd l :e6 2 1 �f1 c4 22 'ili'e2 'it'f5 23 ..ti>g2 .l:tf6 24 Jtc 1 h5 25 h4 when despite Black's op­ tical compensation, White should slowly real­ ize his extra pawn) 20 dxc5 Jtxc5 2 1 �f1 l:.dS 22 'ili'c2 and it is hard for Black to prove full compensation. After 22 . . . .l:txd l 23 .l:taxd l Jtf3 (Ponomariov) 24 h4, White parries the main threat and avoids structural damage, so Black should try the more patient 22 . . . l:.feS. 19 f3 (D)

UNDERSTANDING THE MARSHAU A TTACK

62

B

White is willing to return the pawn to get the bishop-pair and secure some initiative. How­ ever, Black does not have to accept the offer. 19 i.h3 ! ? Black prefers to play for a n attack. Taking the pawn is also reasonable, although Black is going to need to be ready to defend an ending where White has the bishop-pair 'for free' . 1 9 . . . i.xf3 20 lLlxf3 'ii'xf3 2 1 i.d2 'ii'xd3 22 i.xd3 f4 (22 . . . l:txe l + 23 %he l g6 24 a3 c:J;; f7 25 c4 lLlf6 26 c:J;;g 2 ;t Ponomariov) 23 ':xe8 ':'xe8 24 c:J;; f2 ! (24 .1:.e l ':'xe l + 25 i.xe l fxg3 26 hxg3 g6 27 c4 i.b4 28 i.f2 bxc4 29 i.xc4 as gives Black ex­ cellent drawing chances) 24 . . . fxg3+ 25 hxg3 c5 26 c4 lLlb6 27 cxb5 c4 28 i.fl axb5 ( ? ! Ponomariov, who suggests 28 . . . 1:.f8+ 29 ..t>g2 axb5 30 i.e2 i.e7 3 1 i.g4 i.f6 32 i.e6+ ..t>h8 33 1:.fl ;t) 29 i.g2 1:.f8+ 30 i.f3 and although Ponomariov implies that White is close to win­ ning here, this does not look so clear after 30 . . . b4 ! ? Instead 30 . . . i.e7 3 1 i.a5 lLla4 32 b3 i.f6 33 c:J;; g 2 i.xd4? 34 ':'fl +- was Ponomar­ iov-Leko, Tal Memorial, Moscow 2008. The threat of 35 i.d5+ wins material. 20 i.f2 (D) 20 lLlf4! ? This sharp piece sacrifice was played a cou­ ple of times by Sargis ian, who is a frequent as­ sistant to Marshall expert Aronian. a) In Sutovsky-A.Heimann, European Clubs Cup, Kallithea 2008, 20 .. :it'g5 was met ruth­ lessly by 2 1 ':'xe8 l:txe8 22 1:.e l ':'xe l + 23 i.xe l h5 24 'ii'e2 ! h4 25 'ii'e 6+ c:J;; h 7 26 lLlfl i.xfl ? ! 27 i.xf5+ g6 28 'ii'f7+ c:J;;h 6 29 i.d2 ! i.f4 30 'ii'f8+ �h7 3 1 i.e6 ! 1 -0. b) 20 . . . 'ii'g 6 21 1:.xe8 1:.xe8 22 1:.e l 1:.xe l + 2 3 i.xe l h 5 24 i.b3 (24 lLlb3 h4 ! 2 5 g4 i.xg4 ! •••

•••

26 fxg4 'ii'xg4+ 27 c:J;; h l h3 28 i.g3 f4 - Pono­ mariov) 24 . . . c:J;; h7 25 i.f2 (25 lLlfl lLlf4 26 'ii'd 2) 25 . . . lLlf4 26 'ii'e 3 h4 27 lLlfl lLld5 28 'ii'd 3 (28 'ii'e l f4 29 'ii'e4 i.f5 30 i.c2? lLle7 3 1 g4 c:J;;h 8 ! -+) 28 . . . lLlf4 29 'it'e3 lLld5 30 'it'd3 lLlf4 3 1 'it'e3 1/2- 1/2 Ponomariov-Gustafsson, Spanish Team Ch, Cala Mayor 2008. In his annotations, Ponomariov indicated that he had trouble find­ ing improvements for White in this game. 21 gxf4 i.xf4 22 i.g3 'ii'gS (D)

On the surface, it looks like Black should not have enough here, but White has not managed to prove anything in practice. 23 lLln Later Ni Hua tried to improve with 23 i.b3+ c:J;;h 8 24 lLlfl , but after 24 . . . h5 25 :Xe8 ':xe8 26 i.f7 ':'e7 27 i.xh5 i.xfl 28 'ii'x fl i.e3+ 29 c:J;;h 1 'ii'xh5 30 'ii'g 2 (Black also easily held after 30 i.e5 l:te6 3 1 1:.el f4 32 ':'e2 1:.g6 33 1:.g2 'it'h3 34 'it'e2 'ii'f5 1/2- 1/2 in Landa-Amonatov, Aeroflot Open, Moscow 2(09) 30 . . .f4 3 1 i.f2 'ii'dS 32 i.h4 1:.e8 Black had sufficient compensation for

15 iLe3:

OTHER LINES

the pawn in Ni Hua-Sargisian, Dresden Olym­ piad 2008. 23 h5! 24 �f2 h4 25 IIxeS IIxeS 26 lIel hxg3+ 27 hxg3 lIdS (D) •••

w

63

This idea is always a possibility i n the Ad­ ams Variation. Black doubles rooks on the e­ file without pinning himself on the a2-g8 diag­ onal. Another idea is 1 8 . . . �h3 19 �d 1 'iif5 20 'iie2 lIe7 2 1 'iif3 ? ! (2 1 'iih 5 ! ? looks like a better try) 2 l . . .'iid 3 22 �e2 ? ! 'iic 2, which was very awkward for White in Staak-Jacot, COIT. 2007 . 19 a4 IIfeS 20 axb5 axb5 The pressure on e3 forces White to trade off his light-squared bishop. 21 �xd5 'iixd5 (D)

w

2S �e2 White cannot play 28 gxf4? 'iig 2+ 29 'ite3 lIe8+, but he could consider 28 'iie 2 ! ? with the idea 28 . . . �xfl 29 'iie 6+ �h8 30 gxf4 'it'g2+ 3 1 'iite 3. Instead White heads for a draw. 2S �xfl+ 29 IIxfl lIeS+ 30 'iitd 1 'iixg3 31 'iixf5 1Ie1+ 32 IIxe1 'iixf3+ 33 l:te2 'iifl + 34 1Ie1 'iif3+ '/2- 1f2 Efimenko-Sargisian, Dresden Olym­ piad 2008. ..•

B) lS 'iifl (D)

This is a typical positional situation where Black's activity and bishop-pair allow him to hold the balance. 22 'iig2 White got nowhere after 22 f3 �d7 23 ttJe4 �f5 24 c4 bxc4 25 ttJc3 'iie6 26 �f2 'it'd7 27 IIxe7 �xe7 28 'iixc4 �e6 29 'iia4 .::tb 8 30 d5 '/2- '/2 in Chuprov-Sargisian, Moscow 2009. 22 ... 'iixg2+ 23 'iit xg2 h5 24 lIaS h4 25 .::teal hxg3 26 hxg3 �e6 (D)

w

This has been tried in a couple of high-level encounters recently. lS ...l:te7 ! ?

64

UNDERSTANDING THE MARSHAU A TTACK

27 :a8 .i.dS+ 28 ..tn .i.xg3 29 .i.gS f6 30 .i.xf6 :e1+ 31 :xe1 :xa8 32 fxg3 gxf6 33 lbe4 rl;;f7 (D)

At first glance it looks like White is a healthy pawn up, but both of Black's pieces are very ac­ tive. White has to watch out for . . . :a2 because it is not easy to defend the b2-pawn - :e2 will run into . . . .i.c4. 34 lbd6+ 'iit g6 3S :e8 35 rl;;f2 :d8 ! dislodges the knight, and 36 lbe4 :h8 shows White's difficulties in coordi­ nating his forces. 3S .. ,lla l + 36 :e1 :a8 37 :e8 :al + 38 :e1 1/2- 1/2 B acrot-Aronian, FIDE Grand Prix, Nalchik 2009.

C) 18 a4 (D)

White continues along normal lines and of­ fers to transpose back into Chapter 1 .

18 i.fS Black tries to repeat by chasing the white queen around. 1 8 . . . .:e6 transposes to Chapter 1 , and this may be a better idea. 19 'ii'n The only alternative to this (besides repeat­ ing moves) is 1 9 .i.d l , but after 1 9 . . ...xd l 20 'fixf5 lbxe3 2 1 'ii'd 3 'fic2 22 :xe3 'ifxb2 23 :bl .a2 24 axb5 axb5 25 lbe4 i.e7 26 ':be l :d8 the ending was just equal in R.Leyva­ Pupo, Match (game 2), Holguin 1 992. 19...i.h3 20 i.dl 'iffS 21 'ife2 (D) •••

21 ... cS With White's pieces a little bottled up, Black tries to open the position. We have seen this idea before - Black wants either to take on d4, giv­ ing White an isolated d-pawn, or to open things up before White can coordinate his forces. Tactical excursions do not work: 2 1 . . .lbf4 22 'fif3 lbd3 23 "xf5 i.xf5 24 i.c2 ! lbxe l ? 25 i.xf5 +- and 2 1 . . .i.f4 22 'ii'f3 .i.xe3 23 'fixf5 .i.xf5 24 fxe3 :xe3 25 :xe3 lbxe3 26 i.f3 ':c8 27 axb5 axb5 28 :a6 i.d7 29 lbe4 ± Milos-Martins, Campinas 1 994 should both be avoided. 22 lbO! (D) White would like to meet 22 . . .cxd4 with 23 lbxd4 ±. Instead 22 axb5 can be answered by 22 . . . cxd4 ! , and here: a) 23 cxd4 axb5 24 lbn lbb4 ! gave Black the initiative in Volokitin-Lastin, Budva 2004. b) 23 lbc4 lbf4 ! 24 'fif3 dxe3 25 lbxd6 exf2+ 26 'fixf2 %be l + 27 'ii'xe l 'fic5+ 28 'ii'f2 'ii'xd6 29 'ii'xf4 'fic5+ 30 'ii'f2 'ii'xb5 leads to equality. 22...lbxe3! ?

15 JLe3:

OTHER LINES

This looks best. Alternatives : a ) 2 2 . . . .tg4? loses t o 23 lt:lh4 +-. b) 22 . . . h4? ! 23 'it'xa6 bxc3 24 bxc3 It:lxc3 25 'it'xd6 lt:lxd 1 26 lt:lh4 .l:.xe3 27 .l:.axd 1 lhe 1 + 28 .l:.xe 1 'it'c2 29 dxc5 'it'xa4 was Morovic-Ad­ ams, Match (game 2), Santiago 1 997. Here 30 c6 ! would give White a winning advantage. c) 22 . . . .ltf4 23 'it'd2 lt:lxe3 24 fxe3 .th6 25 axb5 axb5 and now 26 .l:.a5 was given by Ad­ ams, while 26 'it'f2, as played in some corre­ spondence games, looks even stronger. d) 22 . . . c4 was suggested by Adams. Black wants to play . . . It:lf4-d3 . After 23 axb5 axb5 24 It:lh4 (the immediate 24 'it'd2 was suggested by Bennedik) 24 . . .'it'd7 25 'it'd2 White intends .to and It:lg2, when it is not clear that Black has enough compensation. After 22 . . .lt:lxe3 23 fxe3 cxd4 Black will meet 24 cxd4 with 24 . . . .th4 and 24 lt:lxd4 with 24 . . . 'it'g5 . In both cases Black's bishop-pair and piece activity should give him enough compen­ sation for the pawn.

Section 3 .4 : Wh ite Avoids 18 a4 1 e4 eS 2 It:lf3 It:lc6 3 .tbS a6 4 .ta4 lt:lf6 S 0-0 .te7 6 .l:.el bS 7 .tb3 0-0 8 c3 dS 9 exdS lt:lxdS 10 It:lxeS lt:lxeS l l .l:.xeS c6 12 d4 .td6 13 .l:.el 'it'h4 14 g3 'it'h3 IS .te3 .tg4 16 'it'd3 :ae8 17 It:ld2 .l:.e6 (D) While 1 8 a4 is the main line, White has sev­ eral other moves which Black must be ready for. The first three of these are not very dangerous, while Line D is important for transpositional

65

reasons as well as for a general understanding of the Marshall Attack. Line E has been fash­ ionable lately because of a discovery by Kram­ nik, but I think Black can solve his problems with careful play. 65 A: 18 .tdl ? ! 66 B: 18 .tgS?! 66 C: 18 c4? ! 67 D: 18 .txdS E: 18 'it'n 69

A) 18 .tdl ? ! This exchanges a piece, but Black will no longer suffer from the pressure along the a2-g8 diagonal. 18 .txd l ! This i s much better than 1 8 . . . .ltf5 ? ! 1 9 'it'n .l:.fe8 20 'it'xh3 .txh3 2 1 .tb3 ! , Ljubojevic­ Nikolic, Amsterdam 1 988. By pinning Black's knight, White will have time to play It:ln and .td2. The text-move gains time. 19 .1:.axdl fS This is the thematic reaction to the exchange of light-squared bishops. Black clearly intends a quick . . . f4. 20 It:lf3 After 20 f4? .l:.fe8 2 1 lt:ln (2 1 .tf2 lt:lxf4 is winning for Black), 2 1 . . . .txf4 ! wins back the pawn and gives Black the initiative. 22 .tf2 .td6 23 ':xe6 ':xe6 24 :te l f4 25 ':xe6 'it'xe6 '+ Hellwing-Nunn, Lloyds Bank Masters, London 1 990. 20....:tg6! Not 20 . . . f4? just yet, because White has the strong response 2 1 lt:lg5 . •••

UNDERSTANDING THE MARSHALL A TTA CK

66

21 'if0 After 2 1 lLlg5 'ifh5 the knight is in trouble, while 2 1 .i.g5 ? ! f4 22 .i.h4? is refuted by 22 . . . fxg3 23 fxg3 .i.xg3 24 .i.xg3 l:.xg3+ 25 hxg3 'ii'x g3+ 26
B

If 30 'ifxc6? then 30 . . . g6 ! secures Black's back rank and gives him a winning attack. 30 'ifxg2+ 31 �e1 'ifgl+ 32 �e2 'ii'g2+ 33 �e1 'ifxg3+ 34 �e2 'ife3+ 35 �O (D) .••

24 l:.h6 24 . . . fxg3 25 hxg3 :xg3+ 26 fxg3 l:.xfl + 27 :xfl (27 �xfl 'iff3+) 27 . . . 'ifxe5 =+= is also good. The queen and knight are definitely stronger than White's rooks and bishop. 25 'ifg2 f3 26 'ii'h l 26 g4 l:tg6 ! . 26 l:te6 27 h4 l:txe5 28 'ifh3 Ljubojevic-Nunn, Amsterdam 1 98 8 . Now 28 . . .l:tfe8 is best, when White's bishop remains passive and Black's control of the e-file gives him a clear initiative. •••

•••

B) 18 .i.g5?! (D) This move is very rare. 18 l:.fe8 19 l:.xe6 l:.xe6 20 lLle4 After 20 a4 .i.e2 2 1 'ii'c 2 h6 22 .i.xd5 cxd5 23 .i.e3 Black has 23 . . . l:te4 ! ! intending . . . l:tg4 -+ as pointed out by Erenburg. 20 .i.f5 21 f3 .i.xg3! This shot gives Black at least a draw. The fol­ lowing moves are forced. 22 hxg3 .i.xe4 23 fxe4 l:tg6 24 'iff3 l:txg5 25 �f2 lLlf6 26 l:.gl 'ifh2+ 27 %lg2 'ifhl 28 e5 lLlg4+ 29
35 'ifc1+ Black takes the perpetual check. 35 ... lLlh2+ ! ? 36 �g2 'ife2+ 3 7 'iff2 'ifxf2+ 3 8 �xf2 lLlg4+ 39 �f3 h5 gives Black winning chances ac­ cording to Erenburg. 36
•••

•••

C) 18 c4? ! (D) This move would be good if it were not for Black's thematic reply. 18 .i.f4! This threatens 1 9 . . . l:th6 ! . The only question now is if White can equalize (he probably can­ not). •••

15 i.e3:

OTHER LINES

67

22 fxe6! 22 . . . ..txe6 ? ! is only good enough for equal­ ity after 23 gxf4 lZ)xal 24 :xal cxb5 intending . . . :d8 =. 23 gxf4 lZ)xal 24 :xal l:txf4 25 f3 cxb5 Black is much better because his bishop is strong on the open board and White has much weaker pawns. •••

B

D)

19 'ifn 1 9 cxd5 ? is sharp, but ultimately just bad. After 1 9 .. J:�h6 20 'ife4 'ifxh2+ 2 1 ..tt f l Black has a pleasant choice: a) 2 1 . . .f5 22 dxc6+ ..tth 8 23 'ifd5 and now the spectacular 23 . . . .:.e6 ! ! overloads the white queen, winning on the spot, since 24 gxf4 :e4 ! is a textbook example of line-blocking. b) 2 1 . . .oltxe3 22 l:txe3 :f6 23 f3 oltf5 24 'ife5 was Timman-l.Johannsson, Reykjavik 1 976. Now Black should play 24 ... ..th3 + ! (less good is 24 . . . 'ifxd2? ! 25 ..tt g l cxd5 26 ..txd5 'ii'x b2 27 :ae l , when White still has some chances be­ cause the f6-rook is rather clumsy) 25 ..tte l 'ifh l + 2 6 ..tt f2 'ii'g 2+ 2 7 ..tt e l 'ifg l + 2 8 ..tt e 2 'ifxa l +. 19 lZ)xe3 20 'ifxh3 oltxh3 21 cxb5 lZ)c2 Black could also play 2 1 . . .lZ)d5 22 :xe6 fxe6 23 gxf4, when 23 . . . axb5 gave him a slight advantage in Hornung-Rogers, Doeberl Cup, Canberra 1 987, while 23 . . .:xf4 and 23 . . . cxb5 are also good. 22 ..txe6 (D)

18 ..txd5 This is very committal and there is no real reason for White to make this capture so early. Black no longer has problems on the a2-g8 di­ agonal and now Black controls the e4-square. Even though this line should not be dangerous for Black, he should be aware of the line, be­ cause while some variations may transpose to lines considered earlier that are satisfactory for Black, there are some independent positions that he should be aware of. At the very least, the play that comes will give some insight into typ­ ical play in the Marshall Attack. 18 cxd5 (D) •..

•••

B

19 'ifn 1 9 a4 is also possible, and it will often transpose. After 1 9 . . . f5 White should play 20 'iffl 'ii'h 5, transposing to the main line, be­ cause neither 20 axb5 f4 nor 20 f4 .:tfe8 2 1 axb5 ..txf4 ! i s satisfactory for White. Black can also play 19 . . . bxa4, when White probably does best to play 20 'iffl , again transposing to the main lines. Other moves are likely to lead to trouble: a) 20 1::tx a4 ? ! f5 and here:

UNDERSTANDING THE MARSHALL A TTACK

68

a l ) 2 1 'iVfl ? f4 22 'iVxh3 .i.xh3 23 gxf4 l:g6+ ! (even stronger than Nunn's 23 . . . l::txf4 24 tLlfl l:[g4+ 25 tLlg3 h5) 24 'iiih l �g2+ 25 <;jtgl �e4+ 26 'iitf l .i.d3+ 27 ::te2 .i.xf4 is winning for Black. a2) 2 1 f4 .i.xf4 ! (2 1 . . .l::tfe8 22 l::txa6 g5 23 'iVfl ! is unclear) transposes to the note to White's 20th move in Section 2.2, which is good for Black. b) 20 c4 and then: b l ) 20 . . . .i.b4 21 cxd5 .i.xd2 22 .i.xd2 l::tb 6 is drawish according to Nunn. After 23 l::t xa4 l::t x b2 (intending . . . l::t x d2 and . . . .i.f3 ) White should play 24 lIa3 (not 24 �f4? l::tx f2 ! ) al­ though Black should be fine here. b2) 20 . . . f5 ! ? looks even stronger. 2 1 cxd5 (2 1 c5 f4 ! +) 2 1 . . .lIg6 22 f4 .i.f3 ! 23 lte2 (23 tLlxf3 lhg3+ ! ) 23 . . . :xg3+ 24 hxg3 'iVh l + 25 �f2 .i.xe2 26 l::txh l .i.xd3 gives Black a clearly better ending because of his bishop-pair. 19 'iVh5 (D)

the f6-rook is a bit awkward, while 24 'iii g 2 l::tf6 25 .i.d2 ;l;; looks even more accurate) 2 1 'iVf2 and now 2 1 .. .l::tfe8 gives Black good play, while 2 1 . . .g5 intending . . .f5 is another Nunn idea. After 20 a4 Black has a choice between: 68 D l : 20 f5? ! 69 D2: 20 bxa4 •••

.••

01) 20 f5? ! (D) •••

•••

w

20 a4 White has a few alternatives, but none should put off Black: a) 20 f4 l::tfe8 21 'iVf2 and now 2 1 . . .'iVf5 in­ tending . . . b4 was given by Nunn. This looks good, while 2 1 . . .f5 is possible as well. b) 20 'iVg2 l::tfe8 2 1 a4 b4 was another sug­ gestion by Nunn. This one has since been seen in practice: 22 c4 (22 ::tac l �e2 ! Nunn) 22 . . . dxc4 23 tLlxc4 �f3 24 'iVfl �d5 with more than enough for the pawn. c) 20 f3 .i.h3 ! (20 . . . l::tf6? ! 2 1 'iVe2 .i.xf3 22 tLlxf3 'iVxf3 23 'iVxf3 l::txf3 and now 24 .i.d2 was a little better for White in Anand-Mannion, Lloyds Bank Masters, London 1 986 because

This is probably the less promising of the two options. 21 f4 2 1 axb5 ! ? also looks better for White. 2 1 . . .f4 22 .i.xf4 .i.xf4 23 ':xe6 .i.xe6 24 gxf4 .i.h3 25 'iVd3 l::t xf4 (25 . . . .i.f5 26 'iVf3 .i.g4 27 'iVg3 ! avoids a repetition) 26 bxa6 (after 26 tLlfl .i.xfl 27 l::t x fl ':'f3 28 'it'd l axb5 Black probably has enough to draw) 26 . . . l::tg4+ 27 'iVg3 l::tx g3+ 28 hxg3 .i.c8 29 a7 .i.b7 30 a8'iV+ .i.xa8 3 1 l:[xa8+ 'iiif7 32 tLlfl ;l;; Zednik-Peetoom, corr. 2005 . 21...l::tfe8 2 1 . . .bxa4 22 'it'g2 is a position we covered in Line A of Section 2. 1 , which is also shaky for Black. 22 axb5! Other moves are less threatening: a) After 22 'iVf2, 22 . . . g5? ! 23 fxg5 f4 24 gxf4 is insufficient for Black, but 22 . . . b4 (Nunn), 22 . . . h6 and 22 . . . bxa4 all look better. b) 22 .i.f2 l::te 2 23 'iVg2 h6 ! 24 axb5 l::tx d2 25 bxa6 .i.b8 26 :xe8+ 'ili'xe8 27 h3 .i.h5 28 a7 .i.xa7 29 ':'xa7 'iWe2 gave Black good compen­ sation for the pawns in Kogan-Mitchell, corr. 1 963 .

15 i.e3:

OTHER LINES

22 Jbe3 23 :xe3 :xe3 24 bxa6 ii.b8 25 'ifb5 'ife8 26 'ifxd5+ 26 'ifxe8+ :xe8 27 a7? ii.xa7 28 l:ba7? :e l + 29 �f2 :e2+ is winning for Black. 26 � h8 (D) ••

•••

69

21 :xa4 f5 22 :xa6 22 f4? 'ii'e 8 is a familiar trick which we saw in note 'c' to White's 22nd move in Line A of Section 2. 1 . Also bad is 22 f3 i.h3 23 'iVf2 :fe8 intend­ ing . . .f4 =+=. Note that 24 f4? fails to 24 . . . :xe3 25 :xe3 'iVd l +. 22 f4 This is safe enough, but Black has a fighting alternative in 22 . . . i.h3 . White then has : a) 23 'ife2 'iVxe2 24 l:txe2 f4 25 :xd6 fxe3 (25 . . . :xd6 26 ii.xf4 :xf4 27 gxf4 :g6+ is only a draw) 26 :xe6 exd2 27 l:lxd2 ii.xe6 +. The bishop should be stronger than the pawns. b) 23 :xd6 ! ? :xd6 (worse is 23 . . . i.xfl 24 :xe6 intending ii.f4 ±) 24 'iVb5 is difficult to assess. 23 ii.xf4 :xe1 24 'fi'xe1 i.xf4 25 gxf4 i.e2! (D) •••

w

27 'ifa8 (not 27 ttJfl ? ii.f3 ! -+) and now: a) 27 . . . :e l +? 28 :xe l 'ifxe l + 29 ttJfl +- . b) 27 . . . h6? ! 28 a7 i.xa7 29 'ifxe8+ :xe8 30 h3 ! leaves Black in trouble after both 30 . . . :a8 3 1 �g2 and 30 . . . ii.xd4+ 3 1 cxd4 ii.xh3 32 d5 :e2 33 l:td l intending ttJc4. c) 27 . . . :e2 28 ttJfl �g8 (Nunn) gives Black reasonable drawing prospects. One idea is 29 . . . i.a7 to trade queens (after 29 b4, for exam­ ple), since 30 'ifxa7?? loses to 30 . . . i.f3 .

02) 20 bxa4 (D) .•.

26 :a7 26 :b6? loses to 26 . . . 'ifg4+ 27 �h l :e8 -+. 26 'ifg4+ 27 �hl :e8 28 'ifal White intends 29 :a8, so Black has nothing better than 28 . . . i.f3+ 29 ttJxf3 'iVxf3+ 30 �gl 'ifg4+ 1/2- 1/2 Lovakovic-Baron, COIT. 2002. •••

E) 18 'iVn This move just looks like a transpositional device, but White has a devilish idea in mind. 18 'fi'h5 (D) 19 f3 ! ? This shocking innovation o f Kramnik's put 1 8 'iffl on the map. Instead 1 9 i.xd5 cxd5 just transposes to Line D, while 1 9 ii.d l ? ii.xd l 20 •••

As is often the case, this is the stronger move.

70

UNDERSTANDING THE MARSHALL A TTACK

w

back the piece. Black has a choice between moving his bishop, pinning the f-pawn, or just taking the bishop on offer. We have: 70 E l : 19 �f5?! 70 E2: 19 1:If6?! 71 E3: 19 lLlxe3 ...

•••

•.•

1:Iaxd l f5 + gives Black a strong attack and is not worth considering. The only real alternative is 1 9 a4 �h3 ( 1 9 . . . f5 and 1 9 . . . bxa4 20 l:txa4 f5 2 1 f4 transpose to the main lines of Chapter 2) 20 �d l 'ii'f5 which is like the Spassky Variation, but White has left out axb5 . This is likely to transpose back to Chap­ ter 1 , but Nunn points out one line where the omission of the exchange of a-pawns is impor­ tant because a back-rank trick is not available to White. The line goes: 2 1 'ii'e2 �f4 ( ! Nunn, while Black can also consider 2 1 . . .c5 22 axb5 axb5 , transposing into Line D2 of Chapter 1 ) 22 'fj'f3 :fe8 23 lLlfl (the point is that the analo­ gous 23 gxf4 l:tg6+ 24 �h 1 �g4 25 'ii'e4 does not work because there is no weak back rank) 23 . . . �xfl 24 lhfl lLlxe3 25 fxe3 �xe3+ 26 �g2, when White has just a tiny edge because Black's pawns are more likely to be vulnerable. We now return to 1 9 f3 (D):

1 9 . . .�h3 has been suggested as well, but this has not seen any takers at grandmaster level. After 20 'ii'f2 �f5 White could continue with 2 1 a4 or maybe even 2 1 c4 ! ? However, 1 9 . . . l:txe3 100ks playable. After 20 fxg4 'ii'xg4 2 1 lhe3 lLlxe3 22 'iif2 lLld5 ex­ changing queens with 23 'iif3 'ii'xf3 24 lLlxf3 does not look like anything special for White after 24 . . . :e8.

EI) 19 �f5?! (D) •••

This is probably Black's worst choice. 20 lLle4! Returning the pawn to seize the initiative. 20 �xe4 21 fxe4 1:Ixe4 22 �dl After 22 �f2, 22 . . . :tfe8? ! 23 l:txe4 l:txe4 24 :te l 'iig 6 25 :txe4 'iixe4 26 'iie l 'iixe l + 27 �xe l lLle3 28 �d2 lLlc4 29 �c l left White a little better in Perelshteyn-Onishchuk, USA Ch, Stillwater 2007, but Black should be OK af­ ter the superior 22 . . . lLlf6. The text-move is a zwischenzug devised by Bacrot that causes Black difficulties. 22 'iig6 23 �c2 f5 24 �d2! (D) This is even stronger than grabbing the ex­ change with 24 �xe4 fxe4 25 'ii' h 3, because af­ ter 25 . . . h5 Black has counterplay. •••

••.

After 1 9 f3 ! ? the bishop on e3 is hanging, but if Black captures it, White can quickly win

15 i.e3:

OTHER LINES

71

i.xd4 26 �xh5 i.xf2+ 27 ..ti>xf2 lDf6 28 i.f3 ..ti>f8 = Yakovenko-Svidler, Moscow 2007 . 20 i. h3 Taking the pawn gives White an easy edge after 20 . . . �xf3 2 1 lDxf3 "ikxf3 (2 1 . . .l:txf3 ? 22 i.d l ) 22 'ii'xf3 :xf3 23 �c l ! ? (perhaps better than 23 i.d2 because White may want to play i.c2 in response to . . . lDb6, and 23 i.c 1 avoids . . . lDc4 coming with tempo) with the idea �g2. 21 'ii'C2 i.cs 22 a4 i.d3 23 axbS axbS 24 "ikg2! (D) •••

24 J:tg4 2S l:te2 l:tC6?! This is bad, but Black's position was uncom­ fortable in any case. 26 llae1 "ikC7 27 :eS+ �f8 2S l:[dS +- Bacrot-Yakovenko, European Clubs Cup, Kallithea 2008. ••

B

E2) 19 :C6?! (D) •••

Now 25 lDe4 is a threat. 24 i.bS 25 g4 Mikhalevski suggests that 25 l:tad 1 , again intending lDe4, is even better. 2S 'ii'g6 26 �xdS cxdS 27 lDb3 White was clearly better in Naiditsch-Sar­ gisian, European Clubs Cup, Kallithea 2008. •••

•••

E3) 19 lDxe3 (D) •••

This pinning move looks a bit artificial. 20 "ikg2 Other moves are less troubling to Black: a) 20 "ike2? ! i.xf3 2 1 lDxf3 l:txf3 22 i.xd5 (White cannot contemplate either 22 �d2? �xg3 or 22 �f2? lDf4 ! ) 22 . . . "ikxd5 23 �f2 l:tf6 24 b3 "ikf5 25 l:tad l h5 was already a little better for Black in Anand-Ivanchuk, Bilbao 2008 . b) 20 i.d l lle8 (20 . . . i.f5 ! ? may be better) 2 1 i.f4 :xe l 22 'ii'xe l :e6 23 i.e5 i.h3 24 "ikf2 (24 f4 ! ? - Mikhalevski) 24 . . . i.xe5 25 f4

UNDERSTANDING THE MARSHAU A TTA CK

72

Taking the piece is the most obvious move, of course. 20 'iVf2! This is the point of Kramnik's idea. White wins back the piece. 20 ttJd5 21 fxg4 'iVxg4 22 iff3 (D)

a) 23 . . . h6? ! 24 �xe6 fxe6 25 ttJd2 a5 26 a4 b4 27 c4 ttJc7 28 i..c 2 was better for White in Naiditsch-Ivanisevic, Subotica 2008. b) 23 ... .l:Ife8 24 �xe6 �xe6 25 a4 b4 26 c4 .l:Ie3 27 cxd5 (D).

.••

B B

White is confident that despite returning the extra pawn, any resulting opposite-bishop end­ ing will favour him. There is still pressure on the a2-g8 diagonal and White has the possibil­ ity of opening the queenside with a4. 22 'iVg5 Black prefers to keep queens on the board. Another way to do this is with 22 . . . 'iVg6 23 �xe6 (23 i.. xd5 cxd5 24 'iVxd5 i.. x g3 25 �xe6 fxe6 gives Black the initiative) 23 . . . 'i!Vxe6 24 'iVe4 �e8 25 ne 1 'iVxe4 26 ttJxe4 '>fi>f8 as given by Marin, although White still has some slight pressure. Heading into the ending immediately with 22 . . .li'xf3 23 ttJxf3 (D) leaves White with a nagging initiative, but it may be tolerable: •••

27 . . Jhb3 ! (better than 27 . . . �xf3 28 i..c4 cxd5 29 '>fi>g2 �e3 30 i.. xa6 ;1;) 28 ttJd2 �xb2 ! (not 28 . . . �d3 ? ! 29 ttJc4 ± Marin) 29 ttJc4 �c2 30 ttJxd6 b3 3 1 �e 1 (3 1 �b 1 b2 and the b-pawn will cost White his rook) 3 1 . . .h5 ! (not 3 1 . . .g6? 32 ttJf5 ! gxf5 33 d6 +-) 32 dxc6 b2 33 .l:Ie8+ '>fi>h7 34 �b8 nxc6 35 �xb2 l:txd6 with equality ! 23 nxe6 23 ttJe4 ? ! 'iVe7 ! is awkward for White. 23 fxe6 24 ttJe4 'iVg6 (D) •••

B

25 'iVd3! This is a slight improvement on Kramnik's original 25 'iVe2 since White avoids the ... i.. f4e3 manoeuvre in line 'b' below because his knight remains protected. After 25 'iVe2 Black has :

15 i.e3:

OTHER LINES

a) 2s . . . lbf4 ? ! 26 'ikc2 lbh3+ 27 �g2 'ikg4 (D).

2S il.xe6+ ! 'ikxe6 29 'ikb3 lbf4+ 30 gxf4 'ikxb3 3 1 axb3 il.xf4 32 l:txa6 ± was the stem game Kramnik-Aronian, Rapid match (game 1 ), Erevan 2007. b) 2S ... il.f4 26 �h 1 il.e3 27 il.c2 (27 il.xd5 ! ? exdS 2 S lbcs l:leS 2 9 .:te l as 3 0 �g2 may still give White chances of a slight edge according to Mikhalevski) 27 . . . 'ii'fS ! looks OK for Black, who may follow with . . . eS or even . . . 'ii'h 3. 25 .i.e7 Instead 2s . . . lbf4 ? ! 26 'ii'c 2 is Kramnik-Aro­ nian above. After 2S . . . hS 26 .i.c2 .i.f4 27 l:le 1 h4 2s lbf2 'ikxd3 29 lbxd3 .i.e3+? ! (29 . . . hxg3 30 hxg3 il.xg3 3 1 l:txe6 l:tf6 32 ':'xf6 gxf6 33 il.b3 as 34 �g2 il.d6 3S �f3 �g7 36 a4 is given by Mikh­ alevski - White still has a nagging edge but Black should hold) 30 'it>g2 .i.d2 3 1 l:te2 lbe3+ 32 'it>h3 hxg3 33 �xg3 (not 33 l:txd2? ':'f6 ! , which suddenly turns the tables - Black threat­ ens mate and 34
73

2S . . . a5 ! ? is another possibility. 26 l:tel 'fIr5 (D)

27 'ike2 e5 28 dxe5 'ikxe5 29 lbd2 'ii'xe2 30 l:txe2 Now, after 30 . . . l:teS 3 1 �fl �f8 32 l:te6 l:tdS 33 lbf3 ltd6 34 lbd4 White won material in Naiditsch-Onishchuk, European Clubs Cup, Kallithea 200S, but 30 . . . il.cS+ 3 1 'ii? g 2 should hold for Black. The immediate exchange of queens given in the note to Black's 22nd move looks tenable as well.

•••

Conclusions A s i n the Main Line o f the previous chapter, the Pawn Push variations are on very shaky ground theoretically. They are dangerous, however, be­ cause any slip by White could prove fatal. The Adams Variation is similar to the Spassky Vari­ ation of course, and is similarly holding up quite well. Of White's I Sth-move deviations, Kramnik's IS 'ii'f l combined with the shocking I S . . . 'ii' h S 1 9 f3 ! ? has given Black some head­ aches, but I think Black can equalize here too if he is careful.

4 M odern Va riation : 1 5 :e4

1 e4 eS 2 �f3 �c6 3 i.bS a6 4 i.a4 �f6 5 0-0 i.e7 6 l:r.e1 bS 7 i.b3 0-0 8 c3 d,5 9 exdS �xdS 10 �xeS �xeS 1 1 l:r.xeS c6 12 d4 i.d6 13 :e1 �h4 14 g3 �h3 1S l:r.e4 (D)

This unusual move has been rather contro­ versial for a long time. White moves his rook for a fourth time and puts it on an exposed square. The purpose of the move is to prevent both . . . i.g4 and . . . f5 while preparing :h4 to evict the black queen, after which White may even go on the attack with �h5 . IS gS ! Practice has shown that this sharp idea is Black's best move. The pawn cannot be taken because after 1 6 i.xg5 ? �f5 Black wins a piece. The manoeuvre l:[h4 is prevented, and Black is ready to attack the wandering rook with . . .f5 or . . . i.f5 . Other moves have been found lacking for Black. 1 5 . . . i.f5?? and 15 . . . f5 ?? both lose the queen to 16 :h4. Some other tries: a) 15 . . . �f6 ? ! 16 :h4 �d7 ( 1 6 . . . �f5 1 7 i.c2 ! ) 1 7 i.g5 ±. b) 15 ... �f5 ? ! 16 �d2 �g6 17 l:r.e l i.g4 1 8 f3 i.e6 1 9 �e4 ± Ricard-Wallner, COIT. 2005 . c) 1 5 . . . i.d7 ? ! 1 6 c4 ! ? ( 1 6 l:r.h4 �e6 1 7 �d2 is also possible) 1 6 . . . bxc4 1 7 i.xc4 l:r.ae8 1 8 �d2 �f6 1 9 l:[h4 �f5 20 �f3 �e4 2 1 i.d3 h6 22 i.e3 ± Nicevski-Plachetka, Naleczow 1 979. •••

d) 15 ... �d7 ? ! 16 �d2 i.b7 (after 16 ... f5 1 7 :e l f4 1 8 'fih5 ! i t i s White who i s attacking, while 16 . . .�f6 1 7 :h4 i.b7 1 8 �f3 :ae8 1 9 i.g5 ± was Tal-Krogius, USSR Ch, Leningrad 1 97 1 ) 1 7 .:te l c5 1 8 dxc5 ( 1 8 �e4 c4 1 9 i.c2 i.e7 20 �g5 i.xg5 2 1 i.xg5 l:r.ae8 gave Black some counterplay in Nunn-I.Sokolov, Wijk aan Zee 1 99 1 ) 1 8 . . . i.xc5 1 9 �e4 i.e7 20 i.g5 l:[ad8 2 1 i.xe7 �xe7 was Lalic-Turner, Spon­ don 2000, and now Lalic gives 22 'fih5 �f6 23 'fic5 ±. e) 1 5 ... i.b7 ? ! 1 6 l:[h4 'fie6 17 �d2 f5 (or 17 . . .l:[fe8 1 8 �e4 i.e7 1 9 �c5 ! i.xc5 20 dxc5 �e7 2 1 'fid3 g6 22 �d2 h5 23 i.xd5 cxd5 24 �d4 ± Nunn-Hebden, Lloyds B ank Masters, London 1 990) 1 8 'fih5 h6 19 �f3 i.e7 20 lIh3 c5 2 1 dxc5 i.xc5 22 i.f4 l:[ae8 23 :d l :e7 24 i.g5 l:[d7 25 l:[e l �b6 26 l:[e2
w

With .:th4 prevented, Black is ready to attack the rook with . . . f5 or, more often, . . . i.f5 . Hav­ ing spent so much time moving the rook al­ ready, often instead of retreating White will leave the rook on e4, offering an exchange sac­ rifice. Positionally speaking, this is justified by

MODERN VARIA TION:

15 l:e4

75

the weakening of Black ' s kingside, especially on the light squares. It is very important that Black considers this when taking the rook - it is more important to be able to fight for the initia­ tive, and this may or may not involve taking on e4. We examine: A : 16 'it'e2 75 77 B: 16 'iVf3 C: 16 'iVe1 80 D: 16 'iVn 81 Other moves are possible but they are much less common: a) 1 6 a4? ! is slow. 1 6 . . . f5 1 7 R.xd5+ ( 1 7 :e3 �h8 1 8 R.xd5 cxd5 i s the same) 1 7 . . . cxd5 1 8 .l:.e3 �h8 1 9 'iVfl (or 1 9 axb5 f4 20 gxf4 'iVh4 +) 1 9 . . . 'iVxfl + ( 1 9 . . . 'iVh5 20 'iVe2 'iVg6 is also possible, but the text-move looks stronger) 20 �xfl f4 gives Black a powerful initiative. b) 1 6 tLld2 f5 ( 1 6 . . . R.f5 transposes to Line B after 1 7 'iVf3 or 1 7 R.xd5 cxd5 1 8 .l:.e3 .l:.ae8 1 9 'iVf3 ) . Now 1 7 R.xd5+ cxd5 does not help White, while 17 .l:.e5 R.xe5 1 8 dxe5 f4 also fails to stem Black's initiative, so White should re­ treat the rook: b 1 ) 1 7 l:.e2 f4 1 8 'iVfl 'iVh5 gives Black good chances. b2) 17 l:.e3 f4 ( 1 7 . . . �h8 is also possible) 1 8 gxf4 'iVh6 also gives Black a strong attack. b3) 1 7 .l:.e 1 f4 1 8 'iVe2 R.g4 1 9 'iVfl and now 1 9 . . . fxg3 looks good but it is not so clear: 20 hxg3 R.xg3 2 1 'iVxh3 R.xf2+ 22 �h2 R.xh3 23 .l:.e5 g4 24 .l:.xd5 cxdS 25 R.xd5+ �g7 26 R.xa8 .l:.xa8 27 tLle4 is difficult to assess. Black could also try 19 . . . 'iVh5 !? with good attacking chances.

A) 16 'iVe2 (D) 16 f5! This is the soundest move and should lead to a draw. Other moves invite White to make a promising exchange sacrifice: a) 16 . . . tLlf6 ? ! 17 tLld2 (not 17 .i.xg5 ? .i.g4) 17 . . ..i.f5 1 8 f3 c5 (worse is 1 8 . . . tLlxe4 ? ! 1 9 tLlxe4 R.xe4 { 1 9 . . . .i.e7? 20 tLlf2 wins Black's queen } 20 fxe4 .i.e7 21 R.e3 ± Sax-AtaIik, Bled 2000) 19 'iVf2 c4 ( 1 9 . . . cxd4 20 cxd4 l:.ac8 2 1 l:te 1 R.b4 2 2 tLle4 tLlxe4 2 3 l:.xe4 R.xe4 24 fxe4 'iVg4 25 e5 gives White good compensation .•.

according to Ponomariov) 20 .i.c2 h6 21 b3 cxb3 ? ! (2 1 . . .l::t fc8 is a better try) 22 axb3 .l:.fc8 23 R.b2 R.b4?! (23 . . . .i.g6 is better but after 24 R.d3 tLlxe4 25 tLlxe4 R.e7 26 c4 White has ex­ cellent compensation for the exchange) 24 l:.e5 R.xc2 25 cxb4 R.g6 was Ponomariov-Adams, Linares 2002. Now White can play 26 d5 ! l:.c2 27 'iVd4 with a winning advantage. b) 1 6 . . . R.f5 ! ? and here: bI ) 1 7 tLld2 and now rather than transpos­ ing to line 'a' by 17 . . . tLlf6 ? ! , Black could try 1 7 . . . R.xe4 1 8 tLlxe4 'iVe6 1 9 'iVf3 f6. b2) 1 7 R.xd5 cxd5 1 8 .l:.e3 R.g4 1 9 'iVfl 'iVh5 20 tLld2 l:.ae8 2 1 ':xe8 .l:.xe8 22 f3 R.h3 23 'iVf2 f5 24 tLlfl f4 25 .i.d2 'iVg6 + Czaeczine­ Jackova, Women's Bundesliga 2003/4. b3) 17 f3 and now: b3 I ) 1 7 . . . tLlf6? ! 1 8 R.xg5 ? (White should instead prefer 1 8 tLld2 ! , transposing to line ' a ' ) 1 8 . . . R.xe4 1 9 R.xf6? (White has t o try 1 9 fxe4 tLlxe4 ! 20 'iVxe4 .l:.ae8 2 1 tLld2 or 1 9 tLld2) 1 9 . . . R.xbl ! 20 .l:.xbl 'iVf5 -+ Novitsky-Azarov, Belarus Ch, Minsk 2004. b32) 1 7 . . . R.xe4 seems to force a draw: 1 8 fxe4 :ae8 1 9 tLld2 f5 20 e5 f4 2 1 tLle4 R.xe5 22 dxe5 :xe5 23 tLlf6+ l:.xf6 24 'iVxe5 fxg3 25 R.xd5+ cxd5 26 'ii'x g5+ �f7 27 'iVxd5+ �e8 28 'iVa8+ rJi;f7 29 'iVb7+ ..t>e8 1/2- 1/2 Bologan­ Onishchuk, Poikovsky 2003 . 17 R.xd5+ White must make this exchange. Instead 1 7 .l:.e6? .i.xe6 ! 1 8 'it'xe6+ �h8 1 9 'it'xd6 .l:.ae8 gives Black a winning attack: 20 R.d2 f4 2 1 R.xd5 cxd5 2 2 f3 g4 0- 1 Smirin-Grishchuk, Eu­ ropean Clubs Cup, Panormo 200 1 . 1 7 cxd5 1 8 l:.e6 (D) 18 f4! •••

••.

76

UNDERSTANDING THE MARSHAll A TTACK

b) 20 'ife5 ? ':'aeS 21 'ifxg5+ �hS -+. c) 20 'ii'd 3 ? :aeS 21 tDd2 (2 1 i.d2 i.e2) 2 l . . .l:le l + 22 tDfl i.f3 -+. 20 'lixfi +! A piece down, Black exchanges queens. 21 �xfi l:tae8 (D) •••

w

This discovery of Anand's completely solves Black's problems. Other moves are risky, but may be playable: a) I S . . . i.c7 19 l:lc6 ( 1 9 i.xg5 is also un­ clear after 19 . . . i.xe6 20 'ii'xe6+ l:tf7 21 tDd2 'ifh5 or 19 . . . f4 20 l:th6 'ii'f5 21 i.h4 i.d7) 19 . . . f4 20 l:hc7 f3 21 'iffl 'ifxfl + 22 �xfl .i.h3+ 23 �e l (23 �g l l:taeS 24 i.d2 l:te2 in­ tending . . . l:tfeS) 23 . . . l:taeS+ 24 �d l (24 i.e3 ? l:he3+ 2 5 fxe3 f2+) 24 . . . l:te2 2 5 i.xg5 l:tfeS was not so clear, despite White's big material advantage, in Mezera-Schulze, corr. 2003 . b) I S . . . i.xe6 ! ? is supposed to be bad, but some correspondence games have shown that matters are not so clear. 1 9 'ifxe6+ �g7 and here: b l ) 20 'ifxd6 l:taeS 2 1 'ifd7+ 'iith S 22 .i.d2 l:te2 (with compensation - Ponomariov) 23 c4 l:txf2 24 �xf2 'ii'x h2+ 25 �e l 'ii'h l + 26 �e2 bxc4 should lead to a perpetual check. b2) 20 i.xg5 'ii' h 5 ! . Black has won two correspondence games from this position, so White will need to find improvements : b2 l ) 2 1 'ii'd 7+ l:tf7 22 i.f6+ �g6 23 'ii'xd6 'ifd l + 24 ..t>g2 l:txf6 25 'ii'xd5 l:tafS 26 'ii'f3 ? 'ii'c 2 and Black is much better, V.Vaiser-H.Koch, corr. 2005 . b22) 2 1 'ifxd6 'ifxg5 22 'ife5+ �g6 23 f4 'ifdS 24 'ife2 l:leS 25 'iff2 l:ta7 26 tDd2 l:tae7 27 �fl 'lics 2S tDf3 'ii'c4+ 29 ..t>g l l:te2 30 'iffl 'ii'd 3 3 1 tDe5+ ':'Sxe5 32 fxe5 'ife3+ 33 'iith 1 ':'f2 34 'ifh3 ':'xb2 0- 1 Koch-Efendiev, corr. 2002. 19 l:txd6 i.g4 20 'lin Other moves are just bad: a) 20 f3 ? l:taeS 2 1 'ii'f l 'ii'x fl + 22 'ifilxfl fxg3 -+.

Despite the extra material, White must be very careful. 22 i.d2 22 :h6? runs into 22 . . . :f6 ! ' 22 i.h3+ 22 . . . ':'e2? fails to 23 gxf4 gxf4 24 l:txd5 +-, and 22 . . . fxg3 23 hxg3 :e2? (for 23 . . . .i.h3 + ! 24 'iit g l see the main line) to 24 f3 ! . 23 �gl fxg3 24 hxg3 l:te2 25 .i.e3 (D) 25 f4 ':'feS gives Black at least a draw, while 25 l:th6 ':'exf2 ! 26 tDa3 :g2+ 27 �h l ':'xd2 2S l:txh3 l::tff2 29 b4 �g7 ! (Ponomariov) leaves White completely tied up. •••

B

25 l:txe3! 26 fxe3 :n + 27 �h2 g4! 28 l:txd5 •••

MODERN VARIA TION:

15 :e4

77

Instead 28 tDd2 l:.xal is certainly not worse for Black. After 28 ':'xd5 a draw was agreed in Ponomariov-Anand, Linares 2002.

B) 16 'ii'f3 (D)

This has been White's most popular move historically, but lately White has been turning to Line C to look for an advantage. 16 .lt fS White must decide how to offer the exchange: 77 B l : 17 .ltxdS 78 B2: 17 .lt c2 •••

Instead 17 l:.e I? loses to 17 ... l:.ae8 18 .i.d2 tDf4 ! -+, while 1 7 tDd2 is relatively unex­ plored; e.g., 1 7 . . . l:.ae8 1 8 .i.xd5 ( 1 8 .ltc2 tDf4 ! ) 1 8 . . . cxd5 1 9 l:te3 l:txe3 ( 1 9 . . .h 6 and 1 9 . . . .lte6 could also be tried) 20 'ii'xe3 (20 fxe3 l:.e8 2 1 a4 i s interesting) 20 . . . f6 2 1 tDfl .i.g6 2 2 f3 l:.e8 23 'ii'f2 .i.d3 24 .i.e3 .ltxfl 25 ':'xfl .i.xg3 26 hxg3 ':'xe3 27 'ii'xe3 '12- '12 Lima-Grishchuk, FIDE World Cup, Khanty-Mansiisk 2007.

a) 20 a4 b4 2 1 c4 dxc4 22 tDxc4 .i.d3 ! 23 ':'xe8+ l:txe8 24 tDe3 .i.f4 25 'ii'g 2 'ii' h 5 26 gxf4 'ii'd l +! 27 'ii'f l (27 tDxd l l:.e l + 28 'ii'f l ':'xfl + 29 'iti>g2 l:.xd l 30 fxg5 b3 is an inferior version of the game for White) 27 . . . 'ii'x fl + 28 tDxfl l:te l 29 fxg5 b3 30 h4 .ltxfl 3 1 f3 .ltc4+ 32 'ifi>f2 l:tfl + 33 'ifi>g2 l:td l was Timman-Short, Tilburg 1 99 1 . Clearly only Black can play for a win. b) 20 b3 'ifi>g7 ! 21 .ltb2 .ltg4 22 'ii'g 2 (after 22 'ii'xd5 Black can draw by 22 . . . .ltf4 ! 23 'ii'x d8 ! .ltxe3 24 'ii'xe8 .ltxf2+) 22 . . . 'ii'xg2+ 23 'ifi>xg2 l:txe3 24 fxe3 l:te8 25 l:te l .ltf5 26 'ifi>f2 .i.d3 27 l:td l '12- '12 Leko-Adams, Wijk aan Zee 200 1 . This is a typical endgame where White's extra pawn is all but worthless. Now back to 18 ... .i.e4 (D), which forces White to sacrifice.

B1) 17 .ltxdS White holds on to his rook for the time being. This line usually leads to an endgame where both sides have chances. 17 cxdS 18 l:.e3 (D) 18 .lte4 This move drastically changes the nature of the position. Black has a decent alternative in 1 8 . . . l:.ad8 1 9 tDd2 l:.fe8, and now: •••

•••

19 l:.xe4! dxe4 20 'fi'f6 This double attack on d6 and g5 ensures that White will get a second pawn for the ex­ change.

UNDERSTANDING THE MARSHALL A TTACK

78

20 Ji'g4 This counterattacking move just about forces the queens off. Worse is 20 . . .�f4? 2 1 �xf4 gxf4 22 'ii'xf4 f5 23 liJd2 intending liJfl -e3 with a big advantage for White. 21 'ii'xg5+ Taking the bishop is bad: 2 1 'ii'x d6? 'ilkd 1 + 22 �g2 'ii'xc 1 +. 21 liJd2 :ae8 22 liJfl (22 'ii'xd6 e3 23 fxe3 l:.xe3 looks rather dangerous for White) 22 . . . �e7 (22 . . . l:te6 23 'ilkxg5+ 'ii'x g5 24 �xg5 :tg6 25 �d2 f5 is similar to the main continuation) 23 'ii'xa6 f5 24 'ii'xb5 f4 was agreed drawn in this unclear position in Timman-Ivanchuk, Linares 1 99 1 . 2 1 'ii'xg5 22 �xg5 f5 23 liJd2 (D) .•

82) 17 .i.c2 (D)

B

•••

White prefers to hold on to his strong light­ squared bishop, but this is risky. 17 ... �xe4 Taking the rook immediately is the most forc­ ing continuation. Instead 1 7 . . . liJf4? ! 1 8 .i.xf4 gxf4 1 9 liJd2 is bad for Black, as is 1 7 . . . l:tae8? ! 1 8 .1xg5 .1xe4 1 9 .1xe4 l:te6 20 liJd2 ±. However, Harding's idea 17 ... .1f4 ! ? (D) is worth considering.

Many games have been contested in this im­ portant ending. White has two pawns for the exchange, so materially speaking he is doing quite well, especially if he can stabilize the po­ sition. Black, on the other hand, must open lines for his rooks. 23 :tae8 Black intends to use his kingside pawns. In­ stead 23 . . . �f7 24 �fl (24 a4, 24 f3 and 24 h4 have all been tried as well) 24 . . . �e6 25 �e2 �d5 26 �e3 l:.ac8 27 .i.f4 ! .i.xf4+ 28 �xf4 was Timmerman-Vitomskis, corr. 1 996. White intends to play l:.c l , b3, 'ite3 and c4+, mobiliz­ ing his pawns. 24 a4 f4 25 axb5 After 25 ]:te l e3 26 fxe3 fxg3 27 hxg3 �xg3 Black's active rooks give him good play. 25 axb5 26 gxf4 �xf4 27 �xf4 :txf4 28 :el �f7 = Timmerman-Tamowiecki, corr. 2000. •••

•••

White has: a) 1 8 liJd2 ? ! and here: a 1 ) 1 8 . . . liJf6 1 9 :te l ! .1xc2 ( 1 9 . . . :tae8 20 liJe4 ! liJxe4 2 1 .1xf4 .1g4 22 'ii'g 2 ± is given by Nunn) 20 gxf4 g4 was Wang Zili-Ye Jiang­ chuan, Chinese Ch 1 988. Now 2 1 'ii'x c6 is criti­ cal, because 2 1 . . .:tae8 can be met by 22 :te5 ! . a2) 1 8 . . . l:tae8 1 9 ':xe8 ? ! (White should try 1 9 'it'd l ! ?) 1 9 . . . ':xe8 20 liJe4 g4 2 1 'it'h 1 .1xc 1

MODERN VARIATION:

(2 1 . . .i.xg3 ! ? 22 fxg3 i.xe4 23 i.xe4 tDf6 is another idea) 22 ':'xc 1 'ii'h 6 23 :fl (23 :e 1 "'d2 ! -+) 23 . . ...g6 ! 24 .l:.e 1 �g7 25 h4 tDf6 26 h5 "'g5 ! 27 f4 gxf3 28 "xf3 i.xe4 29 i.xe4 tDxe4 30 .l:.xe4 "c 1 + 3 1 �f2 "xb2+ 32 "'e2 "'xe2+ 33 ':'xe2 .l:.xe2+ 34 �xe2 �h6 35 �f3 �g5 ! -+ Sacerdotali-Hansson, corr. 2004. b) 1 8 i.xf4 is a better try for White. 1 8 . . . gxf4 ( 1 8 . . . tDxf4 1 9 tDd2 ±) 1 9 tDd2 i.xe4 20 i.xe4 ! fxg3 (20 . . . l:tad8 2 1 i.xd5 .l:.xd5 22 "xf4 is better for White) 21 hxg3 "'h6 has occurred several times in correspondence play. 22 tDb3 .l:.ad8 (22 . . . tDe7 23 �g2 f5 24 i.c2 ±) 23 tDc5 .l:.d6 24 tDxa6 (24 tDd7 :e8 25 tDe5 looks like an improvement) 24 . . .:f6 25 "'d3 "'g5 26 i.g2 l:le8 27 tDc5 :xf2 ! 28 �xf2 :e3 29 "c2 l:hg3 30 .l:.g l "h4 3 1 �e2 "'h2 gave Black an attack in Sakai-Schulze, corr. 2003 . 18 i.xe4 "e6 For 1 8 . . . .l:.ae8 ? ! , see 1 7 . . . .l:.ae8? ! above. 19 i.xg5 f5 (D)

20 i.d3 20 i.xd5 ? ! cxd5 2 1 tDd2 f4 22 i.xf4 i.xf4 23 gxf4 ':'a7 ! (this is a good idea, well known from the Sveshnikov Sicilian ! ) 24 �h 1 .l:.e7 25 :tg l + �h8 26 tDfl "'e4 27 �g2 "'xf3+ 28 �xf3 :e l 29 :h l .l:.e4 =+= Topalov-Adams, Sara­ jevo 2000. 20 f4 This is a tough decision. Black can also throw in 20 . . . h6. After 2 1 i.d2 Black has: a) 2 1 . . ..l:.a7 22 tDa3 .l:.g7 (22 . . . f4 23 :te l "f6 24 g4 ;!;) 23 .l:.e 1 "f6 24 c4 ! gives White the initiative. b) 2 1 . . .f4 22 "h5 .l:.a7 23 tDa3 .l:.g7 24 .l:.e l 'iff6 25 .l:.fl ! ? could be tried. •..

15 l:te4

79

21 'ife4 (D)

White offers an exchange of queens while stepping up the pressure on the light squares. Both sides must tread carefully. 21 .....d7! This move has enjoyed some success in cor­ respondence games. Other moves look worse, but 'c' is of interest: a) 2 1 . . . ..xe4?! 22 i.xe4 :ae8 23 tDd2 .l:.e6 24 �g2 gives White excellent compensation for the exchange in the ending. b) 2 1 . . . ...h3 ? ! 22 i.h4 ! .l:.a7 23 tDd2 (the immediate 23 i.fl is also possible) 23 . . . .l:.g7 24 i.fl ! "'d7 (24 . . . "xh4 25 "e6+ �h8 26 "xd6 also favours White) 25 .l:.e 1 i.e7 26 i.xe7 .l:.xe7 27 'ifbl fxg3 28 hxg3 .l:.fe8 29 .l:.xe7 .l:.xe7 30 tDf3 and with his knight settling on e5, White had good compensation in Kr.Georgiev-Lukacs, Baile Herculane 1 982. c) 2 1 . . .'ilff7 ! ? 22 tDd2 fxg3 ! (after 22 ... .l:.ae8 23 "f3 "g7 24 "h5 White keeps up the pres­ sure) 23 fxg3 .l:.ae8 24 'ilfg2 tDe3 and Black has no problems according to Pavlovic, who con­ tinues 25 i.xe3 .l:.xe3 26 .l:.fl "e6 27 i.e4 .l:.xfl + 28 �xfl i.e7 ! , when Black can fight for the initiative. 22 tDd2 .l:.ae8 23 "hI 23 "'g2 looks more natural, but it runs into 23 . . ....g4 24 i.h4 f3 25 "'fl tDf4 with a strong attack. 23 ..... g7 24 tDe4 i.e7 (D) Black's well-centralized pieces give him good counterplay. 25 i.h4? ! f3 Now the white queen is entombed and White is running out of moves.

UNDERSTANDING THE MARSHALL A TTACK

80

Instead 1 6 . . . lLlf6 1 7 tZXl2 'ii'h5 1 8 'ii'd l 'ii'xdl + 1 9 .i.xd l lLlxe4 20 lLlxe4 .i.e7 2 1 .i.xg5 .i.xg5 22 lLlxg5 ;t was Motylev-Tkachev, Kazan 2005 .

Cl) 16...f5 The move that 1 6 'ii'e l is supposed to prevent. 17 �xd5+ exd5 18 ':e6 (D)

26 �n ? ! b4 27 ':dl bxe3 28 lLlxc3 lLlxe3 29 bxe3 �a5 Now 30 ':c l 'it'h6 ! 3 1 ':c2 �xc3 32 ':xc3 'iVd2 is lights out, so White resigned in Hallen­ gren-Lakatos, COTTo 200 1 .

C) 16 'iVel (D) 18...� e7 ! ? This is Pavlovic 's idea. As mentioned above, 1 8 . . . �xe6 1 9 'iVxe6+ is covered in note 'b' to Black's 1 8th move in Line A. If this line works for Black, then both 1 6 'iVe2 and 1 6 'iVe l are harmless. 19 �xg5 1 9 l:tc6 is met by 1 9 . . . ':a7 . 19 'ii' h5 ! 1 9 . . . f4 20 l:f6 �g4 2 1 l:xf8+ :xf8 22 lLld2 'iVh5 23 .i.h4 :e8 24 'ii'f l l:e2 25 lLlb3 allows White to claim some advantage according to Pavlovic, although I think this looks playable for Black as well. 20 l:e6 20 �e7? �xe6 2 1 'it'xe6+ 'it'f7 22 'iVxf7+ l:xf7 23 �c5 :e8 24 �fl f4 25 g4 f3 26 h3 l:e2 (Pavlovic) looks winning for Black. 20 l:ta7 21 �f4 White must avoid 2 1 'it'c l ? ! f4 22 �h6? �h3 ! -+ and 2 1 �h6 l:te8 22 'iVc l ? ! f4 23 ':f6? �d8 24 ':d6 ':f7 ! (improving over PavloviC's analysis) threatening 25 ... fxg3 26 fxg3 1If2 ! with a mating attack. Black wins after both 25 �xf4 .i.c7 and 25 :xd8 l:xd8 26 .i.xf4 �h3 . 21 ...�xf4 22 ':xe8 l:xe8 23 'it'e6+ �g7 24 'iVxe8 (D) •••

This modem finesse is designed to improve on Line A above. The main point is that 1 6 . . . f5 can be met by 1 7 �xd5+ cxd5 1 8 l:e6, when 1 8 . . . f4? loses to 1 9 l:xd6 because 1 9 . . . �g4 does not come with tempo. However, 1 8 . . . �xe6 1 9 'ii'xe6+ r3;g7 may well be playable, and was covered by transposition under Line A in note 'b' to Black's 1 8th move. Black can also play a little more calmly. Because e l is not the ideal square for White's queen, Black should be able to hold the balance. We examine: C l : 16 f5 80 81 C2: 16 .i.f5 •••

•••

•••

MODERN VARIATION:

15 :e4

81

22...'ifg4 Also possible is 22 . . . fS 23 eS f4 - Anand. 23 .te2 :e7 24 .td2 .:tfeS Now instead of 2S eS ? ! f6 ! , when Black seized the initiative in Anand-Bacrot, Sofia 2006, Anand gives 2S 'ifd l 'ifxd l 26 .:txd l l:xe4 27 .txe4 l:xe4 28 c:J;f2 = .

B

0) 16 'ifn (D)

It looks like the tactics have worked out for White, but Black has a nice resource. 24. ..tc1 ! 2S lLlaJ .txb2 26 'ifc5 :f7 27 .:tel f4 28 'ii'b4 .txaJ 29 'ifxaJ 'ifh3 30 gxf4 'ifg4+ with a draw - Pavlovic. .

e2) 16....tfS (D)

This has been White's most popular move of late. Black can exchange queens and hope to carry his initiative into the endgame, or he can retreat and play for an attack. D 1 : 16...'ifxfi+ 81 83 D2: 16...'ifh6 ! ? D3: 16...'ifhS 8S Also, there is the typical prophylactic move 1 6 . . . h6 ! ?, which is untested but may be worth a try. 17 lLld2 h6 17 . . .lLlf6 1 8 f3 cS ! ? 1 9 'ife3 h6 20 'ii'f2 cxd4 2 1 cxd4 .:tad8 (taking on e4 is another idea) 22 l:e2 .i.d3 23 l:e3 'iffS 24 lLle4 lLlxe4 2S fxe4 .i.xe4 26 'ifxfS .i.xfS 27 .td2 c:J;g7 112- 112 Schek­ achev-Gustafsson, Austrian Team Ch 2004/S . IS f3 �g7 19 a4 .txe4 20 fxe4 .:taeS 21 axbS axbS 22 lLln After 22 'ii'f 2? ! fS ! 23 eS (23 exdS ? ! f4 gives Black a decisive attack) 23 . . . f4 24 'ifg2 'ifxg2+ 2S �xg2 lLle3+ 26 �gl fxg3 27 hxg3 .txeS 28 dxeS .:txeS Black has an attack according to Anand.

01) 16 ... 'ifxn+ 17 �xn (D) 17 ... .tfS 1 7 . . . h6 1 8 lLld2 .tfS transposes to line 'b' in the next note, while 17 . . .fS 1 8 .:te l f4 1 9 lLld2 is better for White, because after 19 . . . fxg3 20 hxg3 .i.xg3 21 lLle4 ± he wins back the pawn with the initiative. ls lLld2 Even in the endgame, White should offer the exchange. The passive 1 8 l:e l .:tae8 1 9 .te3 gives Black a pleasant choice between

UNDERSTANDING THE MARSHAll A TTA CK

82

compensation' in Informator, but it doesn't look like enough. After 27 �f2 l:te8 28 .ib3 c4 29 .id l as 30 .ie2 a4 3 1 a3 .if8 32 lLlh4 .ih3 33 .if3 .ih6 34 lLlg2 i.xg2 3S i.xg2 ± White won handily in Motylev-Sargisian, Aeroflot Open, Moscow 2006. b) l 8 . . . h6 (D) is a typical move.

w

1 9 . . J:te6 intending . . J:t£e8 and 1 9 . . . .ih3+ in­ tending . . . fS . White can also play 1 8 f3 h6, and then: a) 19 lLld2 .ixe4 20 fxe4 lLlc7 (intending . . . cS, . . . cxd4 and . . . lLle6) 2 1 �g2 cS 22 eS .ie7 23 lLle4 cxd4 24 cxd4 as 2S .i.e3 a4 26 .i.d l lLldS 27 .if2 gives White enough compensa­ tion for the exchange but no more than that, Anand-Svidler, FIDE World Ch, San Luis 200S. b) 1 9 .:tel :fe8 20 .ixdS cxdS 2 1 lUe8+ ':xe8 22 �f2 as 23 a3 .id3 is a typical Marshall endgame. The players agreed to split the point here in Peng Xiaomin-Grishchuk, China-Russia match, Shanghai 200 1 . We now return to l 8 lLld2 (D) :

IS J:tadS By protecting the d6-bishop, Black is ready to play 1 9 . . . .i.xe4 20 lLlxe4 h6. Black has also tried: a) l 8 . . . :ae8 1 9 f3 lLlf6 20 lhe8 lhe8 2 1 �f2 g4 22 .id l gxf3 2 3 lLlxf3 lLlg4+ 24 �gl cS 2S .id2 lLle3 26 .ixe3 %he3 is given as 'with ••

Black protects the gS-pawn to lessen the ef­ fect of a knight coming to e4 after Black grabs the exchange. After 1 9 :e l ( 1 9 f3 is Anand­ S vidler in the note to White's 1 8th move above) 1 9 . . . l:tae8 20 lLlf3 there is: bl) 20 ... i.h3+ 21 �gl i.g4 22 lLleS f6 23 lLld3 i.fS 24 .:td l i.g4 2S :d2 Wg7 26 a4 lLlb6 27 axbS axbS 28 i.d l .ifS 29 i.f3 :a8 30 l:txa8 l:txa8 3 1 l:td l ± Volokitin-Ponomariov, Foros 2006. Black does not have anything for the pawn. b2) 20 . . . g4 2 1 i.xh6 (or 2 1 lLlg l .id3+ 22 lLle2 l:te6 23 i.xdS cxdS 24 .ie3 l:tfe8 2S l:ted 1 i.e4 with compensation, Motylev-Onishchuk, Biel 2007) 2 l . . .gxf3 22 .ixf8 l:txf8 is a rather unclear endgame. 23 i.d l .ig4 (23 . . . .ih3+ 24 �g l .i.g2 2S a4 �h8 26 l:te4 b4 27 c4 lLlf4 28 gxf4 :g8 29 cS 1/2- 1/2 Volokitin-Sargisian, Bun­ desliga 200S/6) 24 a4 :d8 2S axbS axbS 26 �gl �g7 27 h3 .i.xh3 28 .i.xf3 i.e6 is proba­ bly a little better for White because of his slight material advantage, but Black drew without dif­ ficulty in Bacrot-Anand, Mainz rapid 2007. We return to l 8 . . . l:tad8 (D) : 19 f3 1 9 l:te l l:tde8 20 lLlf3 g4 is even better for Black than the lines above because after 2 1 .i.h6 gxf3 2 2 .ixf8 l:txf8 (which led to a draw in Kariakin-Grishchuk, Turin Olympiad 2006),

15 :e4

MODERN VARIA TION:

83

21. h6 After 2 1 . . .':feS 22 axb5 (this looks prema­ ture; instead 22 e5 i.e7 23 liJe4 h6 24 h4 is sim­ ilar to the main line) 22 . . . axb5 23 e5 i.e7 24 liJe4 ':as 25 ':bl (not pretty, but White still seems to have good chances) 25 . . . h6 26 h4 liJe6 27 i.xe6 fxe6 2S hxg5 hxg5 29 i.xg5 i.xg5 30 liJxg5 �g7 3 1 �e2 left White with good com­ pensation in Vachier-Lagrave - lenni, Euro­ pean Ch, Plovdiv 200S. 22 'iit g2 l:ld7 23 h4! gxh4 24 eS i.e7 25 gxh4 �g7 After 25 . . . i.xh4 26 liJf3 White wins back the pawn and keeps up the pressure. 26 axbS axbS 27 liJe4 (D) ••

compared to the endgames in line 'b2' just above, Black still has his h-pawn. 19 i.xe4 20 fxe4 20 liJxe4 h6 should be alright for Black. 20 liJc7 (D) •••

•••

B

27 ':a8? ! Now Black's position quickly comes undone. 27 . . . i.xh4 must be a better try. White can re­ gain the pawn with 2S i.xh6+ �xh6 29 lIh l , but after 29 . . .':gS+ 30 <3;f3 <3;g7 3 1 ':xh4 ':hS Black has good chances to hold. 28 ':xa8 liJxa8 29 liJg3 i.f8 30 liJfS+ <3;h8 31 i.xh6 i.xh6 32 liJxh6 �g7 33 liJxf7! White had excellent winning chances in Yakovenko-Zhang Zhong, China-Russia match, Ergun 2006. •••

21 a4! White plays on the entire board. Other moves: a) 2 1 liJf3 is not very testing. 2 1 . . . h6 22 i.e3 <3;g7 23 i.c2 :tfeS 24 <3;f2 a5 gave Black coun­ terplay in Grigoriants-Lalic, Cappelle la Grande 2007 . b) 2 1 <3;g2 c5 (2 1 . . .h6 is more solid) 22 liJf3 (White should seek to punish Black by 22 e5 i.e7 23 dxc5 with the point 23 . . . i.xc5 24 liJe4) 22 . . . h6 23 i.e3 cxd4 24 liJxd4 ':feS 25 .:n ':d7 26 ':f6 (or 26 liJf5 i.fS 27 <3;f3) 26 . . . ':xe4 27 �f3 ':e5 2S liJf5 liJeS 29 i.xf7+? (29 liJxh6+ .<3;g7 30 ':xf7+ ':xf7+ 3 1 liJxf7 ':f5+ 32 �e4 ':xf7 33 i.xf7 <3;xf7 34 i.xg5 is level) 29 . . . ':xf7 30 liJxh6+ <3;g7 3 1 ':xf7+ �xh6 + Z.Almasi-Harikrishna, George Marx Memorial, Paks 2006.

02) 16 :ifh6 ! ? (D) This is an interesting idea recommended by Pavlovic. It has not been tried much in grandmaster practice or in correspondence play, but that should not necessarily discourage Black. Pav­ lovic is an experienced grandmaster and his ••

84

UNDERSTANDING THE MARSHALL A TTACK

B

analysis is very interesting. At first sight this move looks rather odd, because the advance h4 looks like it will exploit the pin on the c I -h6 di­ agonal, but this is not so easy to implement (the immediate 1 7 h4 would be met by 1 7 .. :fi'g6, hitting White' s rook). 17 l:tel This move is the first choice of most analysis engines. It is possible they are right, since other moves do not appear to give White anything clear: a) 17 �d2 f5 ! I S l::te l f4 and here: a l ) 1 9 �e4 .ig4 20 'ii'g 2 (20 .id l .ixd l 2 1 lhd l .ic7 2 2 .id2 .:taeS 2 3 'ii'd 3 l::te6 ! { intend­ ing .. :ii'h5 and . . . l::t h 6 } 24 f3 ? ! fxg3 25 hxg3 .ixg3 26 �xg3 'ii'h 3 +) 20 . . . .ih3 2 1 'ii'h l .ic7 22 .id2 'ii'g 6 with compensation - Pavlovic. a2) 19 'ii'g 2 .ih3 20 .ixd5+ 'iith S (this is a typical method, essentially gaining a tempo) 2 1 'ii' h I cxd5 2 2 'ii'xd5 .i.g4 2 3 a4 (23 �e4 .if3 24 'ii'xd6 'ii'h 3 25 'ii'e 5+ is a draw) 23 . . . l:tadS (23 . . . l::taeS ! ? - Pavlovic) 24 'ii'c 6 (24 'ii' h 1 l::td eS 25 l:txeS l:txeS 26 �f1 :e l 27 'ii'a S+ 'ii'fS 2S 'ii'xfS+ .ixfS 29 b3 bxa4 30 bxa4 .ih3 3 1 .ib2 l:te2 32 .ic l l:te l with a repetition - Pavlovic) 24 . . .fxg3 25 fxg3 l:tf2 26 l:teS+ �g7 27 h4 l:txeS 2S 'fixeS .ixg3 . Pavlovic stops here, concluding that Black is doing fine. Even trading queens leaves White under pressure: 29 'ii'e 7+ l:tf7 30 'ii'x g5+ 'ii'xg5 3 1 hxg5 .ih3 with excellent play; for example, 32 axb5 ':'f2 (32 . . . axb5 ! ?) with at least a draw. 33 bxa6? loses to 33 . . . l:tg2+ 34 �h l l::te 2. b) 17 f3 (D) has been played successfully at grandmaster level, but there appears to be at least one neat solution to Black's problems here too:

b l ) 1 7 . . . �hS I S �d2 .ih3 ( I S . . . �f4 ! ? 1 9 gxf4 gxf4 20 �f2 l:tgS) 1 9 'ii'e l ( 1 9 'ii'f2 f5 20 l:e l l:taeS 2 1 a4 'fig6 22 ':'xeS ':'xeS 23 axb5 �f4 24 bxa6?! �d3 25 a7 �xf2 26 as'ii' �d3 +) 19 . . . �f4 20 �f1 (20 gxf4? gxf4 2 1 'iith l l::tg S 22 l::te2 'ii'g6 ! 23 'ii'f2 l::t aeS ! 24 �e4 l:txe4 ! 25 fxe4 f3 -+) 20 . . . 'ii'g 7 2 1 'ii'f2 �d3 22 'ii'd 2 �xc l 23 l:txc 1 f5 24 l::te6 ± Kr.Georgiev-Tseshkovsky, Moscow 1 9S5. b2) 1 7 . . . f5 ! is PavloviC's idea. IS l::te2 (worse is I S .:te l f4 1 9 g4 .ixg4 ! 20 fxg4 f3 +) I S . . . f4 1 9 g4 .ixg4 ! 20 fxg4 f3 2 1 l:tf2 l:taeS gives Black an attack: b2 1 ) 22 .i.xd5+ cxd5 23 �h l l:te2 24 l:txe2 fxe2 25 'ii'xe2 l:tf1 + 26 �g2 l:txc l 27 'ii'e S+ 'ii'fS 2S 'ii'x fS+ .ixfS is clearly awful for White. b22) 22 �a3 and now: b22 1 ) 22 . . . 'fih4 ! ? 23 .id2 l::te 6 and now White must avoid 24 �h l ?, as Pavlovic 's sug­ gestion 24 . . . �g7 ! , threatening . . . l::te 2, leaves White with no adequate defence. White should try instead 24 l:te l , when he might hang on. b222) 22 . . . l:.e4 ! may be even stronger. 17 �h8!? Worse is 1 7 . . . f5 ? ! IS 'ii'g 2 ! l::ta7 19 .i.xd5+ cxd5 20 'ii'x d5+ l::taf7 2 1 .ixg5 'ii'x g5 22 'ii'xd6 f4 23 l::te 5 ±. 18 h4 (D) This is an obvious move but it meets a the­ matic rejoinder. Other moves: a) I S .i.xd5 cxd5 19 h4 'ii' h 5 ! 20 .ixg5 f6 2 1 .i.e3 l::t a7 ! 22 �d2 l::t g 7 again with an attack - Pavlovic . This becomes clearer after 23 'ii'g 2 l::t fgS (planning to detonate on g3) 24 �f1 'ii'xh4 -+, intending . . . .ih3 or . . . .ixg3 . b) I S �d2 .ih3 1 9 'ii'e 2 (after 1 9 'ii'd 3 l:taeS Black has the initiative) 1 9 . . . �f4 ! ? 20 'ii'd l ..•

MODERN VARIA TION:

Jtf5 2 1 ltJf3 (after 2 1 gxf4 gxf4 22 �h 1 .l:.g8 Black has a strong initiative) 2 l . . .ltJh3+ 22 �g2 l:lae8 gives Black counterplay - Pavlovic.

B

18 ltJf4! We shall see this tactical justification of Black's play more than once. 19 hxgS 1 9 gxf4? gxf4 is suicidal. 19 ltJh3+ 20 �g2 'ii'g6 Now 2 1 f4 h6 22 gxh6(?) l:.g8, with an at­ tack, was given by Pavlovic. Perhaps 2 1 'ii'e2 ! ? i s more testing, intending 2 1 . . .ltJxg5 2 2 �xg5 'ii'x g5 23 ltJd2 or 2 l . . .Jtf5 22 ltJd2. The best re­ ply is probably 2 1 . . .f6. •••

•••

03) 16 'ii'hS (D) •••

15 ':e4

17 ltJd2 White has tried several other moves: a) 1 7 f3 Jtf5 (worse are both 17 . . . Jth3 ? ! 1 8 'ii'f2 f5 1 9 lte6 f4 20 l:.xd6 fxg3 2 1 hxg3 l:lxf3 22 l:lxd5 ! , as given by Nunn, and 1 7 . . . f5 1 8 .l:.e2 Jtd7 1 9 a4 .l:.ae8 20 axb5 axb5 2 1 :a7 , which was better for White in Komiagina­ Shadrina, Russian Women 's Ch, Oriol 2006) 1 8 ltJd2 transposes to the main line. b) 17 .l:.e l ? ! Jth3 18 Jtd l g4 19 'ii'd 3 f5 20 c4 l:lae8 21 l:lxe8 l:lxe8 22 Jtd2 ltJf4 ! 23 gxf4? Jtxf4 24 �e2 (24 'ii'c 3 b4 ! wins for Black after 25 'ii'xb4 �fl ! 26 Jtxf4 'ii'h 3) 24 . . . l:lxe2 ! was LJohnson-Cleghorn, COIT. 1 960. If 25 'ii'x e2, then 25 . . . �xh2+ 26 �xh2 Jtfl + wins. c) 17 �d l 'ii'g 6 1 8 ltJd2 Jtf5 19 .l:.e l :ae8 20 ltJf3 Jtd3 (or even 20 . . . f6 ! ?) 2 1 'ii'g 2 Jte4 gives Black compensation. d) 17 Jtd2 ? ! Jtf5 (also possible is 17 . . . f5 1 8 l:le l f4) 1 8 .l:.e l Jth3 1 9 'ii'e 2 Jtg4 20 'ii'f l l:lae8 2 1 ltJa3 Jth3 22 'ilkd3 ltJf4 ! 23 �xf4 (23 gxf4? loses to 23 . . . gxf4 24 �h l f3) 23 . . .gxf4 24 ltJc2 Jtg4 25 .l:.xe8 ':xe8 26 ':'e l �e2 27 'ii'd2 and now 27 . . .'ilkf3 gave Black a strong initiative in Smolensky-Heffner, COIT. 1 985, while the di­ rect 27 . . . f3 may be even stronger. e) 1 7 'ii'e 2? ! 'it'g6 ! 1 8 .l:.e8 Jtf5 1 9 l:.xa8 %has 20 ltJd2 Jtd3 2 1 'ii'e l ltJf4 ! 22 gxf4 gxf4+ 23 �h l �h8 24 'it'g l 'it'f6 25 f3 lte8 ! 26 a4 (26 ltJe4 l:lxe4 ! ) 26 . . . 'it'h4 and now not 27 ltJe4? l:.g8 -+ Ernst-Hebden, Gausdal 1 987. If White tries 27 axb5 then 27 . . . l:.e 1 (worse are 27 . . . axb5 28 Jtxf7 l:.e 1 ? 29 l:.a8+ and 27 . . . cxb5 28 l:lxa6 l:.g8 29 l:.a8 l:.xa8 30 �xf7) 28 bxa6 .l:.xg l + 29 'itxg l 'it'e l + 30 'itg2 'ii'e 2+ 3 1 �gl Jtxa6 =+= gives Black good winning chances. 17 JtfS 1 7 . . . ltJf6 ? ! was played in Vachier-Lagrave ­ Fressinet, Paris 2007, which was drawn after 1 8 f3 ? ! ltJxe4 1 9 fxe4 Jth3 20 'it'f3 'ii'xf3 2 1 ltJxf3 h6 22 Jte3 �g7 23 a4 Jte6 24 Jtc2 l:.fb8 - if anyone is better at this point, it is Black. How­ ever, Renet gives 1 8 'it'e2 ! 'it'g6 1 9 J:.e3 ltJg4 20 l:.e8 Jtd7 2 1 l:.xa8 .l:.xa8 22 ltJe4 ±. The main alternative for Black is 1 7 . . . f5 . This has a dubious appearance, but may stand up to analytical scrutiny: a) 1 8 l:.e l was given as dubious in New in Chess Yearbook 81, but White may be able to fight for an advantage even with this move. 1 8 . . . f4 (D) and here: ••.

As in the previous line, Black prefers to play the middlegame, but putting the queen on this square has been seen a lot more in practice.

85

86

UNDERSTANDING THE MARSHALL A TTACK

w

a 1 ) 1 9 �d 1 ? ! f3 ! 20 ttJe4?! �h3 2 1 �d3 .l:tae8 22 �d2? .l:txe4 ! ! 23 �xe4 �xg3 ! 24 �xf3 �xh2+ 25 ..ti>xh2 �g4+ 26 ..ti>gl �xf3 27 �e6+ ..ti>g7 28 �e5+ .l:tf6 29 iVh2 .l:th6 0- 1 Vachier-Lagrave - Svid1er, Donostia 2009. a2) 1 9 ttJe4 �h3 ? ! ( l 9 . . . �g4 is more to the point - Nunn; definitely not 1 9 . . . fxg3 ? 20 hxg3 �g4 2 1 �xg5 �xg3 22 ttJxg3 'iWxg5 23 l1e5 'iWh4 24 .l:txd5 and White wins, Z.Almasi-Gyi­ mesi, Hungarian Team Ch 2005/6) and instead of 20 'iVd3 ? ! l:1ae8 2 1 �d2 ? ! �f5 22 c4 (Sam­ buev-Khrushchov, Geller Memorial, Moscow 2006) 22 . . . bxc4 23 �xc4 �g6 24 f3 fxg3 +, White can improve with 20 �d l ! g4 21 �d3 .l:tae8 22 �d2 �c7 23 �c2 ±, as given by Renet. b) 1 8 �d 1 iVh6 (worse is 1 8 . . . g4? 1 9 .1:te 1 f4 20 ttJe4 �c7 2 1 �b3 ..ti>h8 22 gxf4 ttJxf4 23 �xf4 �xf4 24 'ii'g 2 .l:ta7 25 I:te2 l:te7 26 ttJg3 'iWg5 27 .l:txe7 �xe7 28 iVxc6 +- Motylev­ Beliavsky, Wijk aan Zee 2006) 1 9 %:te l f4 20 ttJe4 �c7 (D) (20 . . . �h3 ? ! 2 1 �e2 .l:tae8 22 'ii'h5 ! 'ii'xh5 23 �xh5 .l:te6 24 �d2 ± Shaba­ lov-Vescovi, Sao Paulo 2009) and now:

b 1 ) 2 1 �f3 �h3 22 �d3 .l:tf7 23 �d2 ':'af8 24 �h 1 .l:tg7 25 �f3 .l:tgf7 26 �h 1 .l:tg7 27 �f3 .l:tgf7 was drawn in Svidler-Leko, MorelialLin­ ares 2007 . A weird sequence and outcome. b2) 2 1 �d2 �f5 (not good is 2 1 . . ...ti>h8 ? ! 22 'iVe2 �f5 23 'ii'h 5 ! �xh5 24 �xh5 �xe4 25 ':'xe4 fxg3 26 fxg3 ttJf6 27 .l:te7 �xg3 28 hxg3 ttJxh5 29 ..ti>g2 ± Vallejo Pons-Sargisian, Team event, Calvia 2007) 22 ttJc5 and here: b2 1 ) 22 . . J:t.f7 ? ! 23 �e2 .l:taf8 24 �h5 �f6 25 g4 ! �g6 26 'iVh3 (Kosten gives 26 'iWh6 %:te7 27 .l:txe7 'iWxe7 28 �f3 �d6 29 .l:te 1 +-) 26 . . . .l:te7 27 .l:txe7 'iVxe7 28 �f3 .l:te8 29 'iVn ± Shirov-Akopian, Karlovy Vary 2007 . b22) 22 . . . �d6 ! 23 �f3 ..ti>h8 24 ttJb7 �b8 25 a4 .l:ta7 26 axb5 (26 ttJc5 �h3 27 'ii'e2 .l:.af7 gives Black good counterplay) 26 . . . .l:.xb7 27 l1xa6 fxg3 28 fxg3 (Galkin-P.H.Nielsen, Euro­ pean Ch, Dresden 2007) 28 . . . �g6 ! is OK for Black; e.g., 29 c4 (29 'iWg2 .l:tbf7) 29 . . . .l:.bf7 30 .l:.a3 �d6 ! 3 1 c5 �h5 32 cxd6 �xf3 33 d7 �h5 34 'iWh3 ! �f6 with a likely draw. We now return to 1 7 . . . �f5 (D) :

18 f3 ttJf6! This move became well known after Leko used it to defeat Kramnik in a famous game from their 2004 world championship match. Other moves look worse: a) 1 8 . . . �xe4?! 19 fxe4 ttJe3 ( l 9 . . . ttJb6 20 e5 �e7 2 1 'iWf5 ..ti>g7 22 ttJe4 ±) 20 'ii'f3 �xf3 (also bad is 20 . . . ttJg4 2 1 ttJn intending �d 1 +-) 2 1 ttJxf3 ttJc4 22 ttJxg5 ±. b) 1 8 . . J1ae8? ! and now: b 1 ) 1 9 a4 �xe4 20 ttJxe4 l:te6 2 1 axb5 axb5 22 ttJxg5 .l:.g6 23 f4 'iWg4 24 �d 1 iVf5 25 �e2 ± Osorio-Grabner, COIT. 2005 .

MODERN VARIA TION:

15 hIe4

87

b2) 1 9 .l:.xe8 ':'xe8 2 0 ltJe4 (this is a typical idea - White returns the pawn and hopes that his bishop-pair and Black's weakened kingside will ensure some advantage) 20 . . . i.. xe4 2 1 fxe4 .l:.xe4 22 i.. d l g4 and now instead of 23 'ir'f2 .l:.e6 24 a4 b4 25 i..d2 bxc3 26 bxc3 c5, which was rather unclear in Elyakim-Krempel, COIT. 1 990, White could play 23 a4. If Black tries to simplify with 23 . . . ltJe3 White has 24 'ir'e2 ! i.. x g3 ! ? (what else?) 25 i.. xe3 i.. f4 26 i.. xf4 .l:.xe2 27 i.. xe2 ±. We return to 18 ... ltJf6 (D):

23 'ir'f2? White follows his preparation into the abyss. Instead 23 'ir'd l leads to a draw after 23 ... i..e 2 24 'i!i'e l (24 i.. c 2? 'i!i'h5 25 �e l i.. xf3 26 'iVf1 i.. d5 27 'i!i'xf6 i.. x g3 ! gives Black a winning at­ tack - Leko) 24 . . . i..d 3 25 'ir'd l with a repeti­ tion. 23 .. J:te2 24 'i!i'xe2 Also losing is 24 bxa6 ':'xf2 25 'it>xf2 'ir'h6 26 'it>g l (26 'it>g2 g4) 26 . . . i.. x g3 27 hxg3 'ii'h 3 28 a7 'i!i'xg3+ 29 'it>h l g4 ! 30 a8�+ 'it>g7 with a winning attack. 24 i.. xe2 25 bxa6 'ilVd3! (D) .•.

White now has an important choice to make: 87 031 : 19 .1:.el 032: 19 'ir'g2 88 033: 19 a4 89

w

03 1 ) 19 l:tel This move is not especially promising, but as Krarnnik (in)famously chose it in his world championship match with Leko, we shall give it detailed coverage. 19 ... .l:.ae8 20 .l:.xe8 .l:.xe8 21 a4 'ir'g6! 22 axb5 White can also play 22 ltJe4 ltJxe4 23 fxe4 i.. xe4 24 i.. x g5 ! (24 axb5 ? i..d 3 ! 25 i.. xf7+ 'ir'xf7 26 'it'xd3 .l:.e l + 27 'it>g2 'ir'd5+ 28 'it>h3 .l:.g l 29 'ir'e2 'ir'f5+ 30 g4 .l:.xg4 ! and Black forces mate) 24 . . . bxa4 25 i..c4 and here Knaak recommends 25 . . . i.. d5 26 i..xd5 cxd5 27 'ir'f6 a3 28 'ir'xg6+ hxg6 29 bxa3 .l:.c8 30 i..d2 ;!; while Leko gives 25 . . . .:.b8 26 .l:.e l i..d5 27 i..xd5 cxd5 28 i..c 1 h5 with counterplay. 22...i.. d3 (D)

26 'it>f2 After 26 a7 Black wins with 26 . . . 'i!i'e3+ 27 'it>g2 i.. xf3 + ! 28 ltJxf3 'ir'e2+ 29 'it>gl ltJg4 ! 30 a8'i!i'+ 'it>g7 3 1 'ii'x c6 'ir'f2+ 32 'it>h l 'ir'f1 + 33 ltJg l ltJf2#. 26 i..xf3 27 ltJxf3 ltJe4+ 28 'it>el ltJxc3 29 bxc3 'ir'xc3+ 30 'it>f2 'ir'xal 31 a7 h6 32 h4 g4 0- 1 Krarnnik-Leko, World Ch match (game 8), Brissago 2004. .•.

88

UNDERSTANDING THE MARSHALL A TTACK

03 2) 19 �g2 (D)

B

19...�g6 This move is best. 1 9 . . . tDxe4 ? ! 20 tDxe4 �xe4 2 1 fxe4 gives White excellent compensa­ tion for the exchange, and 1 9 . . . �h3 ? ! 20 �f2 l:tae8 2 1 l:txe8 ':xe8 22 a4 also looks insuffi­ cient for Black. 20 l:te3 After 20 l:te l l:tae8 2 1 lhe8 l:txe8 22 tDn Black could try either 22 . . . �d3 or 22 . . . c5 . 20 l:tae8 Grishchuk's move 20 . . . tDd5 ! ? is another idea for Black. After 2 1 l:te l ltae8 22 �f2 (D) (White may investigate 22 tDe4 ! ? h6 23 �d2) Grishchuk had quite a surprise in store.

a) 28 �d l ? ! looks good, but hits a snag: 28. ..�h3 29 �e2? (29 �c2 is complex and might survive) 29 . . . �e7 ! 30 �d3 �h4 3 1 �e2 �g5 32 'ili'e5+ �xe5 33 dxe5 �f2 0- 1 Shomoev­ Grishchuk, Russian Team Ch, Dagomys 2008. b) 28 a4? also runs into trouble: 28 . . . �g6 ! (28 . . . �h3 29 tDe4 �g6 30 �xf7 �g2+ 3 1 Wgl �xf7 32 tDxd6 �f6 is unclear - Golubev) 29 h4 �e7 30 �h2 �xh4 3 1 �xf7 �xf7 32 �xh4 �g6 33 �h2 �h3 -+. c) 28 tDe4 is safest. 28 ... �xe4 29 fxe4 f3 30 �g5 (there is nothing else) 30 . . . .l:!.xg5 3 1 l:tg l is about equal. 21 tDe4 2 1 .l:txe8 transposes to the note to White's 20th move. 21...tDxe4 (D) Black should avoid 2 1 . . . tDd5 ? ! 22 �xd5 cxd5 23 tDxd6 'ili'xd6 24 l:txe8 lhe8 25 �xg5 �e6 26 g4 �d3 27 .l:td l 'iVe l + 28 l:txe l l:txe l + 29 Wf2 l:te2+ 30 Wg3 l:txg2+ 3 1 Wxg2 ± Anand­ Aronian, Amber Rapid, Monte Carlo 2007 .

w

•••

B

22 . . . tDf4 ! 23 gxf4 gxf4+ 24 Wh l l:txe l + 25 �xe l Wh8 26 �g l (after 26 �f2 l:tg8 intend­ ing . . . �h3 Black has a strong attack) 26 . . . �h5 27 �f2 l:tg8 and here:

22 g4! This is the point of White 's 2 1 st move. With­ out this resource White would just be worse. 22 ...tDg3 22 . . . �xg4 23 fxe4 gives White a strong, mobile centre. 23 hxg3 �d3 (D) Previously, Aronian had played 23 . . . �bl against Anand with success: 24 'iVe2 l:txe3 25 �xe3 h6 26 �e l �c2 27 �xc2 �xc2 28 �e4 �d l + 29 Wg2 Wg7 30 �e3 �xg3 ! 3 1 Wxg3 l:te8 32 �xe8 ! 112-112 Anand-Aronian, Wijk aan Zee 2007. However, White can probably im­ prove here, perhaps with 24 a4 b4 25 �c4 ! ?

MODERN VARIA TION: 15

l:Ie4

89

:'a6 is clearly good for White, but Black could investigate 1 9 . . . 'ifg6 20 l:.e l l:.aeS with the idea 2 1 axbS .td3 22 'iff2 lhe l + 23 'ifxe l l:.eS . 20 liJxe4 After 20 fxe4? ! i.e6 =+= White does not have enough for the exchange because Black is well developed and ready to strike back with . . . fS . 20 :ii'g6 (D) Following 20 . . . .txe4?! 2 1 fxe4
24 .td2 24 'ii'd2 ! ? is also possible. Black will have to find an improvement on 24 . . . :'xe3 2S 'ifxe3 .txg3 (2S . . . h6 ! ?) 26 'ifxgS l:.eS 27 .te3 'ii'x gS 2S .txgS
033) 19 a4 (D)

White looks to open a second front. 19 )iJxe4 In practice, Black always captures the rook. 1 9 . . . l:taeS ? ! 20 l:txeS l:txeS 2 1 axbS axbS 22 .•

21 liJxd6 Black must also be ready for 2 1 liJxgS, which is critical. 2 1 . . .l:taeS ! (Black can also try 2 1 . . .cS, but White can fight for an advantage with either 22 .tdS or 22 liJe4) 22 liJe4 (22 axbS .td3 23 "d l axbS and now 24 l:.a7 is unclear, while White should avoid 24 l:ta6? h6 2S liJe4 i.xe4 26 fxe4 .txg3 ! =+= Harding) 22 . . . .txe4 23 fxe4 l:txe4 (not 23 . . ...xe4? 24 .th6 ±) 24 .tc2 :'feS (24 . . . fS 2S .txe4 fxe4, with compensation, was given by Marin but the text-move is more con­ vincing) and now White has: a) 2S i.d2 i.xg3 26 .txe4 l:txe4 27 hxg3 'ifxg3+ 2S 'ifg2 'ifxg2+ 29
UNDERSTANDING THE MARSHAU A TTACK

90

+) 26 . . . bxa4 ! ? 27 .i.d2 (27 l:ha4? loses to 27 . . . 'ii'g4, because of 28 'ii'f l .i.xg3 ! ) 27 . . . h5 28 'ii'f3 h4 29 .i.e 1 c5 30 .i.f2 cxd4 3 1 cxd4 hxg3 32 hxg3 :f4 33 'ii'c 3 :g4 34 'iii>g 2 = is one line given by Harding. 21 'ii'xd6 22 .i.xg5 (D) •••

23 . . .:fe8 24 .i.f4 .i.e6 1/2- 1/2 Moura-Harding, corr. 2006-7 . Harding gives 25 .i.c2 .i.f5 , when 26 .i.b3 .i.e6 repeats. He also gives 26 .i.d l :ad8, although White could still try 27 'ii'd2 with an interesting fight ahead. 24 axb5?! Opening files for Black's rooks looks wrong. After 24 .i.f4 the position is still unclear. 24 axb5 25 :xa8 :xa8 26 'iii>f2 .i.c4 (D) •••

22 'ii'g6! This is a strong move, not offering White the chance to demonstrate an improvement on 22 . . . :fe8 23 :e l 'ii'g 6 24 .i.e7 :a7 25 .i.c5 :aa8 26 .i.e7 :a7 27 .i.c5 1/2- 1/2 Shirov-Leko, Tal Memorial, Moscow 2006. 23 'ii'c 1 (D) After 23 .i.e7 :fe8 24 :e l .i.e6 25 :xe6 ! ? (25 .i.xe6 l:he7 2 6 :e4 'ii'h 6 i s fine for Black) is an interesting idea of Harding's. White sacri­ fices a second exchange, but his position looks easier to play. He gives 25 . . .fxe6 26 .i.c5 :ab8 27 'ii'c 1 bxa4 28 .i.xa4 :ec8 29 .i.c2 'ii'h5 30 g4 'ii'h4 3 1 'ii'e 3 :e8 with unclear play. •••

23 .i.d3 •••

27 .i.xc4 White would rather not play this, but 27 .i.d l 'ii'd 3 is awful and 27 .i.c2 'iWe6 28 .i.e3 'iWh3 29 �gl :a2 30 .i.f2 h6 ! stops 'iWg5+ and threatens . . . :xb2. The rest of the game is a gem so I give it in full. 27 bxc4 28 g4 :e8 29 .i.f4 1M3 30 �g3 'ii'e2 31 'ii'b l?! 'ii'e l + 32 'ii'x e1 ':'xe1 33 .i.d6 :gl+ 34 'ittf2 :bl 35 .i.a3 'ittg7 36 'ittg3 �g6 37 h3 h5 38 'itth 4? (D) 38 h4 retains drawing chances. .••

38 :g1 ! 39 .i.c5 :g2 40 .i.a3 f6 •••

MODERN VARIA TION:

15 l:te4

91

Zugzwang. 41 gxh5+ �f5 42 f4 l:.g8 43 .i.d6 �e6 44 h6? ! 'ittxd6 45 �h5 f5 46 h7 l:.h8 47 �g6 �e7 48 'ittg7 'itte 8! (D)

w

After 59 h7 :f8 60 'ittg 7 :h8 ! 6 1 'itt g 6 (6 1 'ittx h8 'ittf7 is the pseudo-stalemate trick again) 6 l . . .'ittf8 62 'itth 6 'ittf7 Black wins easily.

49 'ittg6 49 'iltxh8 'ittf7 puts White in zugzwang. Black will keep moving his king from f7 to f8 until White is forced to push his b-pawn. After Black captures on b3, White's c-pawn will be free to move and Black will quickly queen and mate. 49 �f8 50 h4 'itte7 51 'ittg7 'ilte8 52 'ittg6 �f8 53 h5 �e7 54 �g7 'ilte8 55 'ittg6 'ittf8 56 h6 �e8 57 �f6 57 �g7 �e7 58 �g6 'iltf8 does not help White. 57 :xh7 58 �g6 :f7! (D) 0- 1 Shirov-Aronian, Tal Memorial, Moscow 2006. •••

•••

Concl usions The Modem Variation leads to some o f the most unusual and interesting positions in the Marshall Attack. Both sides have unique tactical and posi­ tional problems to solve. 1 6 'ii'e2 is not consid­ ered dangerous following Anand's discovery of 1 8 .. .f4 ! . The most popular move, 1 6 'iff3, leads to interesting play in both of its main lines and Black has his share of the play. The most topical move, 1 6 'iffl , can be met by heading into an un­ clear endgame with 1 6 . . . 'ifxfl + or with the more ambitious 1 6 . . . 'iVh5, which leads to a very deli­ cate rniddlegame, with 16 . . . 'ifh6 a rare but inter­ esting alternative.

Pa rt 2 : Other Li nes after 8 c3 d S 9 exd S ttJxd S 1 0 ttJxeS ttJxeS 1 1 l:xeS c6

5 Refi ned Rook- Lift : 1 2 d3 1 e4 e5 2 tt:Jf3 tt:Jc6 3 ii.b5 a6 4 ii.a4 tt:Jf6 5 0-0 ii.e7 6 l:el b5 7 ii.b3 0-0 8 c3 d5 9 exd5 tt:Jxd5 10 tt:Jxe5 tt:Jxe5 11 l:xe5 c6 12 d3 (D)

B

the direct 1 3 . . . 'ii'h 4, while the next chapter con­ siders the modem preference, 1 3 . . . .ltf5 . 13 'ii' h4 Black continues as he does in the first few chapters. However, White can exploit a tactical detail that was not available to him in Chapter 4. 14 g3 'ii'h3 15 l:e4 (D) •••

This is a more modest-looking continuation than 12 d4. At first glance it does not seem like this should be dangerous for Black, but White maintains control over the e4- and c4-squares, which allows for some tactical possibilities. White has not grabbed as much space, however, and the pawn on d3 may become vulnerable it­ self. White will often be more than willing to offer this pawn in order to fight for the initia­ tive. It will all come down to the specifics of each position. Despite the similar appearance at first glance of this position to those we have al­ ready seen, the play is surprisingly different from that in the first four chapters. 12 ii.d6 13 l:el (D) Now Black faces an important decision. In this chapter we look at the most natural move, •••

REFINED ROOK-LIFT:

This move is the point of White's play. He intends l:th4 often with lbd2-e4 to follow. Compared to the previous chapter, 1 5 . . . g5 is not possible, because after 16 .i.xg5 'iff5 the e4rook is protected and White will 'remain two pawns up. Therefore Black must find another way to create counterplay. There are four main continuations: A : lS .lt d7 93 94 B : ls lbf6 C : lS 'ifd7 96 99 D: lS 'iffS •••

•••

•••

•••

15 . . . �b7 could also be played, although after 1 6 lbd2 Black should probably play 1 6 . . . 'ifd7, transposing to Line C. Instead, 1 6 . . . l:tae8? ! 1 7 l:th4 ! ? (less ambitious i s 1 7 'ifn 'ifd7 1 8 l:txe8 l:txe8 1 9 lbe4 c5 20 �d2 .ltfS, as in Leko­ Adams, Belgrade 1 995, when Adams gives 2 1 f3 f5 22 lbf2 ;1;; ) 1 7 . . .'ifd7 1 8 lbe4 �e7 1 9 �g5 f6 20 'ifh5 ! (or even 20 �e3 f5 2 1 lbc5 'ii'c 8 22 lbxb7 'ifxb7 23 l:td4 'it>h8 24 �xd5 cxd5 25 'iff3 ± Timoshenko-Fang, Vienna 1 998) 20 . . . fxg5 2 1 'ifxh7+ 'it>f7 was A.lvanov-Fang, Lowell 1 993, and now 22 �d l ! gives White a winning attack.

A) lS .ltd7 (D) •••

This move was tried a couple of times by Kamsky. Black just gets on with his develop­ ment. 16 lbd2 Surprisingly, the consistent 1 6 l:.h4 is un­ tried, although after 16 . . . 'iff5 White can still

1 2 d3

93

continue with 1 7 lbd2, because 1 7 . . . 'ii'x d3 ? loses to 1 8 .ltc2 - a typical idea in the 1 2 d3 variation. 1 7 . . . :tae8 is better, with the idea 1 8 lbe4 �e7 to pester White ' s rook. This is an­ other common theme. 16 gS 1 6 . . . .:.ae8 is less explored, but it is probably just as good (or bad ! ) as the main continuation. a) If White continues aggressively with 1 7 l:th4 Black gets good play after 1 7 . . . 'ii'f5 : a l ) 1 8 lbe4 �e7 1 9 l:.h5 'ii'g 6 20 l:te5 f6 ! (20 . . . f5 2 1 lbd2 f4 22 lbe4 'iith 8 23 'ifh5 is better for White according to Lalic) 21 .ltxd5+ (2 1 ':'xd5 cxd5 22 �xd5+ 'it>h8 intending . . . f5 favours Black) 2 l . . .cxd5 22 ':'xd5 �c6 23 l:.d4 f5 24 lbd2 �c5 25 l:tf4 (it may be better to sac­ rifice the exchange with 25 lbb3, although Black is not worse) 25 . . . 'ife6 gives Black a strong ini­ tiative. a2) 1 8 lbn 'ii'g 6 1 9 .i.d2 ( 1 9 'iff3 l:te l 20 :e4 l:.e8 2 1 l:.xe8+ .ltxe8 gives Black good compensation, and the greedy 22 �xd5 cxd5 23 'ii'xd5 ? �f8 ! intending . . . �c6 gives Black a crushing initiative) 1 9 . . . �f5 (or 1 9 . . . �e7 20 l:te4 �f5 2 1 l:.e5 .ltxd3 22 .ltxd5 cxd5 23 lbe3 �d6 24 l:txe8 l:txe8 25 lbxd5 l:.e2 26 .ltf4 'ii'e4 ! with compensation - Lalic) 20 �xd5 (20 d4 �e7 traps the rook) 20 . . . cxd5 2 1 lbe3 and Black has compensation after either 2 1 . . . .ltxd3 22 lbxd5 �e2 or 2 l . . .�e6 (Murei-Geller, Am­ sterdam 1 987). b) White should probably prefer 17 l:txe8 l:.xe8 1 8 lbe4 .i.c7 1 9 .ltd2 intending 'ifn , when Black still has to prove he has enough for the pawn. We now return to the position after 16 . . . g5 (D) : •••

UNDERSTANDING THE MARSHALL A TTA CK

94

17 l:te2 This prophylactic move has been White's most popular choice. Other moves such as 17 'ji'fl and 1 7 a4 are also possible, while practice has seen 17 lDf3 h6, and here: a) I S :d4 tries to quell Black's play with an exchange sacrifice. I S . . . i.c5 1 9 l:txd5 cxd5 20 i.xd5 l:tadS 2 1 d4 was Mukhutdinov-Berzinsh, Naberezhnye Chelny 1 993. Now Lalit gives 2 l . . .i.g4 ! with the idea 22 dxc5 i.e6. b) I S lDe5 is critical. I S . . . i.f5 ( I S . . . i.xe5 1 9 l:txe5 ±) and now : b 1 ) 1 9 lDg4 can be met by 1 9 . . . i.xg4 ! (worse is 1 9 . . . i.f4?! 20 lDxh6+ 'ji'xh6 2 1 gxf4 i.xe4 22 fxg5 'ji'g6 23 dxe4 'ji'xe4 24 i.c2 'ii'h4 25 'ji'f3 :feS 26 i.d2 :t Lalit) 20 l:txg4 (20 'it'xg4 'it'xg4 2 1 l:xg4 lDf6 22 %:td4 i.c5 also traps the rook) 20 . . . lDf6 2 1 l:e4 lDxe4 22 dxe4 i.e5, when White has compensation for the exchange but I do not think Black is worse. b2) 1 9 lDxc6 ! was queried by Ivanchuk in his lnformator notes, but Lalit points out that it is not bad at all. 1 9 . . .lDf6 (with an attack Ivanchuk) 20 'it'f3 ! i.xe4 2 1 dxe4 lDg4 22 'ji'g2 'it'xg2+ 23 �xg2 l:tfeS 24 f3 gives White the better chances with two pawns for the exchange in the ending. 17 f5? ! (D) This is consistent, but it just looks bad. In­ stead 17 . . . 'it'h5 I S l:e 1 ! :t is given by Ivanchuk.

(24 . . . l:adS and 24 .. .11f6 are also good) 25 i.xe3 l:aeS 26 l:e l i.f3 27 d6 i.c6 2S 'ii'e2 'iVh3 29 'ii'f l 'ii'e6 30 i.d l 'ji'd5 3 1 f3 'ji'xd6 + Anand­ Kamsky, Dortmund 1 992. 18 f4 I S . . . bxc4 1 9 dxc4 f4 20 'iWfl ! +- wins for White (but not 20 cxd5 ? fxg3 2 1 hxg3 l:txf2 ! 22 l:txf2 'iWxg3+ 23 l:tg2 'ji'e3+ 24 1:[f2 'ji'g3+ with a draw). 19 'iWfi! 1 9 cxd5 ? fxg3 20 hxg3 (after 20 fxg3 i.c5+ 21 �h l i.g4 Black has the initiative - Anand) 20 . . . l:xf2 ! again leads to a draw. 19 fxg3 20 hxg3 (D) •••

•••

B

•••

w

20 'iWh5 20 . . .lDf4 should also fail against accurate de­ fence. 2 1 c5+ �hS 22 cxd6 'ji'h5 23 gxf4 i.h3 24 'it'xh3 'ii'x h3 25 ':e3 'ji'h4 26 l:.g3 gxf4 27 lDf3 'ii'f6 2S l:tg2 ± is one line given by Ivan­ chuk. 21 cxd5 i.g4 22 dxc6+ c:j;g7 23 l:te4 This was Ivanchuk-Kamsky, Tilburg 1 992. White is winning. •••

B) 15 lDf6 (D) This move also looks a little doubtful. The main line sees White sacrificing an exchange for tremendous compensation. If Black wants to make 1 5 . . . lDf6 playable, probably it will have to be with the positional play within the notes rather than by following the main line. 16 l:th4 'iWf5 17 lDd2 This developing move is best, because Black cannot play 1 7 . . . 'ji'xd3? I S l:td4. Other moves •••

18 c4! I S lDe4?! is a thematic idea, but I S . . . fxe4 1 9 dxe4 i.g4 20 exd5 c5 ! 2 1 'ji'fl 'ji'h5 ! gives Black a strong initiative for the pawns. 22 l:e3 (or: 22 l:d2 i.xg3 ! 23 hxg3 i.f3 ; 22 l:.e l c4 23 i.c2 i.xg3 !) 22 . . . c4 23 i.c2 i.c5 24 i.d2 i.xe3

REFINED ROOK-LIFT:

have also been tried, but they look less danger­ ous: a) 17 �f4 �e7 ! leaves the h4-rook feeling rather uncomfortable. After 1 8 �c2 'it'g6 19 d4 .i.fS 20 .i.xfS 'fixfS 2 1 .i.eS tt)dS 22 g4 'it'g6 23 l:.h3 f6 24 �g3 fS 2S gxfS ':xfS Black had more than enough for the pawn in Ragger­ Beliavsky, Graz 2008. b) 17 d4 'it'g6 ! ? 18 �c2 �fS 19 .i.xfS 'it'xfS (the exchange of light-squared bishops com­ bined with the awkward position of the h4-rook usually leaves Black with sufficient compensa­ tion for the pawn) 20 �e3 (20 �f4 �e7 ! trans­ poses to line 'a' above) 20 . . . l:.ae8 2 1 tt)d2 tt)dS 22 tt)f1 and now 22 . . . f6? ! 23 a4 gS 24 l:.hS 'it'd7 2S axbS axbS 26 �d2 ! <J;; g 7 (Z.Almasi­ Lukacs, Austrian Team Ch 1 99617) 27 tt)e3 gives White a huge advantage. Almasi pro­ poses 22 . . . 'ii'g 6 ! 23 'ii' h S 'ii'x hS 24 l:IxhS fS , and 22 . . . .i.e7 ! ? 23 l:.hS 'it'd7 also looks quite playable for Black. c) 17 �c2 'ii'g 6 1 8 tt)d2 ( 1 8 d4 is line 'b' ) 1 8 . . . �g4 1 9 tt)f3 ( 1 9 f3 �fS 20 tt)e4 can be met by either 20 . . . �e7 2 1 �gS l:.ae8 22 tt)xf6+ �xf6 23 �xf6 'ii'xf6 with compensation, or 20 . . . l:.ad8 2 1 d4 l:.fe8 22 tt)xf6+ 'ii'xf6 23 �e4 h6 again with compensation, as in Y.Geller­ Naiditsch, Moscow 2(08) 1 9 . . . l:.fe8 20 ..tg2 tt)dS 2 1 d4 �fS 22 �xfS 'ii'xfS with compensa­ tion, Leko-Adams, Madrid 1 998. 17 g5 As mentioned above, 1 7 . . . 'ii'xd3 ? loses to 1 8 l:Id4, but Black has tried a couple of less radical ideas: a) 1 7 . . . ':'e8 is natural, but Polgar found a way to seize the initiative by returning the pawn immediately: 1 8 tt)e4 tt)xe4 1 9 l:.xe4 ( 1 9 dxe4 ! ? •••

12 d3

95

i s also possible because 1 9 . . . ':'xe4? loses to 20 'ii'x d6, and Polgar gives 1 9 . . . 'ii'g 6 20 �c2 .i.e7 2 1 l:.f4 as unclear) 1 9 . . . l:.xe4 20 dxe4 'ili'xe4 2 1 �c2 ! (not 2 1 'ii'xd6?? �h3) 2 1 . . .'fie7 22 �gS f6 23 �e3 �e6 24 'ii'f3 'ii'd7 2S :d l ':'d8 26 �e4 with strong pressure for White. 26 . . . �xa2 27 .i.b6 �b3 28 l:.d4 ! cS 29 �xcS 'ii'e6 30 c4 1 -0 was the abrupt finish of 1.Polgar-Svidler, Wijk aan Zee 200S . b) 1 7 . . . tt)g4 could use further tests. 1 8 f3 ( 1 8 tt)e4 �e7 1 9 f3 tt)eS 20 l:.f4 'ii'd7 2 1 tt)f2 cS 22 ':'e4 tt)g6 gave Black some compensation in Yakovenko-Smikovsky, Russian Team Ch, Omsk 2(0 1 ) 1 8 . . . tt)e3 ! ? ( 1 8 . . . tt)eS ? ! 1 9 d4 tt)g6 20 �c2 ±) 1 9 'ii'e2 tt)dS 20 c4 tt)c7 2 1 d4 'ii'g 6 22 'ii'e4 �e6 23 'ii'xg6 hxg6 24 cS �e7 2S ':e4 l:.fd8 gave Black counterplay against White's d-pawn in Smeets-Hebden, European Union Ch, Liverpool 2008. White may improve here somewhere but at least Black has some counter­ play, whereas in the main line Black's kingside is weak and the situation looks rather grim. We now return to 17 . . . gS (D):

w

lS l:.h6! This aggressive move looks funny, but it is clearly the strongest. Instead 1 8 tt)e4 gxh4 1 9 tt)xd6 'ii'g 6 intending . . . �g4 gives Black good counterplay according to Lalit, while Ivan­ chuk's creative 1 8 l:.d4 is probably too risky af­ ter 1 8 . . . �cS 1 9 tt)f3 h6 ! 20 �f4?! (trying to exchange bishops on d6 through tactical means, but 20 h4 is unclear at least, and a better try) and here: a) 20 . . . gxf4?! 2 1 ':'xf4 'ii'g 6 22 tt)eS 'ii'g7 23 'ii'f3 ! �d6 24 l:.xf6 �xeS 2S ':'xf7 ':'xf7 26 'ii'xc6 l:.b8 27 'ili'e8+ ..th7 28 'ii'xf7, with four

UNDERSTANDING THE MARSHALL A TTACK

96

pawns for a piece, is the point of White ' s odd bishop move. b) 20 . . . iLb7 2 1 iLd6 iLxd6 22 l:lxd6 c5 23 lDd2 ? ! (23 lDe l is better, but after 23 . . . l:taeS Black still has a strong initiative) 23 . . . l:.adS 24 .l:.xdS .l:.xdS 25 iLc2 lDg4 26 'it'e2 c4 + Ivan­ chuk-Timman, Wijk aan Zee 1 999. c) 20 ... l:teS ! 21 iLd6 i.a7 intending . . . c5 and . . . i.b7 gives Black more than enough for the pawn. We now return to the position after IS l:th6 (D) :

White has scored very well here. The g-pawn is falling and White will have the bishop-pair, a safer king and two pawns for exchange. Some examples: a) 2 l . . .g4 22 iLg5 iLf5 23 'it'd2 ± gives White a bind. b) 2 l . . .i.f5 22 lDxg5 iLxd3 23 lDf3 ! iLc4 24 lDe5 'ii'e6 25 i.xh6 'ii'x h6 26 lDxc4 bxc4 27 i.xc4 ± gives White two pawns for the ex­ change. Black's pawns are weak and his king position is uncomfortable. c) 2 l . . .i.g4 22 'ii'd2 lDf5 23 'ii'x g5 i.f3 24 'ii'f4 iLxe4 25 dxe4 lDg7 26 iLe3 ± was misera­ ble for Black in Anand-Shirov, Mainz rapid 2004. d) 2 l . . .lDg4 22 iLxg5 iLf5 23 iLf4 l:tadS 24 'iWe2 l:.feS 25 :e l h l i.xe4 ? ! (this gives White a huge centre, so shedding another pawn with 29 . . . c4 30 d4 lDd3 3 1 iLxd3 cxd3 32 'ii'x d3 .l:.e7 was a better chance of getting some counterplay) 30 fxe4 lDd7 31 d4 cxd4 32 cxd4 ± Svidler-Adams, Elista Olympiad 1 995.

18 lDg4 Black attacks the rook and threatens to play 1 9 . . . 'it'xf2+. 19 lDe4 This defends the f2-pawn and attacks the d6-bishop. 19 lDxh6 20 lDxd6 'ii'g6 21 lDe4! (D) After 2 1 'ii'd2 'ii'xd6 22 'ii'x g5+ 'ii'g 6 23 'ii'xg6+ hxg6 24 iLxh6 .l:.eS Black was fine in Leko-Grishchuk, Dubai rapid 2002.

C)

•••

•••

15 'ii'd 7 (D) •••

W

B

This voluntary retreat is a bit comical if one considers that it can also arise from the move­ order 1 2 .l:.e 1 iLd6 1 3 g3 'ii'd7 14 d3 'it'h3 1 5 .l:.e4 'it'd7 . A s Nunn points out, it i s as if Black is claiming that White has nothing better than 1 6 l:te l , and indeed 1 6 . . . 'it'h3 would repeat the

REFINED ROOK-LIFT:

position. Nevertheless, 1 5 . . :it'd7 i s not s o bad and White must play carefully to get anything at all. 16 ft)d2 .i. b7 Black wants to play . . . c5, when he hopes that the pressure on White's kingside will offer him sufficient play. The direct 16 . . . f5 ? ! looks insufficient after 1 7 .l:.e l h8 ( 1 7 . . . f4 1 8 'iWh5 ! �h8 19 ft)e4 ± gives White the initiative in addition to the extra pawn), and here: a) After 1 8 f4, 1 8 . . . 'iWa7+ 1 9 d4 was played in Liang linrong-Peng Xiaomin, Chinese Ch 1 992, but this looks like a poor version of the main lines with 1 2 d4 for Black. Therefore he should prefer 1 8 . . . .i.b7 or even 1 8 . . . .l:.e8 1 9 .l:.xe8+ 'iWxe8 intending a standard manoeuvre, . . . .l:.a7 ! . b ) 1 8 'iWh5 i s more enterprising. 1 8 . . . ft)f6 1 9 'iWh3 ! ( 1 9 'iWh4 ? ! c5 20 ft)f3 .i.b7 2 1 ft)g5 h6 22 .i.d2? 'iWc6 23 f3, Benjamin-Kamsky, USA Ch, Los Angeles 1 99 1 , and now 23 ... ft)h 7! wins) 1 9 . . . c5 20 ft)f3 and now : b l ) 20 . . . .i.b7 ? ! 2 1 ft)h4 ! ft)h5 was played in Kotronias-I.Sokolov, Caleta 2009. Here Mikh­ alevski recommends 22 .i.d2 ! with the idea .i.d l . b2) Mikhalevski proposes 20 . . . 'iWc6 ! 2 1 ft)g5 (2 1 ft)h4 c4 ! 22 .i.d l f4 ! 23 g4 lIe8 24 .l:.xe8+ "xe8 25 .i.d2 lIa 7! with counterplay) 2 1 . . ..i.b7 22 f3 c4 23 dxc4 bxc4 24 Ji.c2 :ae8 25 lIxe8 'ifxe8 26 �n h6 27 Ji.xf5 ft)h5 ! and Black has a very active position. 17 l:.el 1 7 ft)f3 commits the knight rather early and 1 7 . . . c5 gives Black good compensation. 17 'ifn c5 1 8 .l:.e l �h8 (D) also gives Black good play :

w

12 d3

97

a) 1 9 .i.d l f5 ( 1 9 . . . l:.ae8 ! ?) 20 .i.f3 f4 2 1 g4 (2 1 'ifg2 .l:.ae8 22 .l:.xe8 'ifxe8 23 ft)e4 looks more challenging) 2 1 . . . .l:.ae8 was rather unclear in Leko-Adams, Tilburg 1 997, although Black looks to have reasonable play. b) 1 9 ft)e4 .i.e7 20 a4 h6 2 1 axb5 axb5 22 :xa8 .i.xa8 23 f3 f5 24 ft)f2 was Polaczek­ Sirota, corr. 2000. Here 24 . . . h4 ! ? gives Black counterplay. c) 1 9 a4 ft)f4 ! ( 1 9 . . . f5 ? 20 axb5 axb5 2 1 :xa8 .i.xa8 2 2 Ji. d 1 g 5 2 3 ft)f3 g4 24 ft)e5 'ifb7 25 f3 ± Ivanchuk-I.Sokolov, Amsterdam 1 996) 20 ft)e4 ? ! (after 20 f3 Black can choose be­ tween 20 . . . c4 and 20 . . . ft)xd3 2 1 'ifxd3 c4, in both cases with unclear play) 20 . . . ft)h3+ 2 1 h l was Z.Almasi-Khalifman, Ubeda 1 997. Now the most accurate sequence is 2 1 . . .ft)xf2+ ! 22 'ifxf2 f5 23 'iWg2 c4 ! 24 .i.c2 cxd3 25 .i.xd3 fxe4 26 .i.xe4 ':'ae8, when 27 .i.d2? ':'xe4 28 ':'xe4 .i.c5 29 .i.f4 fails to 29 ... ':'f7 ! intending . . . lIe7. 17 c5 18 ft)e4 Ji.e7 (D) •••

Black has a very nice-looking position, but he remains a pawn down. Now the passive 1 9 f3 �h8 20 .i.d2 (20 d4 c4 2 1 .i.c2 .l:.ae8 22 ft)g5 is rather strange and after 22 . . . h6 23 ft)e4 f5 24 ft)f2 f4 Black had good play in Simmelink­ Bennedik, corr. 1 998) 20 . . . f5 21 ft)f2 f4 22 ft)e4 fxg3 23 hxg3 'iWh3 24 'iWe2 c4 ! 25 .i.c2 (25 dxc4 ft)xc3 ! 26 bxc3 .i.xe4 intending . . . .i.c5+ ) 25 . . . cxd3 26 .i.xd3 ':'ad8 27 'ifn 'ifh5 28 g4 'ifh4 29 ':'ad l ft)f4 30 .i.xf4 ':'xf4 gave Black a strong initiative in Leko-I.Sokolov, Groningen 1 995, so White usually chooses between: 98 C I : 19 Ji.g5 C2: 19 84 99

UNDERSTANDING THE MARSHALL A TTA CK

98

These moves are closely related because 1 9 a4 i s invariably met b y 1 9 . . . b4 2 0 .i.gS, but there are some interesting differences.

26 . . . tifg6 is given by I.Sokolov, but then 27 .i.e7 ! looks good for White. Black could try to improve with 26 . . . tlfbS . 24 c4 25 .i.c2 b4 26 dxc4 tifxc4 27 ':'e5 lLlxc3! 28 .i.xf5 (D) •••

Cl) 1 9 .i.g5 (D)

B

19 f6 1 9 . . . .i.xgS ? is bad: 20 lLlxcS ! �c7 2 1 lLlxb7 tifxb7 (2 1 . . .lLlf6 22 �f3 :ab8 23 tiffS ! .i.h6 24 lLlcs +-) 22 tiff3 ! (worse is 22 l:.eS ? ! lLlxc3 23 bxc3 .i.f6 24 ':cS l:.ac8 with compensation) 22 . . . ':ad8 23 ':eS wins back the piece and leaves White two pawns up. 20 .i.d2 20 .i.e3 has been played too, but the bishop is more vulnerable here, not just to the dS­ knight but also to . . . fS-f4. 20 /oith8 21 �h5 White has tried other moves as well: a) 21 c4 gave Black excellent play after 2 1 . . .lLlb4 22 .i.xb4 cxb4 23 cxbS fS 24 lLld2 axbS 2S :e6 .i.f6 26 lLlf3 f4 + in Svidler­ I.Sokolov, Groningen 1 996. b) 21 d4 ! ? c4 (2 1 . . .cxd4 22 cxd4 looks in­ sufficient for Black) 22 .i.c2 :ae8 (22 . . .fS 23 lLlgS .i.xgS 24 .i.xgS 'ii'c 6 is another idea) 23 f3 (White could try for more with 23 tifhS ! ?) 23 . . . fS 24 lLlcs .i.xcs 2S dxcS 'ii'c 6 (now Black wins back the pawn and has no problems) 26 ':xe8 ':xe8 27 'ii'f l lLlf6 28 ':e l tifxcS+ 29 'ii'f2 .l:r.xe l + 30 .i.xe l �dS = Acs-Pinter, Hungarian Team Ch 1 997/8. 21 f5 22 lLlg5 .i.xg5 23 .i.xg5 �c6 24 f3 24 tiff3 c4 (24 . . . tifd7 ! ? is a possible im­ provement) 2S dxc4 bxc4 26 .i.c2 and now •••

This moment has gone by without comment by various annotators, but this is a critical posi­ tion. 28 . . . .l:r.xfS ? 29 ':xfS lLle2+ 30 �h l was Emst-I.Sokolov, European Team Ch, Debre­ cen 1 992. Here Sokolov gives 30 . . . �g8 3 1 b3 'iWe6 32 'iWg4 g6(?) 33 :f8+ �xf8 34 �xe6(??) .i.xf3# but instead 34 �xb4+ �g8 3S 'iix b7 should win for White. Even 3 1 .i.h6 ! ? is good because 3 1 . . .gxh6 32 'ii'xh6 .i.xf3+ 33 l:.xf3 'ii'd S 34 :afl lLld4 3S 'ii'f4 lLlxf3 36 tifxf3 leaves White a pawn up for nothing. Fortu­ nately, Black has a better move. 28 lLle2+! 29 ':xe2 l:Ixf5 30 .l:r.e8+ .l:r.f8 (D) •••

••

•••

w

White's kingside is weak and the b7-bishop is very strong. Black has excellent compensa­ tion for the pawn.

REFINED ROOK-LIFT:

12 d3

99

e2) 19 a4 b4 20 .1t.gS (D)

Throwing in the queenside pawn moves soft­ ens up the queenside a bit (for both sides) and creates some new tactical possibilities. 20 .1t.xgS This move failed before, but it works now, as we shall see. Playing by analogy with Line C l looks worse now, because after 20 . . . f6 2 1 .1t.d2 Black has lost the possibility of playing a timely . . . c4. 21 lLlxgS 2 1 lill.c S? is wrong now: 2 l . . .'ifc7 ! (2 l . ..'ifc6? is worse as after 22 lLlxb7 neither 22 . . . 'ifxb7 23 'iff3 .:tad8 24 c4 ± nor 22 . . .bxc3 23 'ii'f3 cxb2 24 .:tabl lLle7 2S 'ii'x c6 lLlxc6 26 %hb2, with an extra pawn for White, is satisfactory) 22 lLlxb7 lLlf6 ! 23 'iff3 l:tab8 24 'ii'fS :xb7 2S 'ii'x gS bxc3 =t. The b-file is opened for Black's rook with great effect - this is the difference be­ tween this and the position without a4 and . . . b4 thrown in. 21 bxc3 22 bxc3 (D) 22 h6 Black kicks away the knight and creates some tuft for his king. Other moves: a) 22 ... lLlxc3 is the first move that comes to mind, but 23 'ifhS h6 24 lLlxf7 l:txf7 2S .1t.e6, with an attack, is given by Adams. Indeed after 2S . . . 'ii'e7 (2S . . . 'ii'c 7 26 .:ta3 ! corrals the knight and leaves White a good pawn up) 26 l:tac 1 lLlxa4 27 .:tc4 ! Black has big problems. b) 22 . . . 'ifc6 23 'iff3 .:tad8 24 d4 (24 'ife4 and 24 .:tac t are also appealing) 24 . . . cxd4 2S cxd4 h6 26 lLle4 i.. a8 27 .:tad I 'ifb6 28 lLlcs •••

•.•

•••

lLlf6 29 'irfS l:d6 30 .:te7 ! led to complications that favoured White in Leko-Slobodjan, FIDE Knockout, Groningen 1 997. c) 22 ... .:tad8 23 lLle4 'ii'c6 24 'ii'h s lLlxc3 2S 'ifxcs .:txd3 26 .1t.c4 'ii'x cs 27 lLlxcs l:td4 28 i.. f 1 ! was J.Polgar-Adams, Tilburg 1 997. This is identical to the main line, except Black has not even played the useful . . . h6 here. 23 lLle4 'ii'c6 24 'ifhS .:tad8 25 'ii'f5 White physically prevents . . . fS . Instead, 2S as ? ! fS 26 lLld2 �h8 gives Black counterplay, and 27 lLlf3 ?? lLlf6 0- 1 was the sudden finish of Almasi-Adams, European Team Ch, Pula 1 997. 25 lLlxc3 26 'ii'xc5 l:txd3 27 i.. c4 'ifxc5 28 lLlxc5 .:td4 29 .1t.fl ! White maintains the initiative in the ending, Z.Almasi-Tseshkovsky, Yugoslav Team Ch, Niksic 1 997. •.•

D) 15 'ii'f5 (D) •••

1 00

UNDERSTANDING THE MARSHALL A TTACK

This is the most popular continuation. Black retreats his queen and intends to develop har­ moniously and put pressure on the d3-pawn with .. :ii'g6 and . . . .i.f5 . 16 tLld2 'ili'g6 (D) As Nunn points out, 1 6 . . . tLlf6 1 7 .:te l 'ili'xd3 1 8 tLle4 'ili'xd 1 1 9 tLlxf6+ gxf6 20 .l:.xd 1 (20 .i.xd 1 intending .i.f3 is good too) gives White the better endgame.

Now White has: 0 1 : 17 a4 02: 17 tLln 03: 1 7 :te l

17 tLlf6 Other moves look even worse: a) 1 7 . . . .i.f5 ? ! 1 8 axb5 ! axb5 ( 1 8 . . . .i.xe4 1 9 dxe4 ±) 1 9 .l:.xa8 .l:.xa8 20 .l:.e 1 .i.xd3 2 1 tLlB ! (Zapata) threatens both 22 tLlh4 and 22 tLld4. b) 1 7 . . . f5 ? ! 1 8 .l:.d4 ! f4 1 9 tLle4 .i.g4 20 'ili'fl (not 20 B ? ! .i.xf3 2 1 'ili'xB fxg3 22 'ili'd 1 ? g2, when Black wins after either 23 :txd5 cxd5 24 .i.xd5+ �h8 25 .i.xa8 .i.xh2+ 26 �xh2 lIfl or 23 .i.g5 :tfl + 0- 1 Rudak-Voinov, Ukrainian Ch, Alushta 2005) 20 . . . .i.e5 2 1 ':xd5 cxd5 22 .i.xd5+ �h8 23 .i.xa8 fxg3 24 tLlxg3 lIxa8 25 axb5 +- Zapata-Pavlovic, GMA Open, Bel­ grade 1 988. 18 .l:.e 1 .i. g4 Again, grabbing the pawn with 1 8 . . . 'ili'xd3 ? is bad in view of 1 9 .i.c2 'ili'd5 20 tLle4 'ili'xd 1 2 1 tLlxf6+ gxf6 2 2 .l:.xd 1 ± with a much better end­ ing. 19 f3 .i.h3 (D) •••

1 00 101 1 02

Instead, 1 7 tLlB is well met by 1 7 . . . .i.g4, and 17 c4 tLlf6 leaves White's position looking rather loose.

01) 17 a4 (D) This is uncommon, but rather dangerous.

20 �hl ! ? A n interesting prophylactic idea. 2 0 tLlfl ? ! i s well answered by 20 . . . :tae8 (Azarov's idea 20 . . . .i.xfl ? ! 2 1 'iti>xfl .i.xg3? 22 hxg3 'ii'x g3 falls short after 23 .l:.e3 !) 21 .i.e3 h5 22 d4 h4 23 .i.c2 (D), and now: a) After 23 . . . .i.f5 24 .i.xf5 'ii'xf5 25 axb5 axb5 26 :ta6 Nunn gave 26 . . . 'ii'd 5 as unclear, while 26 . . . 'ili'c8 intending . . . tLld5 also looks OK. b) 23 . . . 'ili'h5 ! ? 24 .i.f2 (24 g4 tLlxg4 ! 25 fxg4 .i.xg4 and both 26 'ili'd3 .l:.e6 and 26 'ili'd2 .i.B 27 h3 f5 give Black an attack) 24 . . .l:he 1 25 .i.xe 1 .i.xfl 26 g4 (26 'it;xfl hxg3 27 hxg3 { 27 .i.xg3 .i.xg3 28 hxg3 'ili'h 1 + gives Black a winning attack } 27 . . . 'ili'h3+ 28 �gl .i.xg3 29

REFINED ROOK-LIFT:

12 d3

1 01

B

�xg3 'iVxg3+ 30 ..th 1 'iVh3+ 3 1 ..tg 1 l:.eS -+) 26 . . . �xg4 27 fxg4 'iVdS -+ Zontakh-Pavlovic, Podgorica 1 99 3 . 20 'iVh5 (D) Instead, 20 . . . 'iVxd3? 2 1 �c2 'iVdS 22 �e4 is a bad ending for Black. Other reasonable moves are difficult to find. For example, the natural 20 . . . l:.aeS? ! 2 1 axbS axbS 22 IIxeS l:.xeS 23 �e4 gives White a clear advantage. •••

17 f5 (D) This is probably best. Alternatives: a) 17 ... hS IS a4 .ig4 19 'iVe l �f6 20 l:.e3 l:.aeS 2 1 axbS axbS 22 d4 h4 gave Black some play in A.Kuzmin-Shulman, Minsk 1 9S6, but this does not look trustworthy to me. b) 1 7 . . . �f6 I S l:.e l .ig4 1 9 f3 .ifS 20 d4 l:.aeS 2 1 .ie3 �dS 22 .ixdS cxdS 23 'iVd2 .id3 24 l:.ad 1 .ic2 2S l:.c 1 .id3 26 l:.cd 1 .ic2 27 l:tc 1 �d3 2S �f4 �xf4 29 'iVxf4 .ixfl 30 'it>xfl 'iVd3+ 3 1 �f2 h6 32 IIcd l 'iVc4 33 a3 is much better for White, Wedberg-Pinter, Han­ inge 1 9S 5 . c) 1 7 . . . �fS is natural but has hardly been played. After I S l:.d4 .ics 1 9 :xdS cxdS 20 d4 .ib6 2 1 .ixdS l:.adS 22 'iVf3 .ig4 23 'iVg2 l:.feS 24 �e3 'iVd3 ! Black had good play in D .Hansson-Hebden, Lloyds B ank Masters, London 1 9S4. White could try I S .ixdS cxdS 19 l:.d4 with the idea 20 �e3 , putting pressure on the dS-pawn. •••

Now: a) 2 1 'iVe2 l:.aeS 22 �e4 .ixg3 ! 23 hxg3 .ie6+ 24 ..tg l �xe4 2S fxe4 'iVxe2 26 l:.xe2 .ixb3 27 axbS 1/2- 1/2 Goloshchapov-Azarov, Cappelle la Grande 2006. b) Azarov gives 2 1 axbS axbS (2 l . . .�g4?! 22 :te2 �xg3 23 'iVg l ! �f2+ 24 IIxf2 .ixf2 2S 'iVxf2 axbS 26 l:.xaS l:.xaS 27 'iVg3 'iVfS 2S �c2 ±) 22 %baS l:.xaS 23 'iVe2 ;t.

02) 17 �n (D)

IS l:.d4

UNDERSTANDING THE MARSHALL A TTACK

1 02

White keeps his rook in the centre to try to fight for the initiative. 1 8 ':e l f4 gives Black good attacking chances. 18 f4 Black should waste no time. 1 8 . . . �h8 ? ! al­ lowed White to make a strong exchange sacri­ fice with 1 9 .i.xdS cxdS 20 .i.f4 �cS 2 1 ':'xdS 'ii'c 6 22 ':xcs 'ii'x cs 23 d4 ± in Nijboer-Van der Sterren, Wijk aan Zee 1 989. 19 ':xd5 cxd5 20 �xd5+ .i.e6 21 .i.xa8 ':xa8 (D)

B

•••

w

22 'ii'n After 22 a4 Black could try either 22 . . . b4 or 22 . . . .i.g4. 22 ':f8 23 'ii'e4 .i.f5 24 'ii'd 5+ �h8 25 a4 This was Timman-Hubner, Tilburg 1 985 . Af­ ter 2S . . . .i.xd3 26 axbS �e4 27 'ii'd4 fxg3 28 lLlxg3 �xg3 29 fxg3 axbS Hubner assesses the position as equal. Both sides must take a little care though, Black because of his pawn deficit and White because his king is a bit loose. •••

Knockout, Groningen 1 997. With 'ii'd6 coming next, Black' s position is critical. b) 1 8 . . . .i.hS looks better: 19 lLle4 ':ae8 20 �g2 (after 20 lLlxd6 .i.xf3 ! Lalit gives 20 g4 fS 2 1 lLlxd6 fxg4 ! +) 20 . . . �c7 (20 . . . .i.b8 should be similar, but after 2 1 .i.d2 fS 22 lLlf2 �h8 23 �xd5 cxdS 24 d4 f4? { 24 . . . ':e4 ! ? } 2S g4 .i.xg4 26 lLlxg4 hS 27 'ifbl 'ii'g S 28 h4 'ii'xh4 29 lLlh2 White was close to victory in Kr.Georgiev­ Bryson, Dubai Olympiad 1 986) 2 1 .i.d2 �h8 22 lLlf2 fS 23 .i.xdS cxdS 24 ':xe8 ':xe8 2S d4 'ii'e6 26 'ii'f 1 'ii'e 2 27 'ii'xe2 lhe2 28 :d l h6 29 h3 gS was Mekhitarian-Greenfeld, Sao Paulo 2008. Black has typical compensation in the ending and the game was soon drawn. We now return to 1 7 . . . fS (D):

03) 17 ':el (D) 17 f5 1 7 . . . 'ii'x d3 ? loses to 1 8 .i.c2, trapping the queen, but 1 7 . . . .i.g4 is an alternative. After 1 8 f3 Black has: a) 1 8 . . . .i.h3 ? ! 1 9 lLle4 :ae8 ( 1 9 . . . .i.c7 20 �e3 l:tae8 2 1 'ii'd 2 was also clearly better for White in Svidler-Adams, FIDE Knockout, Gro­ ningen 1 997) 20 :e2 ! hS 2 1 .i.gS .i.c7 22 'ii'd2 �h8 23 llae l f6 24 lLlf2 ':xe2 2S ':xe2 iHS 26 .i.xdS cxdS 27 .i.f4 ! .i.xf4 28 'ii'xf4 �xd3 29 ':d2 was played in Anand-Adams, FIDE •••

Black intends . . .f4 with an attack. White can block the f-pawn with 1 8 f4, ignore it with 1 8 a4, or force matters by 1 8 c4. 1 03 D31 : 18 f4 1 04 D32: 18 a4 1 06 D33: 18 c4

REFINED ROOK-LIFT:

Other moves are less challenging: a) 1 8 �e4? is bad in view of 1 8 . . . fxe4 1 9 dxe4 -ltg4 20 'ifd4 ? ! llae8 2 1 �d2 'ifi>h8 ! with the idea 22 exd5 c5 -+. We shall see an im­ proved version of this idea for White in Line 033 . b ) 1 8 �f3 ? ! allows Black to develop his ini­ tiative very naturally. 1 8 . . . f4 19 �e5 -ltxe5 20 ':xe5 fxg3 2 1 hxg3 (2 1 fxg3 �g4 22 'ifc2 ':ae8 23 iLf4 iLe6 ! gives Black good play) 2 1 . . .iLg4 22 'ife l (22 'iffl l:.ae8 23 l:.g5 'ife6 and now 24 c4? ! iLh3 ! 25 cxd5 'iff6 ! gives Black a strong attack and 24 iLf4 iLf3 ! also leaves Black with a strong initiative) 22 . . . -ltf3 23 -ltd2 llae8 was good for Black in Leko-Svidler, Dortmund 1 998. c) 18 'iff3 'ifi>h8 19 -ltd 1 (too greedy is 1 9 -ltxd5 ? ! cxd5 20 'ifxd5 llb8 intending . . . iLb7) 19 . . . f4 20 g4 h5 21 h3 �f6 22 'ifg2 (22 'ifxc6? ! hxg4 23 'ifxa8 f3 24 h4 gives Black a strong attack after either Adams 's 24 . . . g3 or LaliC 's 24 ... �f5) 22 ... hxg4 23 hxg4 iLxg4 24 ':e6 'ifh5 (D) and here:

w

c 1 ) 25 ':'xf6? -ltxd l 26 llxd6 llae8 27 �e4 iLf3 28 'ifh2 ':'e5 -+ is given by Adams. After 29 'ifxh5+ (or 29 iLxf4 llxf4 30 �g3 llh4 and again Black wins) 29 . . . llxh5 30 �g3 fxg3 3 1 fxg3 llh l + 3 2 �f2 llh2+ 3 3 �e3 c5 ! White gets mated - Lalic. c2) 25 -ltxg4 �xg4 26 llxd6 (or 26 'ii'h l 'ifxh l + 2 7 �xh l llf6 2 8 llxf6 �xf6 2 9 f3 in­ tending �e4 = Lalic) 26 . . . llae8 27 �e4 �e5 28 f3 ? ! (28 �d2 �f3+ 29 �fl �h2+ 30 �gl �f3+ =) 28 ... �xf3+ 29 'ifi>f2 �h4 30 'ifh l g5 gave Black a very strong attack in J.Polgar­ Adams, Dos Hermanas 1 999.

12 d3

1 03

031) 18 f4 (D)

We have seen this blockading idea before, but the situation is different here because Black can capture on f4. At first it seems as if this move is just an oversight, but White hopes that he can use the time Black spends establishing material equality to grab the initiative, mainly by getting a firm grip on the dark squares. 18...iLxf4 19 'iff3 After 1 9 �f3 Black should avoid retreating with 1 9 . . . iLc7 20 �e5 iLxe5 2 1 llxe5 -lte6 22 c4 bxc4 23 dxc4 �b6 24 -lte3 llab8 25 'ifd6 iLf7 26 'ifxg6 hxg6 27 llc5 ± (Lalic) and play either the solid 1 9 . . . -ltxc l 20 llxc 1 f4 2 1 �e5 'ifh6 22 'iff3 fxg3 23 'ifxg3, which looks a little better for White, or the speculative 1 9 . . . iLxg3 ! ? 20 hxg3 'ifxg3+ 2 1 �h l f4. 19...�b8! This retreat looks funny at first but there is a tactical reason that Black has lined up all of his pieces on the back rank. 20 iLxdS+ cxdS 21 �b3 (D) This is consistent and clearly best. White wants to play -ltf4. Grabbing material with 2 1 'ifxd5+ �h8 2 2 'ii'x a8 ? just rebounds after 22 . . . 'ii'b 6+ 23 d4 -ltb7 (this is the real point be­ hind 1 9 . . . iLb8 ! - the rook is not hanging on f8) 24 ':e6 'ifc7 +. And after 21 �fl -ltb7 22 ':e7 iLc6 23 'ife3 (Popovic-Pavlovic, Yugoslav Team Ch, Cetinje 1 993) Black should play 23 . . . ':f6 ! intending . . . -ltd6 + (23 . . . d4 also looks good). 21 'iff7 Also possible is 2 1 . . . -ltb7 22 -ltf4 -lta7+ (White is much better after 22 . . . -ltxf4? 23 'ii'xf4 •••

1 04

UNDERSTANDING THE MARSHAU A TTA CK

b2) 23 . . . �xf4 attempts to keep some life in the position. 24 tbxd4 (24 'ii'f3 �e3+ 25 %be3 �b7 26 'ii'f4 dxe3 27 tbc5 �aS with counter­ play) 24 . . . .i.b7 25 '6'xf8+ �xf8 26 tbe6+ �gS 27 tbxf4 g5 2S tbe6 'ii'h 5 29 tbd4 f4 gives Black sufficient counterplay. 22 i.a7 23 i.f4 i.d7 24 l:te5 b4! ? 25 l:tae1 l:taeS 26 ..tr>n bxc3 The game was drawn here in Leko-Adams, Linares 1 999. After 27 bxc3 i.xd4 2S cxd4 l:txe5 the position is equal. If White tries for something with 29 dxe5 , then 29 . . . d4 ! opens a diagonal for Black's queen and bishop, and it is White who should be careful. •••

d4 24 tbxd4 ':aeS 25 ':xeS 'ii'x eS 26 'ito>f2 ! ± Kotronias-Beliavsky, European Ch, Istanbul 2003) 23 d4 l:taeS 24 'ii'f2 lIe4 25 tbd2 and now instead of 25 . . . h6 26 tbxe4 dxe4 (Am.Rod­ riguez-L.Perez, Cuban Ch, Matanzas 1 995), Black should just play 25 ... .:eeS, when it is not so clear that he is really any worse. If White tries to get his knight to e5, Black will be able to play . . . .:e4 again. 22 tbd4 After 22 iLf4 d4 ! (D) White has:

03 2) IS a4 (D)

IS lIbS Black should take a move in order to secure his queenside because I S . . . f4 1 9 tbe4 is annoy­ ing. Opening the a-file should favour White of course, but Black's rook is much less vulnera­ ble to tactical ideas now that it is off the h i -aS diagonal. Black does have an alternative in I S . . . �hS ! ?, ignoring the threat of axb5 . White has: a) 1 9 �xd5 cxd5 20 axb5 f4 2 1 tbfl looks risky after 2 l . . .d4 with the idea . . . �b7. b) 1 9 tbf3 f4 20 .i.xd5 (20 tbe5 'ii'f6 2 1 iLxd5 fxg3 ! , as pointed out by Pavlovic, gives Black a strong attack) 20 . . . cxd5 2 1 tbe5 iLxe5 22 ':xe5 fxg3 23 hxg3 i.g4 gives Black com­ pensation. Pavlovic gives the funny line 24 'ii'f l l:taeS 25 i.f4 l:txe5 26 i.xe5 i.f3 27 'ii' h 3 'ii'f5 2S 'ii'xf5 l:txf5 29 g4 l:txe5 30 axb5 h5 ! with a •••

a) 23 tbc5 iLxf4 24 gxf4 (24 'ii'xf4 dxc3 25 bxc3 �b7 =) 24 . . . l:ta7 25 cxd4 �b7 gives Black compensation. If 26 '6'f2, then 26 . . . �d5 is fine for Black, while 26 tbxb7 :'xb7 27 d5 l:td7 2S lIe6 ':fdS 29 ':ae l ':xd5 30 ':eS+ ':xeS 3 1 :'xeS+ 'ii'xeS 3 2 'ii'x d5+ 'ii'f7 is level. b) 23 'ii'x aS gives Black a choice: b l ) 23 . . . �b7 is sufficient. 24 'ii'x bS ':'xbS 25 �xbS 'ii'd5 26 l:te2 'ii'f3 ! 27 lId2 (27 tbxd4 'ii'h l + 2S �f2 'ii'xal ) 27 . . . 'ii'h l + 2S �f2 'ii'xh2+ 29 �fl 'ii'h l + with a draw.

REFINED ROOK-LIFT:

strong attack, because 3 1 bxa6 will get White mated after 3 1 . . .hxg4. c) 1 9 axb5 must be critical, when 1 9 . . . lLlf4 ! is PavloviC's idea - Black continues the attack at all costs. Then 20 lLlc4 lLlh3+ 2 1 �g2 f4 22 lLlxd6 'it'xd6 is critical. 23 f3 (23 'ii'f3 iLd7 24 ':xa6 ':'xa6 25 bxa6 fxg3 26 'ii'x g3 ':'xf2+ 27 �h l 'iWf8 28 iLe3 :f3 29 'ii'g 2 ':'f2 30 a7 ':'xg2 3 1 �xg2 'it'a8 32 �g3 c5 gives Black enough play) 23 . . . cxb5 24 g4 iLxg4 ! 25 fxg4 lLlg5 26 d4 f3+ 27 �h l lLlh3 28 iLe3 ':'ae8 29 'ii'd 2 'ii'g 6 30 iLc2 'ii'xg4 and Black is not worse - Pav­ lovic. If we continue with the forced sequence 3 1 ':'g l 'ii'h 5 32 ':'g3 'ii'd 5 33 ':'fl f2+ 34 lIg2 ':'f3 , White must find the desperate resource 35 iLh6 ! to have a chance of holding the position. 19 axb5 axb5 (D)

1 2 d3

1 05

b4, utilizing the b8-rook, gave Black coun­ terplay in Copar-Bohak, corr. 1 993) 2 l . . .h5 22 'it'g2 f4 23 lLle4 f3 24 'it'fl lLlf4 25 iLxf4 iLxf4 26 �h l iLh6 27 lLlc5 iLf5 28 d4 b4 29 l:te7 bxc3 30 bxc3 'ii'g4 3 1 ':'aa7 ':'fe8 32 ':'ab7 ':'a8 33 ':'a7 ':'ab8 112- 112 Lastin-Smimov, Russian Ch, Krasnodar 2002 seemed dynamically bal­ anced throughout. d) 20 ':'a7 �h8 (or 20 . . . iLc5 2 1 ':'c7 iLb6 22 ':'ce7 f4 23 d4 fxg3 and now both 24 fxg3 iLg4 and 24 hxg3 iLg4 leave the position un­ clear) 2 1 lLlf3 f4 22 lLle5 iLxe5 23 ':'xe5 fxg3 24 hxg3 iLg4 25 'ii'd 2? ! lLlf4 26 gxf4 iLf5+ 27 �h2 'iWh5+ 28 �g2 'ii' h 3+ 29 �gl 'ii'g4+ 30 �fl 'it'h3+ 3 1 �e l ? (funny things can happen in rapidplay ; 3 1 �g l = ) 3 1 . . .'ii' h l + 32 �e2 iLg4+ 33 �e3 ':'xf4 34 l:.e4 c5 35 �xf4 'ii'f3+ 36 �e5 'ii'f6+ 37 �d5 ':'d8+ 3 8 �xc5 'ii'd 6+ 39 �xb5 ':'b8+ 40 �a4 'ii'b 6?? (40 . . . iLd7+ is mate after 41 ':'xd7 'ii'a6# or 41 �a5 'ii'b6#) 4 1 ':'e8+ ':'xe8 4 2 'ii'e 3 and White even went on to win in Morozevich-Grishchuk, Dubai rapid 2002 ! 20 fxe4 21 dxe4 iL g4 22 'ii'd4 iLf3 22 . . . .:.be8 23 ':'a6 ! causes trouble - this is the point of White throwing in 1 8 a4. 23 exd5 (D) •••

B 20 lLle4! This idea of Anand's is the point of White's play. Other moves are less testing: a) 20 c4 f4 ! gives Black good play : 2 1 lLle4 (2 1 cxd5 ? fxg3 22 fxg3 iLxg3 23 �h l iLg4 24 :e6 iLxe6 25 dxe6 ':f2 ! -+ Lalic) 2 l . . .fxg3 (2 l . . .bxc4 is also possible) 22 fxg3 iLg4 23 'ii'c 2 bxc4 24 dxc4 lLlb4 25 'ii'd 2 iLe5 26 lLlf2 iLd l ! intending . . . iLxg3 is one amusing idea. b) 20 lLlf3 f4 2 1 lLle5 iLxe5 22 ':'xe5 fxg3 23 fxg3 iLg4 24 'ii'e l iLh3 25 iLe3 (Lalie's sugges­ tion 25 ':'g5 can be met by 25 . . . .:.be8 ! 26 'it'd l 'ii'f7, when Black has the initiative; for exam­ ple, 27 iLf4 h6 28 ':'h5 iLe6 ! ) 25 . . . .:.fl + 26 'ii'x fl iLxfl 27 ':'xfl ':'f8 28 ':'xf8+ was agreed drawn in Anand-Adams, Dos Hermanas 1 999. c) 20 'ii'f3 �h8 21 h4 (after 21 iLxd5 cxd5 22 'ii'xd5 iLb7 intending ... f4 Black gets a strong attack, while 2 1 'ii'g 2 f4 22 lLlf3 fxg3 23 hxg3

23 c5 Black has invested a second pawn in the at­ tack, but White's b3-bishop is cut off and . . . c4 is a possibility, stranding the d5-pawn. Of course, White's kingside is vulnerable too. 24 'it'h4 ':'be8 25 iLd2 Other moves are also possible, but Black has just enough play to hold the balance: a) 25 ':'xe8 ':'xe8 and now : •••

UNDERSTANDING THE MARSHAll A TTA CK

1 06

a 1 ) 26 i.e3 'ii'd 3 is tough to meet, because 27 %le I %lxe3 ! 28 fxe3 'ii'd2 wins for Black Anand. a2) 26 i.f4 and now 26 . . . c4 27 i.xd6 cxb3 , with compensation, is given by Anand, while 26 . . . 'ii'e4 27 l:tn c4 28 i.a2 (28 i.xd6 'ii'xh4 29 gxh4 cxb3 leads to an equal ending) 28 . . . i.c5 ! gives Black good counterplay. Instead 28 . . . i.h 1 29 'ii'h 3 i.xf4 30 i.b1 '±' is given by Lalic, al­ though 30 . . . 'ii'f3 probably leaves White with nothing more than perpetual check. b) 25 i.e3 l:te5 ! is given by Anand (instead 25 . . . 'ii'f5 ? ! 26 %lac 1 i.e4 27 i.d 1 was Anand­ Adams, Dortmund 2000, when Black doesn't have enough for the remaining pawn after taking one back on d5) . Then 26 'ii'h 3 c4 27 i.d 1 (27 i.a2? .l:.h5 ! is very good for Black) 27 . . . i.xd5 28 'ii'g4 (28 l:ta6 'ii'e4 29 f3 l:txf3 30 i.xf3 'ii'xf3 3 1 l:txd6 'iVh 1 + 32 �f2 'ii'f3+ is a draw­ ing line given by Anand) 28 . . . %lxe3 ! 29 ':xe3 (or 29 fxe3 'ii'f6 30 'ii'e 2 i.c5 with excellent play) 29 . . . 'ii'f7 30 'ii'e2 i.c5 3 1 i.c2 i.xe3 32 fxe3 i.f3 should give Black enough counterplay. 25 i.e4 26 %le2 'ii'f5 Threatening . . . 'ii'f3 . 26 . . . c4 ! ? is also possible. 27 i.f4 c4 28 l:txe4 l:txe4 29 i.c2 i.xf4 30 i.xe4 'ii'xe4 31 gxf4 112- 112 Anand-Khalifman, FIDE Knockout, New Delhi 2000.

18 f4! 19 lLle4 White must cover g3. Accepting the piece gets White throttled: 1 9 cxd5 ? fxg3 20 dxc6+ �h8 2 1 hxg3 i.xg3 22 fxg3 'ii'x g3+ 23 �h 1 i.g4 24 l:te2 i.xe2 25 'ii'xe2 l:tf2 is winning for Black. 19 fxg3 20 fxg3 This looks the safest. Recapturing with 20 hxg3 ? ! i.g4 is similar, but White cannot defend along the second rank, and the f2- and h 1 squares are more vulnerable, while the ugly 20 lLlxg3? is good for Black after 20 . . . i.g4 2 1 'ii'c 2 (2 1 'ii'd2 i.b4) 2 l . . .lLlb4. 20 i.g4 21 'ii'c2 bxc4 22 dxc4 (D) 22 'ii'x c4 i.c7 23 'ii'c 2 �h8 24 i.xd5 cxd5 25 lLlf2? (25 'ii'x c7 is better, although Black has excellent compensation for the pawns af­ ter 25 . . . dxe4 26 dxe4 %lac8) 25 . . . %lac8 (even 25 . . . i.xg3 26 hxg3 i.d 1 ! is good for Black) 26 'ii'a4 i.d7 27 'ii' h4 was played in Hellers­ Wahls, World Junior Ch, Adelaide 1 98 8 . Now 27 . . . i.b6 ! 28 i.e3 l:tc2 ! gives Black a killing initiative. •••

•••

•••

•••

033) 18 c4 (D)

22 l:tae8 Tactical fishing with 22 . . . i.b4? is nonsense. After 23 i.d2 lLlf4 24 i.xf4 i.xe 1 25 lhe 1 i.f5 the cleanest kill is 26 c5+ (26 'ii'g 2 was played in Smagin-Hebden, Moscow 1 986 and here 26 . . . %lae8 gives Black some hope, but as we can see it is not terribly relevant because 26 c5+ just wins for White) 26 . . . �h8 27 i.e5 ':ae8 28 'iWc3 ! . After 28 . . . i.xe4 29 l:txe4 (Livie-Kris­ tjansson, COTTo 2003) the rook is immune be­ cause of mate on g7 and White has both a decisive material and positional advantage. 23 cxd5 i.f3 24 dxc6+ �h8 (D) •••

This forcing move is critical, but it looks like Black is holding his own in the complications that follow.

REFINED ROOK-LIFT:

12 d3

107

Balatonlelle 2000) White has 28 'iti>g2 ! +-. It is very strange that less than a year earlier Pokorna had played the correct 26 . . . ':xe4 ! herself. See the next note. 27 .i.f4 27 �f2? �h5 28 �xf3 'ii'xf3 0- 1 Maros­ Pokorna, Slovakian Team Ch 1 999/00. 27 ':xel+ After 27 . . . ':fxf4?, 28 �f2 ! miraculously turned the tables in Andrijevic-Pavlovic, Yugo­ slavia 1 988. 28 l:txel 'ii'xc2 29 �xf3 (D)

w

•••

This position is critical for 1 8 c4. White is a lot of material up, but he lags behind in devel­ opment and his knight is caught in a nasty pin. The pendulum of theory has swung back and forth in this line. It now seems that Black is no worse, and in fact White should be careful. 25 �d5 This is the main move, but the position is very complicated and the alternatives must be carefully examined: a) 25 .i.d2? loses to 25 . . .l:.xe4 26 ':'xe4 .i.xe4 27 'ii'c 3 'ii'f5 0- 1 A.Ivanov-Agapov, USSR 1 984. b) 25 c7? ! .i.xe4 26 c8'ii' ':'xc8 ! (stronger than LaliC 's 26 . . . .i.xg3 ? ! 27 'ii'xe4 .i.xh2++ 28 �h l ':'fl + ! 29 ':'xfl 'ii'xe4+ 30 �xh2 ':'xc8, which is unclear) 27 'ii'xe4 .i.c5+ 28 .i.e3 (both 28 �h l l::tf l + 29 �g2 ':g l + 30 'iti>h3 �h5+ 3 1 'ii'h4 �xh4+ 3 2 'iti>xh4 ':xe l -+ and 2 8 'iti>g2 ':f2+ 29 'iti>h3 �h5+ 30 �h4 �f5+ 3 1 'ii'g4 ':xh2+ 32 'iti>xh2 �xg4 -+ are just losing for White) 28 . . . �xe4 29 �xc5 �f3 30 �xfS ':xfS =F. c) 25 .i.f4 .i.xf4 26 tt'lf2 (26 c7? 'ii'b 6+ 27 'ii'f2 .i.xc7 28 tt'lc5 .i.e5 =F) 26 . . . ':xe l + (the computers are hot for this, but if Black wants to try to win, there is 26 . . . �xc2 27 .i.xc2 .i.e3 with compensation, as given by Lalic) 27 ':xe l �g5 28 'ii'c 3 ! �xg3 29 'ii'xf3 ! �xf2++ 30 �h 1 ! 'ii'c 5 3 1 ':e5 ! ':xf3 32 ':xc5 �xc5 33 c7 ':'fS 34 .i.e6 =. Ah, technology. 25 .i.xg3 26 hxg3 ':xe4! As Lalic points out, 26 . . . .i.xe4? is a big mis­ take because after 27 l::txe4 ':'xe4 (A.Gara­ Pokorna, European Girls Under- 1 8 Team Ch, •••

B

Originally it was thought that with :c 1 com­ ing, White was much better here. 29 �xb2 Lalic claimed that Black was not worse, and it seems he was right. 30 e7 30 ':c I ? fails to 30 . . . ':xf4. After 30 c7 the position is unclear. Black could try 30 . . . �f6, 30 . . . h6, or even 30 . . . h5 ! ? with the idea 3 1 �xh5 �h7 . .••

Concl usions 1 3 . . . �h4 may not give Black equality, but i t re­ mains a decent practical choice. Black should certainly choose 1 5 . . . 'ii'd7 or 1 5 . . . �f5, as other continuations look distinctly dubious. With pre­ cise play White may hold an edge, but over the board many players will have trouble dealing with Black's active piece-play.

6 El ite Eq ua l izer: 1 2 d3 �d 6 1 3 l:.e l �f5

1 e4 eS 2 liJf3 liJc6 3 ..tbS a6 4 ..ta4 liJf6 5 0-0 ..te7 6 l:[el bS 7 ..tb3 0-0 8 c3 dS 9 exdS liJxdS 10 liJxeS liJxeS U l:[xeS c6 12 d3 ..td6 13 :e1 ..tfS (D)

w

b) With 1 5 'it'f3 ! ?, White puts pressure on the d5-pawn and may try to exchange bishops with ..tf4. After 15 . . . lIeS ( l 5 . . .'it'd7 1 6 ..tf4 ! ) 1 6 ..te3 ..txd3 1 7 liJd2 ..tg6 I S ..tf4 l:[xe l + (Black may be better off acquiescing to the ex­ change of bishops with I S . . . 'ii'c7 1 9 .l\.xd6 'ii'xd6 because 20 liJb3 b4 gives him some counter­ play) 1 9 lIxe l ..tfS 20 lIe5 f6 (20 . . . 'it'a5 ! ?) 2 1 l:[xd5 'it'eS Black had some compensation for the pawn in Hammer-Moradiabadi, Reykjavik 200S, although this is of a rather negative na­ ture - only White can play for a win here.

A) 14 liJd2 (D)

Foregoing the direct 1 3 . . . 'it'h4 in favour of this positional move has become very popular, especially at top level. Black takes aim at the d3-pawn, hoping to make White's development more difficult. In fact, leaving the d-pawn to its fate with 1 4 liJd2 is quite common, but White's real try for an edge comes from 1 4 'it'f3, when Black can either play 14 . . . 'it'h4 after all, or try to take advantage of White's backward devel­ opment with 14 . . . l:[eS. l OS A : 14 liJd2 B: 14 'it'f3 1 09 14 ..txd5 is not considered so dangerous, but it should not be ignored. After 14 . . . cxd5 White has : a) 1 5 ..te3 'ii'h4 1 6 g3 'it'h3 1 7 'it'e2 ..tg4 I S 'ii'f l 'ii'h5 1 9 liJd2 l:[aeS ( 1 9 . . . l:[feS ! ? gives Black the chance to avoid exchanging dark-squared bishops) 20 f3 ..th3 2 1 'it'f2 'ii'g 6 22 liJb3 'it'xd3 23 ..tc5 ..txc5 24 liJxc5 'ii'g 6 25 :lad l ;!; Bolo­ gan-Tkachev, Enghien les Bains 200 1 .

This move tends to lead to drawish positions. Heavy exchanges often leave a simple oppo­ site-coloured bishops position with a symmet­ rical pawn-structure. 14 liJf4 Taking the pawn with 1 4 . . . ..txd3 is also pos­ sible. 15 liJf3 ..tg6 16 .l\.g5 'iWd7 17 liJe5 ( 1 7 ..th4 �:tfe S I S ..txd5 cxd5 1 9 .l\.g3 i s recom­ mended by Marin) 1 7 . . . ..txe5 I S lIxe5 l:.feS 1 9 l:[xeS+ l:[xeS i s quite solid for Black, although •••

EUTE EQUAUZER:

12 d3 J.d6 13 :el i.f5

of course White can hope to exploit his bishop­ pair eventually. After 20 'iWd2 h6 2 1 -tf4 .l::td 8 22 l:.d l -tf5 23 -tg3, as in R.Byme-Angantys­ son, Reykjavik 1 982, Black can play 23 . . . liJxc3 ! 24 -txf7+ (not 24 'iWxd7? liJe2+ 25 'iitf 1 liJxg3+ 26 hxg3 .l::txd7 +) 24 . . . 'iitxf7 25 'iWxd7+ lhd7 26 .l::tx d7+ -txd7 27 bxc3 with a drawn ending. 15 liJe4 Other moves also offer little: a) 1 5 d4 'iWg5 (not 1 5 . . . liJxg2? 1 6 '&t>xg2 'iWh4 17 liJf1 'iWh3+ 1 8 'iitg 1 �g4 1 9 l:.e3 ! +­ Nunn) 16 g3 (after 16 'iWf3, rather than 16 . . . -tg4 1 7 'iWxc6 ! , Black should play 1 6 . . . 'ikg6 ! ? with the initiative) 1 6 . . . liJh3+ 1 7 �g2 liJf4+ 1 8 'iith 1 ? ! liJd3 1 9 ':'e2 l:.ae8, with compensation, is given by Nunn. b) 1 5 'ikf3 (D) and here:

1 09

b) 1 6 'iWf3 -tg6 1 7 .l::te 2 �c7 ( 1 7 . . . c5 ! ?) 1 8 �g5 �xe4 1 9 1he4 'iWxg5 20 'iWxd3 ':'ad8 2 1 'iWf3 c5 was completely level in J.Polgar-Adams, Enghien les B ains 2003 . 16......d7 (D)

17 liJxd6 The ambitious 1 7 lle3 ? ! can only lead to trouble. After 1 7 . . . �xe4 1 8 l:he4 l:.ae8 1 9 '"g4 'iWxg4 20 l:hg4 (Ki.Georgiev-Nunn, Dubai Olympiad 1 986) Black can grab the pawn with 20 . . . liJxb2 as 2 1 i.h6 can be met by 2 l . . .i.e5 +. 17......xd6 18 i.c2 'iWg6 19 i.xd3 i.xd3 20 i.e3 112- 112 Anand-Leko, World Ch, Mexico City 2007.

b I ) 1 5 . . . 'iWd7 is solid enough. 1 6 liJe4 -tg4 1 7 'iWe3 liJxd3 1 8 liJxd6 'ikxd6 1 9 'iWe7 llad8 20 'iWxd6 ':'xd6 112- 112 Smirin-Adams, Tilburg 1 992. b2) 15 ... liJxd3 16 'iWxf5 liJxe l 17 liJe4 g6 ( 1 7 ... -te7 ? 1 8 -th6 ! gxh6 19 .l::t xe l �g5 20 f4 1 -0 was Hage-Robak, corr. 1 995, because after 20 . . . -th4, 2 1 liJg5 ! hxg5 22 �c2 mates) and White does not seem to have more than 1 8 'ikg4 liJd3 1 9 -tg5 -te7 20 .l::td 1 'iitg 7 2 1 �xe7 'iWxe7 22 .l::tx d3 f5 23 'iWh3 'iWxe4 24 l:.d7+ '&t>f6 25 .l::td 6+, with a draw, as given by Marin. 15 liJxd3 16 �g5 White develops with tempo. Other moves are harmless at best: a) 1 6 liJxd6 ? ! leaves White lagging behind in development. 1 6 . . . 'ikxd6 1 7 -te3 'iWg6 1 8 l:.f1 ( 1 8 l:.e2 -tg4) 1 8 . . .liJxb2 1 9 'it'd2 liJd3 + Bluvshtein-Onishchuk, Montreal 2003 .

B) 14 'iWf3 (D)

•••

This is White's main attempt to seek any kind of advantage. White attacks d5 and may

UNDERSTANDING THE MARSHALL A TTACK

110

also with are: Bl: B2:

look to exchange dark-squared bishops 1 5 �xd5 cxd5 1 6 �f4. Black ' s main tries 14 ':'eS 14 'ii'h4 •••

..•

1 10 1 12

14 . . . 'ii'd7 ? ! leads to a rather dreary position for Black. White can execute his positional plan with 15 �xd5 cxd5 16 �f4 (Black will re­ gain his pawn, but his troubles are not over) 1 6 . . . �xf4 ( 1 6 . . J:He8 1 7 lDd2 �xf4 1 8 'ii'xf4 �xd3 amounts to the same thing) 1 7 'ii'xf4 �xd3 1 8 lDd2 �He8 ( 1 8 . . Jlae8 1 9 ':'e3 ':'xe3 20 'ii'xe3 �g6 2 1 .:te l h6 22 lDb3 , as in Smagin­ Geller, Moscow 1 989, is similar) 19 lDb3 ( 1 9 h 3 h6 20 'ii'd4 �g6 2 1 lDb3 'ii'c6 22 lDc5 a5 23 a3 a4 24 .:tad 1 again gives us the same structure, Anand-Adams, Paris (rapid) 1 992) 1 9 . . . 'ii'c 6 20 'ii'd2 �g6 2 1 lDd4, as in Fressinet-Hamdouchi, Belfort 2003. White has a strong knight on d4 against a rather bad light-squared bishop, and Black faces a rather unpleasant defensive task.

'ii'x e2 ':'xe2 22 .:te l .:txe l + 23 �xe l lDb6 was given as equal by Bronstein, but clearly White can keep trying here. b) 1 6 . . .lDf4 17 �c2 (better than 1 7 lDa3 ? ! 'ii'e 2 1 8 �xf4 'ii'xf3 1 9 gxf3 �xf4 2 0 d4 ':'e8, as given by Karpov) 17 . . . .:td8 1 8 lDa3 'ii'e2 1 9 'ii'xe2 lDxe2+ 20
B1) 14 .:teS (D) •••

w

16 'ii'e l+ This check is the most obvious move and is probably best. These alternatives are experi­ mental and have not done so well: a) 16 ... 'ike5 ? ! 17 lDfl �g6 was Kotronias­ Pavlovic, Vrnjacka Banja 2005 . New in Chess Yearbook 86 gives 1 8 �d2 ! ± with the idea 1 8 . . . �h5 ? ! 1 9 g4 �g6 20 :te l +-. b) 1 6 . . . 'ii'd7? ! 1 7 h3 ':'e8 1 8 lDe4 �xe4 1 9 dxe4 'ii'e6 ( 1 9 . . . f5 20 e5 ! ? �xe5 2 1 �e3 ;t in­ tending .:td l , Dikmen-Harding, COIT. 1 999) 20 �e3 'ii'xe4 2 1 'ii'xe4 .:txe4 22 .:td l ! ± Raidna­ Delalleau, COIT. 200 1 . c ) 1 6. . .'ii'e 6 i s a better version of line 'b' , but it still falls short of equality. 1 7 h3 .:te8 1 8 lDe4 �xe4 1 9 dxe4 'ii'e 5 ? ! (better is 1 9 . . . 'ii'xe4 20 'ili'xe4 ':'xe4 but White certainly has some pull after 2 1 �fl because of his bishop-pair) 20 •••

This is a fresh approach. Instead of rushing to a pawn-down ending that often arises in Line B2, Black keeps pieces on the board and hopes to get counterplay based on White's weak back rank. 15 ':'xeS+ 'ii'xeS 16 lDd2 1 6 �d2 is a bit passive, but Black should know what to do: a) 1 6 . . . 'ii'e 5 ? ! 1 7 g3 l:.e8 1 8 lDa3 �xa3? 1 9 bxa3 ? ( 1 9 .:te l ! ± ) 1 9 . . . �h3 20 d4 'ii'e2 2 1

EUTE EQUAUZER:

1 2 d3 J.d6 13 :el J.f5

�fl ii.c5 21 ii.d2 'ii'h 2 22 �e2 I:te6 23 c;i(d3 ! (White ' s king is brave and he will consolidate his extra pawn) 23 .. .1:1f6 24 'it'g3 'it'xg3+ 25 fxg3 ± Shirov-Onishchuk, Poikovsky 2008. 17 ttJn ii. g6 (D)

18 g3 Or: a) 1 8 h3 l::te 8 1 9 �d l and here: a l ) 1 9 . . . ttJxc3 ! ? may be playable. 20 bxc3 'it'xc3 2 1 .:tbl .:te l 22 ii.f4 was given as ;!; by Dolmatov, and this may be true. Black has a few ways to try to defend but White should be care­ ful as well. 22 . . . ii.xf4 (after 22 . . . 'it'xd3 23 'it'xd3 ii.xd3 24 ii.c2 I:txbl 25 ii.xbl ii.xbl 26 ii.xd6 ii.xa2 White can still try a bit) 23 'it'xf4 h6 24 ii.c2 l::tx fl + (this looks safer than 24 . . . .:txbl 25 ii.xbl 'it'b2 26 ttJd2 c5 27 c;i(h2 c4 28 dxc4 ii.xbl 29 ttJxbl 'it'xbl 30 c5, although Black may be fine here) 25 l::tx fl 'it'xc2 26 .:te l 'fixa2 27 l::txc6 ii.xd3 112- 112 Schreber-Simmelink, COIT. 200 1 . White could certainly play on with 28 'it'd6 ! . a2) 1 9. . .ii.h2+ is thematic and solid. 20 �xh2 'it'xfl 2 1 ii.d2 ii.xd3 (worse is 2 1 . . .'it'xd3 ? ! 22 'it'xd3 ii.xd3 23 a4 ;!; Dolmatov-Khalifman, Moscow 1 990) 22 .:tc l ii.e4 ! (not 22 . . . ii.e2? 23 ii.xe2 'fIxe2 24 lIe l - Dolmatov) 23 'it'g3 'it'c4 24 ii.b3 'it'e2 25 l::te l 'ii'x d2 26 l::txe4 :f8 ! with counterplay, Samundsson-JOnsen, COIT. 2002. Black's active queen gives him good play. b) 1 8 �xd5 cxd5 19 'it'xd5 .:td8 20 ii.g5 'it'xa l 2 1 ii.xd8 (D) probably only leaves White on the better side of a draw: b l ) 2 1 . . .ii.f8 22 f4 (after 22 h4 h6 23 ii.b6 'it'xb2 24 'fid8 'it'bl 25 ii.c5 ii.xd3 or 22 h3 'fixb2 23 'fIa8 'it'a3 24 d4 h6 intending . . . �h7

111

B

Black should hold) and now 22 . . . 'it'xb2 was played in Chandler-Hebden, Catalan Bay 2004. Since 23 f5 �h5 24 f6 leaves White with some initiative, I would prefer 22 . . . 'fid l ! ? 23 'it'f3 'it'xf3 24 gxf3 ii.xd3 with a likely draw. b2) 2 1 . . .ii.xh2+ 22 �xh2 'ii'x fl leaves White two pawns up but it is very difficult for him to make any progress. 23 ii.h4 (23 'ii'a 8 h6 looks alright for Black) 23 . . . h6 24 d4 'ii'b l (24 . . . 'ii'e2 ! ) 2 5 'ii'a 8+ (25 'ii'b 3 is a better try) 2 5 . . . �h7 26 'ii'xa6 'ii'x b2 27 d5 'it'xc3 28 ii.g3 'it'd3 29 d6 'it'd5 30 'it'a7 112- 112 Stellwagen-L' Ami, Schagen rapid 2005 . 18 b4! (D) This is much better than 1 8 . . . :e8? ! 1 9 ii.d l 'it'e6 ( 1 9 . . . ttJxc3? 20 ii.d2 +-) 20 ii.d2 'it'h3 2 1 a4 ;!; Dolmatov-Kamsky, Dortmund 1 993. In­ stead 2 1 'fIg4 is given a ' ! ' in Informator, but af­ ter 2 1 . . . 'ii'xg4 22 ii.xg4 ii.xd3 23 ii.d7 l::te 2 24 ii.xc6 ii.e4 ! Black has good counterplay. How­ ever, 2 1 ii.c2 also looks good for White. •••

w

19 �xd5

UNDERSTANDING THE MARSHALL A TTACK

112

Future developments are sure to come after 1 9 c4 lbf6 ( 1 9 . . . lbc7 ! ? was suggested by Rad­ jabov and Bacrot) 20 .id 1 (Nataf suggests 20 'iixc6 ! in lnformator, giving many variations, of which his main line goes 20 . . . .l:.dS 2 1 'iib6 .l:.d7 22 .i.c2 ! lbg4 23 .l:.b 1 'iie 2 24 .id2 lbe5 25 'iie 3 'iih 5 26 .id 1 'iif5 27 .i.e2 .i.fS ;1;) 20 . . . l:teS 21 .i.d2 'iie 5 22 'iix c6 .i.fS, which gave Black good play in Stellwagen-Gustafsson, Bundes­ liga 2007/S . 19 cxd5 20 'iixd5 l:Id8 21 .i.g5! The play becomes very forcing from here on. 21 'iix a1 22 .ixd8 .if8 23 .ia5 23 cxb4 'iixb2 gives Black counterplay. 23 'iib l ! (D) 23 . . . bxc3 ? ! 24 .ixc3 'iid 1 allows White to consolidate with 25 'iie 5 ! .ixd3 26 'iie 1 ;1;. •••

•••

•.•

24 �g2 24 d4 .ie4 gives Black good counterplay. 24 ... .ixd3 25 lbd2 'iic2 26 'iia8 No better is 26 'iig 5 h6 27 'iie 3 .ic5 2S 'iie l ? ! (White should settle for 2S 'iix c5 'iixd2 = ) 2S . . . 'iix b2 29 .i.xb4 .ixb4 30 cxb4 'iix a2 + Vouldis-Gustafsson, Greek Team Ch, Ermioni Argolidas 2005 . 26 'iixd2 27 .i.xb4 h5! 28 'iixf8+ �h7 29 'iixf7 'iie2 Black quickly delivered perpetual check in 'Team Ojjeh' - 'Team Nataf' , corr. 2003 . •••

82) 14 'iih4 (D) This is the main line. Several lines head into endgames where Black is a pawn down but has excellent chances to hold. Aronian has just •••

about perfected the art of drawing these end­ games, but even he has occasionally run into some trouble in doing so. This type of position will certainly not suit everyone. Shirov keeps trying to prove something from the white side, while both Leko and B acrot seem content mak­ ing quick draws in this variation with either colour. 15 g3 1 5 h3? ! l:taeS 1 6 l:txeS l:txeS 1 7 .i.d2 only leads to trouble for White after 1 7 . . . lbf4 ! , while 1 5 'iixf5 is not refuted but allows Black at least a draw after 1 5 . . . 'iixh2+ 1 6 �f1 l:IaeS ( 1 6 . . . 'iih 1 + 1 7 �e2 l:IaeS+ I S .i.e3 lbxe3 1 9 :xh 1 lbxf5+ 20 'iitf3 is equal) 1 7 l:IxeS l:txeS I S .i.e3 'iih 1 + (after l S . . . lbxe3+ 1 9 fxe3 'ii'h 1 + 20 �f2 .i.g3+ 21 �xg3 l:Ixe3+ 22 �f4 'iif 1 + 23 �xe3 'iixf5 24 lba3 Black should probably just deliver per­ petual check) 1 9 �e2 g6 ! 20 'iif3 (20 'iid 7? ':xe3+ mates) 20 . . . 'iic 1 2 1 .ixd5 cxdS 22 'iixd5 .ifS 23 b4 ! 'iVb2+ 24 lbd2 'iWxal 25 lbe4 =, as given by Finkel. 15 'ii h3 (D) •••

ELITE EQUAliZER:

12 d3 J.d6 13 :'el J.j5

Now White can continue his development or grab another pawn: B21 : 16 ll:ld2 1 13 1 14 B22: 16 .i.xd5 1 6 .i.e3 does not give White much after 1 6 . . . .i.xd3 1 7 ll:ld2 'ii'fS 1 8 .:tad 1 (or 1 8 'ii'xfS .i.xfS = Ivanchuk-Aronian, Amber Rapid, Mo­ naco 2008) 1 8 . . . l:.fe8 (Black moves this rook to ensure he can keep his dark-squared bishop af­ ter the following sequence) 1 9 'ii'xfS .i.xfS 20 .i.xdS cxdS 2 1 ll:lb3 .i.e6 22 .i.cs .i.c7 (this would not be possible if Black had a rook on fS), J.Polgar-Leko, Wijk aan Zee 2008. Black's bishop-pair compensates for the isolated pawn.

113

a) 23 . . . .i.b8 24 axbS axbS 2S l:la6 �d7 26 .i.d2 :e6 27 l:le3 l:lg6+ 28 ll:lg3 l:lgf6 29 ll:le4 l:.g6+ 30 ll:lg3 l:lgf6 3 1 ll:le4 l:lg6+ 112- 112 Ago­ pov-Gustafsson, European Team Ch, Kherson­ isos 2007. b) 23 . b4 24 c4 ll:lf4+ 2S .i.xf4 .i.xf4 26 cS+ �h8 (with compensation) 112- 112 Kotronias­ Asrian, European Clubs Cup, Kerner 2007 . 21 .i.xf4 22 gxf4 fxe4 23 dxe4 (D) ..

•••

8 21) 16 ll:ld2 (D)

White will win back the piece and it looks like Black will remain a pawn down, but there is a neat tactical resource. 23 .i.f3+! By giving back the bishop instead of the knight, Black is able to establish material equal­ ity. 24 �xf3 l:txf4+ 25 �g3 l:lfxe4 26 l:lxe4 l:lxe4 27 f3 (D) •••

This usually leads to an endgame where White retains a symbolic advantage. Black has less trouble defending here than he does a pawn down in Line B22. 16 .:.ae8 17 ll:le4 .i.g4 18 'ii'g2 'ii'xg2+ 19 'iitxg2 f5 Both sides are caught in pins. 20 h3 The immediate 20 .i.f4 is similar, because 20 . . . .i.xf4 2 1 gxf4 fxe4 22 dxe4 is met by the same trick: 22 . . . .i.f3+ ! 23 �xf3 l:txf4+ 24 �g3 l:tfxe4 2S l:txe4 .:txe4 26 f3 l:teS 27 c4 bxc4 28 .i.xc4 as . This has occurred several times at high level with all the games being drawn. 20 .i.h5 21 .i.f4 White can also try 2 1 g4 fxg4 22 hxg4 �xg4 23 a4, but Black does not have many troubles: •••

•••

27 l:le5 This is perhaps a tiny improvement on the al­ ready sufficient 27 . . . l:te2 28 c4 bxc4 29 �xc4 •••

UNDERSTANDING THE MARSHALL A TTACK

114

:xb2 30 �xa6 g5 of Nakamura-Aronian, Gi­ braltar 2005, when White may try a little tiny bit with 3 1 �c4. 28 c4 bxc4 29 �xc4 a5 This is the idea behind Black's 27th move ­ White does not even get a passed a-pawn. 30 J:!.et :g5+ 31 'it>f2 'it>rs 32 �n liJe7 = 1.Polgar-Aronian, Wijk aan Zee 2008.

W

822) 16 �xd5 cxd5 (D)

This is more critical. There is a further split, as White can play positionally or grab material. We have: B22 1 : 17 �f4 1 14 1 14 B222: 17 �e3 1 15 B223: 17 'ifxd5

This second pawn offer gives Black enough play to hold the balance. White's pawns are all weak and Black's rooks will be very active. 21 liJd2 2 1 cxd4 :fd8 22 :e3 �f5 23 a4 b4 24 liJd2 :xd4 gave Black compensation in Grishchuk­ Tkachev, Prague rapid 2002. Grishchuk was clearly convinced of the viability of Black's idea, as his game against Kotronias given above was played two years after this one. 21...dxc3 22 bxc3 IUd8 23 :e3 �f5 24 liJe4 :ac8 Black has enough compensation. 25 a4 g6 26 axb5 axb5 27 h4 :c6 28 :a5 :b6 29 :a7 h6 30 'it>h2 b4 3 1 cxb4 :xb4 32 h5 :bd4 33 hxg6 112- 112 B auer-Adams, Senat 2003 .

8222) 17 �e3 (D)

8221) 17 �f4 White aims to return the pawn for the type of positional advantage we have come across be­ fore, but in this case Black does not have to oblige. 17 �g4! 17 ... �xf4? ! 18 'iVxf4 �xd3 19 liJd2 is quite pleasant for White. 18 'iVg2 �xf4 19 'iVxh3 Worse is 1 9 gxf4 d4 20 �g3 ? ! :ad8 2 1 c4? ! (2 1 cxd4 :xd4 2 2 :e3 �xg3+ 2 3 fxg3 :fd8 +) 2 l . . .J:!.d6 ! 22 :e5 f5 ! with a big advantage for Black in Kotronias-Grishchuk, FIDE Knock­ out, Tripoli 2004. 19 �xh3 20 gxf4 d4! (D) ..•

•••

This move allows White to maintain an ex­ tra pawn, but queens usually stay on the board

ELITE EQUALIZER:

12 d3 JJ..d6 13 '1:.el JJ..f5

and B lack should be able to maintain the bal­ ance. 17 ii.xd3 18 'ifxd5 l:Iad8 19 'iffJ 1 9 'ifg2 'iff5 ! ? ( 1 9 . . .'ii'xg 2+ and 1 9 . . . 'ifh5 are covered under Line B 223) 20 �d2 b4 2 1 .i.d4 bxc3 2 2 bxc3 h6 2 3 a4 l:Ife8 24 'ifc6 112- 112 Grishchuk-Khalifman, European Clubs Cup, Rethymnon 2003. 19 ii.c4 We shall take this as the main line because it is Aronian 's preference. 1 9 . . . .i.f5 20 �d2 .i.e6 is another way of reaching the same position, while 19 .. .'ii' f5 20 'ifxf5 ii.xf5 2 1 �d2 l:tfe8 22 ii.d4 f6 23 �b3 r1;f7 was Kasimdzhanov­ Onishchuk, Calvia Olympiad 2004. Black has some compensation here, but it still looks like an inferior version of the main line of B 223 . 20 �d2 ii.e6 21 ii.d4 (D) •••

115

29 f3 l:Ie7, when it is Black who has the initia­ tive. 22 'it'g2 One way for White to force a draw is 22 �e4 ii.d5 23 ii.xg7 �xg7 24 'iff6+ �g8 25 'fi'g5+ �h8 26 'it'f6+ �g8 112-112 Bacrot-Aronian, FIDE World Cup, Khanty-Mansiisk 2005 . 22 'it'h5 23 f3 (D) •••

•••

B

Anand-Leko, Amber Rapid, Monte Carlo 2007 was drawn here immediately. 23 .'ii'g6 24 a3 h5 25 lle2 .:td5 26 .:tae1 l:.fd8 27 �e4 ii.d6 28 'fi'f2 .:tr5 29 �xd6 112- 112 Anand-Aronian, Amber Rapid, Nice 2008. After 29 . . . .:txd6 the opposite-coloured bishops and Black's counterplay against f3 make the position pretty level. ••

With the bishop-pair and an active position, Black has decent compensation for the pawn. This position has been seen in a lot of high­ level games, and White has made little prog­ ress. 21 ii.b8 Alternatives also look OK for Black: a) 2 l . . .h6 22 a3 ii.b8 23 'fi'g2 'fi'f5 24 f3 .:tfe8 25 �e4 ii.d5 26 l:.e2 .:te6 27 .:tae 1 112- 112 Kasimdzhanov-Adams, Linares 2005 . b) 2 l . . .ii.e7 22 'ifg2 'ifh5 23 l:Ie5 'ifg6 24 �e4 l:Id7 25 �c5 ii.xc5 26 l:Ixc5 and now 26 . . . l:Ifd8 ? ! 27 'iff3 h6 28 h4 ii.xa2 29 l:Ic6 'ifc2 30 l:Ixa6 ii.d5 3 1 'it'g4 ± gave White an extra pawn and the initiative in Naiditsch-Ivanchuk, European Clubs Cup, Izmir 2004. Black can improve by 26 . . . .i.xa2 with the idea 27 l:tc6 (not 27 l:txa2? 'fi'b1 +) 27 . . . 'fi'd3 28 l:txa6 ii.d5 •••

82 23) 17 'it'xd5 (D)

B

This is the critical test.

UNDERSTANDING THE MARSHALL A TTA CK

116

17 llad8 18 'ifg2 (D) 1 8 'iff3 �b8 ! ? 1 9 d4 �g4 20 'ifg2 'ifhS 2 1 lLld2 llde8 2 2 l:le3 fS ! ? gave Black compensa­ tion in Voss-Happe, Gennan corr. Ch 1 993-S. •••

a) 2 1 .i.d4 �e6 (2 l . . .�h3 22 'it'f3 'ifxf3 23 lLlxf3 l:.fe8 is also possible) 22 'iff3 'ifg6 23 'ifc6 (23 'iWe4 'ii'h S 24 'ii'f3 'iWg6 2S 'it'e4 'ifhS 26 'iWf3 112- 112 Wang Hao-Yakovenko, Russia­ China match, Nizhny Novgorod 2007) and now Pavlovic's 23 . . . �f4?! 24 lLle4 'ifhS ? ! 2S gxf4 'ifg4+ 26 lLlg3 �dS 27 'ifcs 'iff3 28 �fl .i.c4+ 29 l:le2 l:lc8 fails to 30 'ifeS , so Black should try 23 . . . 'ifc2 ! ? 24 lLle4 .i.e7. b) After 2 1 'ifc6 Black has experienced some problems in practice, but there is some hope. b I ) 2 l . . ..i.e6? 22 'iWxa6 'iWdS 23 'iWb6 �h3 24 f3 fS 2S 'iWd4 'iWc6 26 'ii' h4 �cS 27 �d4 +­ Shirov-Yakovenko, FIDE World Cup, Khanty­ Mansiisk 2007. b2) 2 l . . ..i.h3 (D) and here:

Now we have a final branch. Black can keep queens on and try his luck in the middlegame two pawns down, or he can immediately regain one pawn and try to hold the ending. We have: 1 16 B223 1 : 18...'ifhS B2232: 18 ...'ifxg2+ 1 18

82231) 18 ...'ifhS This is a risky winning attempt, but it ap­ pears quite viable if Black is well prepared. 19 .i.e3 �h3 Black plays for a direct attack. Regaining a pawn with 1 9 . . . .i.xd3 is rather obvious, but Black remains a pawn down and must be care­ ful or else White will take over the initiative. After 20 lLld2 .i.fS (D) White has:

w

b2 1 ) 22 'ifxa6?? is a blunder. Black wins with 22 . . . �xg3 ! 23 hxg3 l:lxd2 ! intending . . . 'iWf3 . b22) 22 �d4 �b8 23 l:le3 (23 lLle4 ? ! 1:bd4 ! 24 cxd4 'iff3 2S lLlf6+ 'ifxf6 26 'ifxf6 gxf6 +) 23 . . . l:ld6 24 'ifb7 .i.d7 ! gave Black compensa­ tion for the pawn in Smeets-Khalifman, Am­ sterdam 2007 . The bishop is heading for the c6-square. b23) 22 a4 looks critical, but an idea of Pavlovic's may resurrect this line for Black: 22 . . . bxa4 ! (22 . . . .i.b8 23 axbS axbS 24 ];1aS .i.d7 2S 'iWcs 'iWg6 26 'ii'g S 'iWd3 ? ! { Kosten's idea 26 . . . 'iWc6 27 .i.d4 f6 100ks like a better try } 27 .i.d4 f6 28 'iWdS+ 'it>h8 29 lLlb3 l:.de8 30 l:.aal ± Stellwagen-Pashikian, Erevan 2(07) 23 'ii'xa4? ! (23 l:.xa4 �d7 24 'it'xd6 �xa4 2S 'ifxa6 and even 23 'ifxa6 100k like better tries) 23 . . . �cS ! 24 �xcS (24 'ifd l ? ! 'ifxd l 2S l:laxd l �xe3 26

ELITE EQUAUZER:

12 d3 i.d6 13 'f:..e l .tf5

l:he3 .i.f5 ! +) 24 . . . ':xd2 (intending . . . 'ii'xc5 and . . . 'ii'd 5) 25 'ii'c 4 ':'xb2 gives Black suffi­ cient play. White cannot play 26 i.xf8?? be­ cause of 26 . . . 'iIi'f3 . 20 'ilhl This looks funny, but after 20 'ii'c 6 i.e6 2 1 iLld2 i.d5 2 2 'ii'xa6 'ii' h 3 2 3 iLle4 f5 Black will win a piece. The position is still very messy though because White will gain several pawns in return. 20 f5 21 i.b6 Shirov 's latest attempt 2 1 f4 was not suc­ cessful after 2 1 . . . .i.c7 22 iLld2 ':xd3 23 iLlb3 i.g4 24 'ili'g2 .i.f3 25 'ili'f2 .i.aS 26 tLlc5 ':d6 27 i.d4 'ii'f7 with compensation, Shirov-Yako­ venko, Russian Team Ch, Dagomys 200S. 21 l:td7 Bacrot also considered 2 1 . . .f4 22 i.xdS f3, but his second Naiditsch found 23 i.b6 ! .i.g2 24 'ilxg2 fxg2 25 iLld2 'ii'g 6 26 .i.d4 'ii'x d3 27 l:tad l , when White will have more than enough for the queen. 22 'ili'd5+ White pins the d6-bishop and the f-pawn with gain of tempo. Instead 22 iLld2 allows Black to attack with 22 . . . f4. 22 �hS 23 iLld2 (D) This allows an interesting shot, but 23 i.d4 allows 23 . . . l:tfdS 24 'ii'c 6 l:tc7 25 'ili'h l f4 26 iLld2 ':'f8 with good counterplay.

11 7

25 .i.xd2 26 .i.d4 ':gS 27 ':e7?! Already White should think about forcing a draw with 27 �h l i.g4 2S i.xg7+ l:txg7 29 'ili'bS+ ':gS 30 'ili'e5+. 27 .i.c1 ! (D) •••

•••

w

•••

•••

•••

B

A clever interference move, which B acrot found before the game ! 2S :el Not 2S ':xc l ? 'it'g5+ mating. 2S i.f4 2S . . . i.d2 repeats, but Bacrot already feels compelled to try for more. 29 :e3 Other moves are worse: a) 29 i.e5 ? i.g4 ! 30 .i.xf4 .i.f3 3 1 'ii'xf3 'ii'xf3 -+. b) 29 'ili'c6? .i.xh2+ ! 30 ..txh2 (30 �h l i.d6 ! +) 30 . . . .i.g4+ 3 1 �g3 f4+ ! 32 ..txf4 l:tf8+ -+. 29 .i.xe3 30 i.xe3 (D) 30 fxe3 'ili'e2 gives Black the initiative. •••

•••

B

23 i.xg3! 24 'ili'xd7 i.f4 25 'it'b7 White should probably try 25 'it'd5 i.xd2 26 i.d4 i.xe l 27 l:txe l ':eS 2S ':xeS+ 'ii'xeS and here both 29 .i.e5 and 29 'it'e5 may be some­ what drawish but White could try to win with no risk at all. •••

30 l:teS •••

UNDERSTANDING THE MARSHALL A TTACK

118

After 30 . . . h6 3 1 'iic7 l:teS 32 �d4 :e7 33 'iig 3 c;fi>h7 Bacrot prefers Black, although the position is still very unclear. This position was seen in practice in Sutovsky-Yakovenko, Poi­ kovsky 200S. Both sides managed to hold . . . Black can also play the forcing 3 0 . . . 'iig4+ 3 1 c;fi>h l f4 32 l:tg l 'iic s 3 3 'iix cs i.xcs 34 i.xf4 i.b7+ 35 l:tg2 l:teS 36 i.e3 . White has two pawns for the exchange but I have the feeling Black is still a little better. 31 i.d4? 3 1 'iic 7 h6 32 �d4 l:te7 33 'iig 3 c;fi>h7 trans­ poses to Sutovsky-Yakovenko given above and looks to be the safest option at this point. 3 1 c;fi>h l is also better than the move played. 31. 'iig4+ 32 c;fi>hl h6! 33 f4 White wants to play ':g l , but the immediate 33 l:tg l loses to 33 . . . 'iix gl + ! . 33 J�e7 34 'iia8+ Or 34 'iic 6 'iixf4 35 'iig6 .i.n 36 'iig 3 'iix g3 37 hxg3 �xd3 +. 34 c;fi>h7 35 �e5 35 l:tg l 'iixf4 is also bad. 35 'iie2 36 ':gl .i.g4 37 ':xg4 fxg4 38 d4? 'iio # (0-1) Smeets-Bacrot, Wijk aan Zee 200S. ••

• •

•••

•••

82232) 18 1Ii'xg2+ This is Black's main choice and is always played by the Marshall's top practitioner, Aro­ nian. Black should hold the endgame, but he will have to work. 19 c;fi>xg2 .i.xd3 20 i.e3 (D) .•.

B

20 i.e4+ •••

This is the latest finesse, by which Black re­ locates his bishop immediately. 20 . . . l:tfeS is the 'old move' . After 2 1 tbd2 Black has tried: a) 2 1 . . .f6 22 i.b6 l:tbS 23 l:txeS+ l:txeS 24 tbb3 i.c4 25 l:td l i.fS 26 i.e3 (White has made some progress; in order to hold endings like this with Black, it is important to know how they can be lost) 26 . . . c;fi>f7 27 tba5 i.e6 (27 . . . i.xa2 2S b3) 2S a3 ':cS 29 B ':c7 30 l:ldS l:tcS 3 1 tbb7 ! ':c7 32 tbd6+ i.xd6 33 l:.xd6 ± Shirov-Kariakin, FIDE World Cup, Khanty­ Mansiisk 2007 . b) 2 1 . . .b4 22 i.b6 (after 22 i.d4 bxc3 23 i.xc3 f6 24 l:tad 1 l:txe 1 25 lhe 1 c;fi>f7 26 f3 i.e7 27 g4 h5 2S h3 hxg4 29 hxg4 i.b5 Black managed to draw without too many issues in Stellwagen-Harikrishna, Wijk aan Zee 200S) 22 . . . l:txe l 23 l:txe l l:.bS 24 i.a5 bxc3 25 i.xc3 f6 26 tbe4 i.fS 27 B. Both sides have played very logically and now it is a matter of how much progress White can make. This position has been seen in a couple of very high level en­ counters : bl ) 27 . . . i.c4 2S b3 (perhaps 2S a3 , as in the next note, is a better try) 2S . . . .i.b5 29 .i.a5 ':cS 30 tbc3 i.c6 3 1 l:.d l �f7 32 tbe2 h5 33 ':c l i.d7 34 l:txcS i.xcs was quickly drawn in Shirov-Leko, MorelialLinares 200S. b2) 27 ... l:tcS 2S l:.d l i.c4 29 a3 .i.b5 30 g4 i.c6 3 1 l:td3 l:teS 32 l:td4 l:lbS 33 h4 .i.e7 34 h5 (White is clearly making some progress) 34 . . . l:.dS 35 l:tc4 i.b5 36 l:tc7 l:td7 37 l:txd7 i.xd7 3S h6 gxh6 39 i.xf6 i.f8 40 i.d4 ± Shirov-Aronian, MorelialLinares 200S. If White can avoid an opposite-bishop end­ ing and slowly gain space, he has chances of success in converting the extra pawn. B e­ cause of this, we often see Black carefully ad­ vancing his kingside pawns - it is difficult for White to win on the queenside alone, and the weaknesses on White's kings ide give Black a target for creating counterplay. Black has an unopposed light-squared bishop, so he will generally arrange his kings ide pawns to help control the dark squares with . . . f6, . . . g5 and . . . h5 . 21 f3 i.c6 22 tbd2 l:tfe8 (D) 23 tbb3 23 �d4 is also possible. After 23 . . .l:he l (Black held without breaking much of a sweat after 23 . . . h5 in Leko-Aronian, Karen Asrian

EUTE EQUAUZER:

1 2 d3 i.d6 13 :el i.f5

119

24 ... :xe1 25 ':'xe1 :d7 26 l:Idl �f7 27 �c5 �c7 28 l:txd7+ �xd7 29 lLld2 h5 30 �d4 �c6 (D)

Memorial rapid, Erevan 2008) 24 :xe 1 f6 25 lLln (or 25 �f2 h5 26 lLle4 .i.fS 27 :e2 �f7 28 lLlc5 .i.xc5 29 .i.xc5 J:d7 30 .i.d4 a5 3 1 b3 1/2 - 1/2 Leko-Naiditsch, Dortmund 2008) 25 . . . h5 26 �b6 l:te8 27 lLle3 �f7 28 'iti>f2 h4 29 l::td 1 �b8 3 0 lLlc2 ! (Black must be careful) 3 0. . . g5 3 1 gxh4 gxh4 32 l::td4 l::th 8 33 lLlb4 �a8 34 f4 a5 35 lLla6 ! (now White's advantage is clear) 35 . . . �e6 36 lLlxb8 ':xb8 37 �xa5 l::tg 8 38 �e3 l::tg 2 39 l::td2 h3 40 l::tx g2 hxg2 41 �f2 �f5 42 �d8 'iti>xf4 43 �xf6 �g4 44 h4 White eventu­ ally squeezed out a win in B acrot-Sargisian, French Team Ch, Evry 2008. Although the endgame is drawish, Black must defend well. White can win against inaccurate play. Usually Black is better off keeping both of his bishops on the board. As long as a pair of rooks re­ mains, Black must be very careful about allow­ ing opposite-coloured bishops. 23 f6 24 �b6 Trading rooks tends to ease Black's defence, but White made little progress after 24 �f2 .i.c7 25 lLld4 .i.d7 26 l::t ad 1 �f7 27 :d2 �c8 28 l:Ied 1 h5 29 lLle2 :xd2 30 :xd2 �e6 3 1 b3 g5 in Leko-Sargisian, Karen Asrian Memorial rapid, Erevan 2008. Black also held easily after 24 lLld4 .i.b7 25 g4 �c5 26 b4 �xd4 27 �xd4 :xe 1 28 :xe l �f7 29 �c5 :d7 30 h4 �d5 3 1 a3 �e6 3 2 �d4 �c4 in Naiditsch-Bacrot, Dort­ mund 2009. •••

This is an ideal defensive stance for Black. The rooks are off the board and Black has gained space on the kingside. It is also difficult for White to do anything on the queenside. White is certainly not going to lose, but he can­ not make any progress either. 31 b3 f5 32 �f2 g5 33 �e3 g4 34 f4 �e6 35 �d4 1/2 - 1J2 B acrot-Aronian, European Clubs Cup, Kallithea 2008.

Concl usions Currently this is Black's main choice against 12 d3. After the main move 14 'ii'f3 , Black can try the relatively unexplored 14 . . . :e8 or play the standard 14 . . . 'ii' h4. The most critical line 15 g3 'ii'h 3 16 �xd5 cxd5 1 7 'ii'x d5 l::t a d8 1 8 'ii'g 2 offers Black a choice. He can head into an endgame which gives him excellent draw­ ing chances, although many may find it rather depressing to defend. The other option is to keep queens on the board while two pawns down. This is riskier, but it also allows Black to keep winning chances.

7 Ea rly Deviations

1 e4 e5 2 lLlf3 lLlc6 3 .ltb5 a6 4 �a4 lLlf6 5 0-0 .lte7 6 l:tel b5 7 .ltb3 0-0 8 c3 d5 9 exd5 lLlxd5 10 lLlxe5 lLlxe5 11 l:txe5 c6 (D)

Section 7.5: 1 2 d4 .ltd6 13 :tel 'ii'h4 14 g3 'ii'h3 15 �xd5 and rarer moves

1 34

Section 7 . 1

In this chapter we look at various lines aris­ ing after 1 1 . . .c6 where White avoids the main lines of the first six chapters. The first part deals with 12 :te l , which can be used as a trans­ positional device. This is in fact a common move-order to reach the 1 2 d3 lines, but here White hopes to avoid the 1 3 . . . .ltf5 line of the previous chapter. The second part is usually just called 'the 1 3 l:te2 line' . White hopes the rook will serve some defensive purpose on the second rank. Next is the Kevitz Variation, char­ acterized by the sequence 1 2 .ltxd5 cxd5 1 3 d4 .ltd6 14 l:te3 . The rook is more active on e3 than it would be on e l or e2, but it can also be a tar­ get itself. The next section looks at the rather fashionable 1 2 d4 �d6 1 3 ':'e l 'iWh4 1 4 g3 'ii'h 3 15 'ii'e 2. White focuses on evicting the black queen from the kingside before developing his queenside. Finally we examine the immediate capture on d5 with 1 5 .ltxd5 . Section 7.1 : 12 l:tel 1 20 Section 7.2: 12 d4 .ltd6 13 l:te2 1 25 Section 7.3: 12 .ltxd5 cxd5 13 d4 .ltd6 14 .l:.e3 1 28 Section 7.4: 12 d4 �d6 13 :tel 'ii' h4 131 14 g3 'ii'h3 15 'ii'e2

1 e4 e5 2 lLlf3 lLlc6 3 .ltb5 a6 4 .lta4 lLlf6 5 0-0 .lte7 6 :el b5 7 �b3 0-0 8 c3 d5 9 exd5 lLlxd5 10 lLlxe5 lLlxe5 1 1 ':'xe5 c6 12 l:tel (D) 1 2 g3 is similar, and after 1 2 . . . .ltd6 1 3 l:te 1 we reach the same position. Black can also try the rare 1 2 . . . .ltf6, aiming for positional com­ pensation rather than a direct attack. After 1 3 l:te 1 c5 1 4 d4 ( 1 4 d3 b4 ! i s given by Nunn) 14 . . . .ltb7 15 dxc5 Black has : a) 1 5 . . . l:te8 1 6 lLld2 lLlxc3 1 7 bxc3 .ltxc3 1 8 c6 (this finesse is often praised, but I think 1 8 l:tb1 'ii'd7 1 9 l:he8+ ':'xe8 20 lLlfl ! ? 'ii'c 6 2 1 f3 'ii'x c5+ 22 �g2 is more critical) 1 8 . . . .ltxc6 1 9 l:txe8+ 'ii'x e8 20 l:tb1 :d8 2 1 'iWc2 .ltxd2 22 .ltxd2 .lte4 23 �xf7 + �xf7 24 'ii'b 3+ .ltd5 25 'ii'b4 'ii'e4 26 'ii'xe4 �xe4 27 :tb2 lld4 gave Black good winning chances despite the oppo­ site-coloured bishops because of the weak a2pawn in Braga-Geller, Amsterdam 1 986. b) 15 ... 'ii'd7 16 lLld2 (White could try 1 6 �e3 or 1 6 'ii'd 3) 1 6 . . . lLlxc3 1 7 bxc3 .ltxc3 1 8 ':'b1 l:.ad8 1 9 lle2 ? ! ( 1 9 l:te3 .ltxd2 20 .ltxd2 'ii'xd2 2 1 'ii'x d2 l:txd2 22 l:te7) 1 9 . . . 'ii'c 6 20 f3 'ii'x c5+ 2 1 'it>g2 .ltxf3+! gave Black counter­ play in Ernst-Dam, Lugano 1 988. Despite these successes for Black, 12 ... .ltf6 has not been popular at all. Of course part of this is due to the scarcity of games with 1 2 g3, but I think there is some suspicion concerning Black's play here. 12 �d6 13 g3 White can also play 1 3 d3, transposing to the main lines with 1 2 d3, or 1 3 d4, heading for the main lines with 1 2 d4. The text-move prevents . . . 'ii'h4, so Black has to find another way to create play. Black can still head straight to the kingside with 1 3 . . . 'ii'd 7, and •••

EARLY DEVIA TIONS

121

.i.d2 lIe2 2 1 g4 .i.xg4 ! 22 hxg4 'ile4 23 'fi'h l (23 'ilxe2 'ifxe2 24 :e 1 may be a better try, but White cannot be happy here) 23 . . . :xd2 24 :e l 'ilxg4+ 25 �fl ':xb2 26 i.xd5+ cxd5 27 liJe5 i.xe5 28 'ifxd5+ �f8 29 dxe5 'ii'h 3+ (Black should just trade queens with 29 . . . 'ifc4+ ! , when he can obviously try to win with no risk) 30 'i1a>g l 'ifg4+ 3 1 ..ti>fl 'ii'h 3+ 112- 112 was Z.Almasi­ Yakovenko, FIDE World Cup, Khanty-Man­ siisk 2007. 14 :xel+ 14 . . ..:.a7 was suggested by Aronian. Play could then transpose to the previous note. 15 'ifxe1 lIa7 16 .i.e3 After 1 6 liJd2 lIe7 1 7 'fifl Black has to play very accurately to obtain compensation accord­ ing to Aronian. After 1 7 . . . 'ife8 1 8 liJf3 Black could play 1 8 . . .f6 or 1 8 . . . i.g4. 16 :e7 17 liJd2 'ife8 (D) •••

after 14 d3 ( 1 4 d4 again reaches the main lines of 12 d4 after the reply 14 . . . 'it'h3), 14 . . . 'ii'h3 trans­ poses to Chapter 5. White may be happy with this though, because he has avoided the lines of Chapter 6. If Black does not want to head into Chapter 5, he has a couple of decent alternatives: 121 A : 13 :e8 1 22 B : 13 i.f5

•••

.•.

.•.

A) 13 ...:e8 (D)

18 liJn Safer is 1 8 i.xd5 cxd5 , but then Black's po­ sitional compensation is clearer because of the bishop-pair. 18 h5 19 a4 .i.e6 20 i.dl Anand embarks on a curious journey with this bishop. 20 i.c2 i.h3 2 1 'ile2 g6 is another possibility. 20 h4 21 axb5 axb5 22 i.f3 i.h3 23 i.xd5 It looks very strange to play i.d l -f3xd5 , but Black seems to have decent play in any event. 23 cxd5 24 'ifdl f5 25 i.g5 25 'ifb3 'iff7 26 i.g5 l:.e4 27 i.xh4 'i1a>h7 28 'ild l 'ilg6 29 i.d8 is similar to the main line. 25 :e4 (D) Aronian sacrifices a second pawn to increase the activity of his pieces. •••

This rare line was used by Aronian to defeat Anand. It looks a bit speCUlative, but has its logic. Black exchanges off White's active rook and may bring his other rook into play with the manoeuvre . . . :a7, which is more typical of the Sveshnikov Sicilian than the Ruy Lopez ! 14 d4 After 1 4 :xe8+ 'ifxe8 1 5 d4 :a7 1 6 liJd2 :e7 1 7 liJf3 f6 Black has reasonable compensation for the pawn. Then 1 8 c;!;>g2 i.g4 1 9 h3 i.h5 20

•••

.••

•••

UNDERSTANDING THE MARSHAll A TTACK

122

This allows a hammer-blow, but after 30 .lth4 l:lxd2 3 1 l:ta8+ .ltfS 32 ltxf8+ ..txfS 33 'fixd2 the calm 33 ... �g8 ! leaves Black with a winning attack.

w

26 .ltxh4 'fig6 27 .i.d8 f4 28 'fid3 (D) Aronian mentions a couple of alternatives, but he remains optimistic about Black's chances. a) 28 .ltb6 "is stupid" according to Aronian. 28 . . . l:.e8 29 .i.c5 .ltc7 and the c5-bishop is out of play. b) 28 .i.h4 f3 29 'fixf3 1::txh4 30 'fixd5+ �h7 3 1 l:ta8 'fie4 (3 l . . .'ifbl allows a perpetual check) 32 'fixe4+ l:txe4 33 �e3 is an odd mate­ rial situation where White has four pawns for a piece while Black has a powerful bishop-pair. Analysis engines tend to favour White here, at least slightly, while Aronian thinks that White is hoping for a draw. I have to agree with the hu­ man here - with two good bishops, Black cer­ tainly should not lose and may even press a little.

28 'fihS 29 �d2? 29 .i.b6 .i.b8 30 .i.c5 may be a better try, but with the white bishop away from the kingside, Black still has compensation for the pawns. 29 l:te2 30 �f3 (D) •••

•••

3O J:te3! 31 fxe3 'ii'xf3 32 'ii'c2 fxg3 33 hxg3 'ii'xg3+ 34 �h1 i.fS 0- 1 Anand-Aronian, Morelia/Linares 2008. ••

B) 13 .ltfS (D) •••

Black plays along the lines of Chapter 6, but here White will play d4, giving the position an original flavour. This line has been holding up very well for Black. 14 d4 'fid7 IS .lte3 The alternative 15 �d2 does not look terri­ bly dangerous. 1 5 . . . .l:.ae8 1 6 .l:.xe8 l:txe8 1 7 �f1 i.g4 1 8 f3 and here: a) 1 8 . . . i.h3 1 9 �f2 (this looks odd) 1 9 . . . 'ii'f5 ( 1 9 . . . h5 ! ?) 20 �e3 'ii'h 5 2 1 .ltxd5 cxd5 22 f4

EARLY DEVIATIONS

'iVh6 23 'ii'f3 �f5 ! and now 24 h4? ! �e4 gave Black excellent play in Bologan-Sargisian, Wijk aan Zee 2007 . 24 lLlxf5 ? ! 'ii'x h2+ 25 �fl (25 'it'g2 l:Ie2+) 25 . . . 'iVh3+ is also bad for White, so 24 Wg l is probably best, but Black still has good compensation after 24 . . . �e4. b) 1 8 . . . �h5 1 9 a4 ( 1 9 �xd5 cxd5 20 �f4?! �xf4 2 1 gxf4 �g6 22 'it'd2 h5 23 l:te l 1:txe l 24 'i!kxe l 'it'f5 25 'it'c l h4 was good for Black in A.lvanov-Geller, New York Open 1 990) 1 9 . . . h6 ( l 9 . . . 'ii'e6 ! ?) 20 axb5 axb5 2 1 �d2 'it'b7 22 �xd5 cxd5 23 �f4 �f8 24 g4 �g6 25 lLle3 �d6 26 lLlg2 (26 �xd6 l:he3) 26 . . . �f8 27 b4 'it'c6 was level in A.lvanov-Friedel, USA Ch, Tulsa 2008. lS l:tae8 1 5 . . . l:tfe8?! is worse because Black may want to push . . . f5-f4 later. 16 lLld2 Neither 1 6 �xd5 cxd5 nor 1 6 a4 �g4 100ks very dangerous for Black. 16 �g4 (D)

123

odd) 22 . . . �xc5+ 23 Wg2 l:te7 with excellent compensation for the pawn. c) 1 7 . . .�f5 is by far the most common move, continuing to bother the white queen. White can repeat moves or play 1 8 'it'c l (D), and then:

...

•••

17 'ii'b l White can also play 17 'ii'c 2. Then: a) 1 7 . . . l:te7 ! ? 1 8 �g5 l:txe l + 1 9 l:txe l h6 20 �e3 l:te8 2 1 c4 �f5 22 'ii'd l lLlb4 gave Black counterplay in Hou Yifan-Cheparinov, Wijk aan Zee 2008. b) PavloviC 's idea 17 ... lLlxe3 ! ? is the sim­ plest plan for Black if it works, and is certainly worth considering if Black wishes to avoid a repetition of moves. 1 8 fxe3 c5 1 9 a4 (Pavlovic also gives the line 1 9 "iic l h5 20 �d l cxd4 2 1 �xg4 'i!kxg4 2 2 exd4 h4 23 l:txe8 l:txe8 24 'it'd l 'ii'g 6, when Black is winning) 1 9 . . . �f5 ! 20 e4 �g6 2 1 axb5 axb5 22 dxc5 (this move looks

c l ) Anand gave 1 8 . . . h5 a ' ! ' in Informator, but it has not been repeated since. 1 9 lLlf3 �g4 20 lLlh4 l:te6 2 1 �d 1 f5 22 �xg4 hxg4 23 �g5 f4 gave Black compensation in Leko-Anand, Cap d' Agde rapid 2003 . c2) 1 8 . . . h6 is a rather typical move. Black is confident in his long-term compensation and simply takes the g5-square under control. 1 9 lLlfl �h3 2 0 'ii'd 2 'ii'f5 2 1 'it'd l 'ii'e4 2 2 f3 'it'g6 23 �xd5 (23 �f2 lLlf4 ! ?) 23 . . . cxd5 24 'it'd2 l:te6 25 a4 bxa4 26 l:txa4 l:tfe8 gave Black good play in lordachescu-Vajda, Serbian Team Ch, Budva 2004. c3) 1 8 . . . l:te7 1 9 lLlf3 (rather dry was 1 9 lLlfl l:tfe8 20 'it'd2 �h3 21 �c2 �xfl 22 l:txfl lLlxe3 23 fxe3 l:txe3 24 �b3 l:t3e7 25 l:tae l 1/2- 1/2 Z.Almasi-Gyimesi, Hungarian Ch, Kazincbar­ cika 2(05) 1 9 . . . �g4 20 lLlh4 l:Ife8 2 1 'it'd2 h6 22 'ii'd 3 g6 23 �d l �h3 (I would prefer 23 . . . �xd l 24 l:taxd 1 lLlxe3 25 fxe3 l:te4 with clear com­ pensation) 24 �f3 g5 25 lLlg2 �f5 was rather murky in Anand-Svidler, World Ch, Mexico City 2007. c4) 18 ... l:te6 1ooks very natural. 1 9 lLlf3 �g4 20 lLlg5 l:tg6 2 1 f3 ? (almost anything is better than this) 2 l . . .�xf3 ! 22 lLlxf3 �xg3 ! 23 hxg3 l:txg3+ 24 �f2 'it'h3 25 �e2 was Shabalov­ Aronian, Calvia Olympiad 2004. Now 25 ... l:tg2+ 26 �f2 l:.e8+ 27 lLle5 'it'f5 ! 28 l:Ifl lLlf4+ 29 'it'xf4 'it'xf4 intending . . . Wh8 and . . . f6 is win­ ning for Black.

UNDERSTANDING THE MARSHAll A rrA CK

124

17 ....i.f5 Like in the Adams Variation, Black seeks a perpetual attack on the white queen. There are a couple of alternatives: a) 17 ... 'fl.e7 ! ? is an interesting idea. After 1 8 ttJe4 :fe8 1 9 ttJxd6 'ilfxd6 20 'ilfd3 "it'f6 2 1 .i.c2 g6 22 'ilfd2 .i.h3 23 .i.d3 'ilff3 24 .i.fl .i.xfl 25 l:txfl h5 Black had good compensation in Frie­ del-Milman, US Chess League (Internet) 2007. b) 1 7 . . . f5 was suggested in New in Chess Yearbook 84. Pavlovic analyses this further: b I ) 1 8 f3 f4 ! 1 9 fxg4 fxe3 20 ttJe4 'iiih 8 2 1 ttJxd6 'ilfxd6 22 .i.xd5 cxd5 and Black i s cer­ tainly OK. b2) 1 8 f4 gives Black good long-term com­ pensation after 1 8 . . . :f7 1 9 .i.xd5 cxd5 20 'ilfd3 'fl.fe7 2 1 ttJb3 .i.f3 22 ttJc5 "it'c6 23 .i.f2 .i.e4, with counterplay. b3) 1 8 'ii'd 3 100ks more testing, but Pavlovic finds that Black has enough here as well: 1 8 . . . f4 1 9 .txd5+ cxd5 20 .txf4 .txf4 2 1 gxf4 l:txf4 22 l:txe8+ 'ilfxe8 23 'ilfe3 'ilffS 24 l:te 1 h6 25 h3 .th5 26 'ilfe6+ (or 26 ttJfl .tg6 27 ttJg3 'iii h 7 with compensation) 26 . . . 'iii h 7 27 'ilfxd5 l:txf2 28 l:tfl l:hfl + 29 ttJxfl .i.g6 30 'ilfg2 b4 with compensation. We now return to 1 7 . . . .i.f5 (D) :

18 .i.e2 1 8 'ilfc 1 transposes to note 'c' to White's 1 7th move above, while 1 8 'ilfd 1 invites a repetition with 1 8 . . . .i.g4. 18 .i.xe2 19 'ilfxe2 f5 20 e4 After 20 ttJf3 f4 2 1 ttJg5 ttJf6 22 'ilfb3+ (22 .i.d2 "it'g4 ! gives Black a strong attack) 22 . . . ttJd5 leaves White nothing better than repeating with 23 'ilfc2.

20 bxe4! This is much better than 20 . . . ttJf6? 2 1 .i.g5 ! ttJg4 22 cxb5 axb5 23 h3 ttJf6 24 .i.xf6 l:txf6 25 ':xe8+ 'ilfxe8 26 ttJf3 ± Iordachescu-Brunello, Reggio Emilia 2007. White is a healthy pawn up. 21 ttJxe4 (D) Black also has good chances after 2 1 'ilfxc4 f4. •••

21. f4 22 .i.d2? (D) This move is the quick choice of most en­ gines, but it is a fatal mistake. The only chance is 22 ttJxd6 fxe3 23 ttJxe8 exf2+ 24 'ilfxf2 l:txf2 25 'iiixf2 ttJe7 26 'fl.e4 'ilff5+ (26 . . . 'ilfxe8 27 'fl.ae 1 ) 27 'fl.f4 'ilfh5 28 l:le 1 "it'xh2+ 29 'iii f3 and now 29 . . . 'ilfh5+ 30 'iti>g2 'ilfxe8 3 1 :fe4 c5 32 l:txe7 'ilfc6+ is given as ''+' by Magomedov, but 33 d5 ! 'ilfxd5+ 34 'iii h 3 is probably a draw. Black could try 29 . . . ttJd5 30 ttJd6 h6 ! ?, which looks like a good winning attempt. ••

•••

22 f3! 23 'ilfd3 l:te2! 24 l:txe2 •••

EARLY DEVIATIONS

Also losing is 24 lDe3 ltxd2 25 'ii'xd2 'ifh3 ! with the idea 26 . . . lDxe3 27 fxe3 f2+. 24 'ii'h3 25 lDe3 After 25 'ii'xf3 .l:.xf3 26 lDxd6 lDf6 (intend­ ing . . . lDg4) 27 .l:.ae l h6 2S .l:.e3 .l:.xe3 29 i.xe3 lDg4 30 lDc4 'ii'xh2+ 3 1 ..tfl 'ii' h l + 32 ..te2 'ii'h 5 Black should win. 2S .l:.f4! (D)

A : 13 i.g4 B : 13 'ii'h4 •••

•••

125

1 25 1 26

•••

A) 13 .i.g4 (D) •••

•••

0- 1 Naiditsch-Gustafsson, European Ch, Dresden 2007. After 26 'ii'xa6 (or 26 gxf4 i.xf4) 26 . . . .l:.h4 ! 27 'ii'a S+ �f7 Black wins.

Section 7 . 2 1 e4 eS 2 lD0 lDc6 3 i.bS a6 4 i.a4 lDt'6 5 0-0 i.e7 6 ltel bS 7 i.b3 0-0 8 c3 cIS 9 exclS lDxclS 10 lDxeS lDxeS 11 :XeS c6 12 d4 i.d6 13 lte2 (D)

Putting the rook on the second rank was a popular plan of defence for a while, but it looks rather clumsy. Here we have:

This is less common than 1 3 . . . 'ii'h4, but it has been used by many of the world's top players. 14 0 jus This is rarer than 1 4 . . . i.h5, but it may well be stronger. 15 i.xd5 ( 1 5 'ii'f l ! ? may be more challenging; for example, 15 . . . i.g6 1 6 'ii'f2 'ii'f6 1 7 g3 i.d3 1 S .l:.e l 'ii'g 6 1 9 lDd2 f5 ? ! 20 f4 ..thS 2 1 lDf3 i.e4 22 lDe5 i.xe5 23 dxe5 ± Kinder­ mann-Nunn, Dortmund 1 99 1 ) 15 . . . cxd5 1 6 lDd2 f5 1 7 'ii'b 3 (or 1 7 lDfl f4) and now: a) 1 7 . . . .i.f7 l s lDfl f4 1 9 .i.d2 'it'd7 20 l:tae 1 a5 gave Black the initiative in Kamsky-Ivan­ chuk, Linares 1 99 1 . b) 1 7 . . .l:teS I S l:txeS+ 'ii'xeS 1 9 lDfl (not 1 9 'ii'x d5+? i.. f7 20 'ii'xd6 'it'e3+ 2 1 �fl l:teS 22 g3 i..c4+ 23 �g2 'ii'e 2+ 24 �h3 g5 ! -+ Nunn) 1 9 . . . i.. f7 20 i.d2 f4 2 1 a4 .l:.bS 22 axb5 .l:.xb5 23 'ii'c 2 i.g6 24 'ii'c 1 i.d3 25 i.xf4 i.xf4 26 'ii'xf4 .l:.xb2 gave Black adequate play in Kot­ ronias-Nunn, Kavala 1 99 1 . 1 5 i.xdS 15 g3 'ii'c 7 16 �f2 1ooks rather odd. 16 . . . 'ii'd7 1 7 .i.xd5 (White cannot play 1 7 lDd2? i..d 3 I S lte l 'ii'h 3 1 9 �gl because of 1 9 . . . i.xg3 ! , when 20 hxg3 'ii'x g3+ 2 1 �h l lDf4 mates quickly) 17 . . . cxd5 I S lDd2 i.d3 19 l:te3 i.g6 (already Black can force a draw with 1 9 . . . 'ii'h 3 20 ..tg l .i.xg3) 20 lDfl a5 2 1 a3 112- 112 Anand-Aronian, World Ch, Mexico City 2007. Black has excel­ lent compensation.

UNDERSTANDING THE MARSHALL A 1TACK

126

IS cxdS 16 lDd2 (D) •••

king too weak in Kotronias-I.Sokolov, Burgas 1 992. 19 lDn .i.xn 20 'ii'xn 'ii'b8 21 g3 'fixb4 Black had excellent play for the pawn in Kotronias-De Vreugt, Kavala 2002.

B) 13 'iIi'h4 (D) •••

16 .i.d3 Or 16 . . . b4 1 7 lDfl ( 1 7 cxb4 .i.d3 1 8 :f2 transposes to the main line, and was the move­ order of Kotronias-De Vreugt) 1 7 . . . bxc3 1 8 bxc3 'fic7 1 9 :e3 :fc8 gave Black plenty of compensation for the pawn in Todorovic-Pav­ lovic, Vrnjacka Banja 1 990. We shall take the 16 . . . .i.d3 move-order as our main line as this makes it easier to cover the deviations. 17 :e2 b4 (D) This minority attack is a typical plan when White has exchanged on d5 . Instead 1 7 . . . 'fic7 ? ! 1 8 g 3 :ae8 1 9 lDfl .i.g6 20 lDe3 hits the d5pawn, which was a bit annoying in Ki.Geor­ giev-Nikolic, Wijk aan Zee 1 988. •••

This is the traditional move. Black is not convinced that having the rook on e2 is a great improvement to White's position. 14 g3 1 4 h3 ? ! .i.g4 1 5 f3 .i.f5 leaves White too weak on the dark squares around his king. After the text-move ( 1 4 g3), Black has a choice between the opportunistic 1 4 . . . 'fih5 and the 'normal' 1 4 . . . 'it'h3 . 1 26 B l : 14 :ifhS 1 27 B2: 14 'ii' h3 ••

•••

B1) 14 'iIi'hS Black pins the rook and threatens . . . .i.g4. Is lDd2 After 1 5 :e4 Black should play 1 5 . . . 'iIi'g6 ! 1 6 .i.c2 .i.f5 1 7 :e2 :ae8 1 8 .i.xf5 'iii'xf5 1 9 'fie l ( 1 9 .i.e3 'fig6 ! intending . . . f5 i s given by Nunn) 1 9 . . . lDf6 20 .i.e3 :e6 2 1 lDd2 l:tfe8 22 'ii'b l 'it'h5 23 'it'd3 c5 with compensation, Hen­ nigan-Motwani, British Ch, Blackpool 1 988. IS .i.h3 1 5 . . . .i.g4 wins back the pawn, but after 16 f3 .i.xf3 1 7 lDxf3 'fixf3 1 8 :f2 or 1 8 'ili'fl White is a little better with the bishop-pair. •••

18 cxb4 :e8 Also possible is 1 8 . . . 'fic7 1 9 lDfl ? ! ( 1 9 g3 .i.c2 20 'ii'f l l:.fc8 ! ? gives Black good com­ pensation after either 2 1 lDb3 .i.xb4 or 2 1 a3 'fib6) 1 9 . . . .i.xfl 20 �xfl .i.xh2 left White's

•••

EARLY DEVIA TIONS

12 7

16 (3 �c7 (D) Prophylaxis against 1 7 lLle4. but White can still play it.

17 a4 1 7 lLle4 is a decent alternative. Black has: a) 17 . . . :tae8 ? ! 1 8 'ii'd 3 ! leaves White much better after both 1 8 . . . :e6 19 .i.d2 :g6 20 g4 ! Beliavsky-Malaniuk. Russian Ch. Minsk 1 987 and 18 . . .f5 1 9 lLlg5 f4 20 lLlxh3 fxg3 21 :xe8 :xe8 22 �g2 gxh2 23 �d2 ± Ehlvest-Geller. Vrsac 1 987. b) 17 ... 'it'xf3 18 lLlg5 'it'h5 19 lLlxh3 'it'xh3 20 �d2 1:.ae8 2 1 'it'fl 'it'd7 22 l:.ae l ':'xe2 23 'it'xe2 was Ehlvest-Nikolic. Zagreb Interzonal 1 987. White has the bishop-pair. but after Ehl­ vest's suggested 23 . . . �d8 with the idea . . . .i.f6 . . . g6 and . . . 1:.d8. Black is very solid. 17 b4 18 c4 lLlf6 19 lLle4! 'ii'g6? ! Nunn gives 1 9 . . . 'it'xf3 2 0 lLlg5 'it'h5 2 1 lLlxh3 'it'xh3 22 .i.g5 'it'f5 23 .i.h4 with the idea :f2 as being very bad for Black. but this may be Black's best at this point. After 23 . . . g5 ! ? mat­ ters are not so clear: 24 1:.f2 (or 24 �c2 'it'g4) 24 . . . 'it'g6 25 �c2 'it'g7. White must at least make sure he does not lose a piece ! 20 lLlf2 �fS 21 �c2 1:.fe8 22 �xfS 'it'xfS 23 1:.xe8+ 1:.xe8 24 �g2 ± Black had nothing for his pawn in Short­ Nunn. Brussels 1 986. although by playing for activity at all costs. Black managed to win any­ way !

1 5 1:.e4 transposes to the Modern Variation of Chapter 4. while 15 'it'fl 'it'h5 1 6 f3 (after 1 6 lLld2 Nunn gives 1 6 . . . lLlf4 ! . while 1 6 . . . �f5 and 1 6 . . . �h3 are also possible) 1 6 . . . �f5 (or 16 . . . .i.h3 17 'it'f2 f5) 17 'it'f2 'it'g6 100ks like good compensation for the pawn. IS �fS (D) ...

•

•••

82) 14 'ii'h3 (D) Is lLld2 •••

16 a4 White has several options here: a) 1 6 �xd5 is harmless at best. 16 . . . cxd5 1 7 f3 1:.ae8 ( 1 7 . . . 'it'h5 1 8 1:.f2 1:.ae8 1 9 lLlfl 1:.e6 20 a4 1:.fe8 also looks very good. Nijboer-Peelen. Amsterdam 1 984) 1 8 lLlfl h5 ! with compensa­ tion. Griinfeld-Pinter. Zagreb Interzonal 1 987. b) 16 'ili'fl 'ili'h5 and here: b l ) 1 7 �xd5 ? ! cxd5 1 8 'ii'e l .i.g4 1 9 :e3 f5 20 f4 g5 2 1 lLlfl :ae8 22 :e5 .i.h3 23 .i.e3 �xe5 24 dxe5 'it'f3 + Tischbierek-Blatny. Leip­ zig 1 988. b2) 1 7 iLd l 'ii'g 6 18 'ii'g 2 :ae8 looks OK for Black. Play is similar to the Spassky Varia­ tion and Black looks to be doing well.

UNDERSTANDING THE MARSHALL A TTA CK

128

b3) 1 7 l:te 1 lLlf4 (this idea is speculative, so 17 . . . l:tae8 is another possibility) 1 8 gxf4 il.xf4 1 9 "'g2 .lth3 20 "'xc6 il.xh2+ 2 1 'iii>h 1 il.g4 22 "'g2 l:tae8 23 l:txe8 .ltb8+ 24 'iii>g l l:txe8 2S lLle4 'iii>f8 ! (avoiding 2S . . . .ltf3 ? 26 lLlf6+) gave Black good play in M.Schlosser-Adams, Oak­ ham 1 990. c) 1 6 lLle4 i.g4 and now: c 1 ) 1 7 lLlxd6? is bad in view of 1 7 . . ....hS ! 1 8 �fl ( 1 8 f3 i.xf3 1 9 'ifi>f2 l:.ad8 20 lLle4 l:tde8 -+) 1 8 . . ....xh2 1 9 �e 1 "'h 1 + 20 �d2 "'h6+ 2 1 �c2 "'xd6 -+ Zuckennan-Pavlovic, New York Open 1 987. c2) 1 7 "'fl "'xfl + 18 �xfl l:tae8 1 9 lLlxd6 .ltxe2+ 20 'iii>g 2 :e6 is a little better for Black. c3) 17 lLlgS is best. After 1 7 . . . il.xe2 1 8 "'xe2 "'d7 1 9 "'d3 White had enough com­ pensation in Geller-Lukacs, Coimbatore 1 987. Here 19 ... g6 100ks best. d) 16 i.c2 i.xc2 17 "'xc2 fS 1 8 c4 ( 1 8 f4 is well met by 18 . . . i.xf4 ! , when 19 gxf4? loses to 1 9 . . ....g4+ and now 20 'iii> f2 lLlxf4 or 20 l:tg2 "'xg2+ 21 �xg2 lLle3+) 1 8 . . ....g4 and here: d 1 ) 19 l:te 1 f4 20 f3 "'h3 21 cxdS (2 1 lLle4 can be met by 2 1 . . .lLlb4) 2 1 . . .fxg3 22 lLlfl ? (22 lLle4) 22 . . . gxh2+ 23 'iii>h 1 (Mokry-Panczyk, Polanica Zdroj 1 984) 23 . . . .ltb4 ! +. d2) 19 l:.e6 and now: d2 1 ) 1 9 . . . lLlf4 leads to a draw: 20 f3 (20 lhd6? l:tae8 2 1 cxbS l:te2 22 "'c4+ 'iii>h 8 23 "'xe2 lLlxe2+ 24 'iii>g 2 f4 2S bxc6 fxg3 26 hxg3 lLlf4+ 0- 1 Ljubojevic-Nunn, Szirak Interzonal 1 987) 20 . . . lLlh3+ 2 1 �g2 lLlf4+ 22 'ifi>g l lLlh3+ 23 'iii>g 2 lLlf4+ '12- '12 HUbner-Timman, Tilburg 1 987. d22) 19 ... f4 ! is given by Nunn. 20 "'e4 (White loses quickly after both 20 cxdS ? fxg3 2 1 hxg3 il.xg3 22 f3 "'xd4+ and 20 l:txd6? fxg3 21 hxg3 { 2 1 fxg3 lLle3 } 2 l . . .l:txf2 ! ) 20 . . . lLle3 ! 2 1 lLlf3 (bad are 2 1 fxe3 ? "'d 1 + 22 'iii>g 2 fxg3 23 hxg3 "'e2+ 24 'ifi>h3 :f2 and 2 1 lLlfl ? lLlxfl 22 'iii>xfl 'it'h3+ 23 'iii>g l fxg3 24 hxg3 il.xg3 -+) 2 1 . . .fxg3 22 "'xg4 gxf2+ (22 . . . gxh2+ ! looks even better: 23 lLlxh2 .ltxh2+ 24 'iii>xh2 lLlxg4+ 2S �g3 lLlxf2 +) 23 'ifi>xf2 lLlxg4+ 24 'iii>g 2 l:tf6 (Nunn) is only a tiny bit better for Black. 16 i.d3 Also possible is 1 6 . . . l:.ae8 1 7 l:.xe8 l:.xe8 1 8 lLlfl hS 1 9 i.xdS ( 1 9 axbS ? ! axbS 20 l:.a6 is worse because of 20 . . . lLlc7 ! 21 l:ta7 { not 2 1 l:.xc6? i.e4 } 2 1 . . .h4 2 2 lLle3 ? hxg3 2 3 fxg3 •••

il.xg3 24 hxg3 l:txe3 2S il.xe3 "'xg3+ 26 'iii>f l .lth3+ 27 'iii>e 2 .ltg4+, de Finnian-Adams, New York 1 996) 1 9 . . . cxdS 20 axbS axbS 2 1 lLle3, as in Sax-Nunn, Brussels 1 988. White is some­ what better because his position is very solid and he has counterplay against Black ' s pawns on bS and dS . 17 :e1 l:taeS IS lLlf3 l:.xe1 + 19 'ii'x e1 (D)

B

Black is active, but it is not easy to improve his position. 19 h6 20 axbS axbS 21 lLleS White could also try 2 1 "'d 1 il.e4 22 il.c2, when Black must still prove that he has enough for the pawn. 21...l:teS 22 "'dl il.xeS 23 dxeS 'ii'fS Black has sufficient activity for the pawn. 24 .ltxdS (after 24 f4 Black can play either 24 . . . hS or 24 . . ....e4) 24 . . . cxdS 2S g4 "'g6 26 il.f4 "'e4 27 h3 hS 28 i.g3 i.e2 29 "'d4 '12- '12 Anand­ Khalifman, Reggio Emilia 1 99 1 . •••

Section 7 . 3 1 e4 eS 2 lLlf3 lLlc6 3 il.bS a6 4 il.a4 lLlf6 5 0-0 il.e7 6 l:tel bS 7 il.b3 0-0 S c3 dS 9 exdS lLlxdS 10 lLlxeS lLlxeS 1 1 :xeS c6 12 il.xdS cxdS 13 d4 il.d6 14 l:te3 (D) This is often referred to as the Kevitz Varia­ tion. Instead 14 l:te 1 "'h4 I S g3 "'h3 is consid­ ered in Section 7.S by way of the move-order 1 2 d4 .ltd6 1 3 l:te 1 "'h4 14 g3 "'h3 I S il.xdS cxdS . 14......h4 14 . . . fS ? ! looks insufficient after I S lLld2 f4 1 6 l:te 1 , as it is not so easy for Black to attack

EARLY DEVIATIONS

in this position. His far-advanced f-pawn cuts off his own dark-squared bishop. Black has tried: a) 16 . . .l:Ia7 17 liJf3 .i.g4 ( 1 7 . . . g5 1 8 liJe5 :g7 1 9 a4 ±) 1 8 h3 .i.h5 1 9 a4 ±. b) 1 6 . . . 'ii'f6 looks insufficient after 1 7 'ii'b 3 'ii'f5 (or 17 . . . .i.e6 1 8 liJf3 ;!;) 1 8 liJe4 ( 1 8 liJf3) 1 8 . . . .i.c7 1 9 f3 with the idea 20 liJf2. Black does not have enough. c) 16 . . . f3 is the most direct, but it does not really work. 17 liJxf3 .i.g4 1 8 :e3 .i.f4 (or 1 8 . . . l:ta7 19 h3 .i.h5 20 g4 .i.g6 2 1 liJe5 .i.xe5 22 dxe5 ±) 1 9 :d3 .i.c7 ( 1 9 . . . .i.f5 20 .i.xf4 .i.xd3 fails to the zwischenzug 2 1 .i.g5) 20 h3 .i.h5 2 1 .i.g5 ! 'ii'd6 (2 1 . . . .i.xf3 22 .i.xd8 .i.xd l 23 .i.xc7) 22 g4 .i.g6 23 ':'e3 .i.e4 24 liJe5 ±. d) 16 ... 'iVg5 1 7 liJf3 'iVh5 1 8 liJe5 ! (worse is 1 8 'ii'b 3 �h8 19 liJe5 .i.e6 20 f3 l::tf6 21 l:te2 l::t af8 - Nunn) 1 8 . . . f3 1 9 gxf3 ! (this is much better than 1 9 liJxf3 ? .i.g4 =+= or 19 l::te 3 .i.xe5 20 l:txe5 'ii'g4 2 1 l:tg5 'ii'e4 with compensation). Black may have some practical chances but ob­ jectively White should be much better. However, the rarer 1 4 . . . 'ii'c 7 is playable: 1 5 h3 .i.f5 1 6 liJd2 and now 1 6 . . . .i.h2+ ! i s a clever move, whose point will become apparent with Black's 20th and 2 1 st moves. 1 7 �h l .i.f4 1 8 l:te l l:tae8 gives Black compensation for the pawn; e.g., 1 9 liJn ( 1 9 liJf3 .i.d6 20 .i.e3 'iVd7 2 1 liJg l l:te6 22 'iVd2 l:tfe8 with compensation) 19 . . . .i.d6 ! 20 .i.g5 'iVd7 ! 21 'iVd2 .i.xh3 ! 22 gxh3 'iVxh3+ 23 �gl 'iVg4+ 24 liJg3 h6 25 .i.e3 ? ! (25 .i.xh6 ! ?) 25 . . . .i.xg3 26 fxg3 'iVxg3+ 27 �h l (27 �n l:te6 -+) 27 . . . 'iVh3+ 28 'iVh2 l:txe3 -+ Mithrakanth-Yurtaev, Calcutta 2000. Now back to 14 . . . 'iVh4 (D). lS h3

129

15 g3 is inconsistent. After 15 . . . .i.g4 ! ? (the less forcing 15 . . . 'ii'h3 1 6 liJd2 .i.g4 also offers Black adequate play) 1 6 f3 (bad is 1 6 'ii'n ? 'ii'h5 17 liJd2 f5 1 8 f4 l:tfe8 1 9 liJb3 .i.h3 20 'iVe 1 .i.xf4 -+, while 1 6 gxh4 .i.xd l gives Black ob­ vious compensation in the endgame) 1 6 . . . .i.xg3 (this should lead to a draw) 17 'iVe2 ( 1 7 hxg3 'iVxg3+ 1 8 �n ? ! .i.h3+ 19 �e2 'iVg2+ 20 �e l l:tae8 2 1 'iVe2 'iVh 1 + 22 �d2 .i.n 23 'iVf2 l:txe3 24 'iVxe3 h5 =+= Nunn) 1 7 . . . .i.f4 1 8 fxg4 f5 ! 1 9 gxf5 ( 1 9 l:th3 ? ! 'iVxh3 20 .i.xf4 fxg4 2 1 .i.e5 11f3 22 liJd2 l:.e3 23 'ii'n :f8 24 'iVxh3 gxh3 '+) 19 . . .l:txf5 20 :e8+ :xe8 2 1 'ii'xe8+ :f8 22 'ii'e6+ �h8 23 .i.xf4 'ii'xf4 is a draw. Black will deliver perpetual check after 24 'ii'e2 or 24 'iVe 1 . lS 'iIi'f4 This is the soundest approach. Black can also try to attack with his kingside pawns: a) 15 . . . f5 ? ! looks natural, but Black has ex­ perienced difficulties: al) 1 6 liJd2 f4 1 7 l:te l ( 1 7 liJf3 ? ! 'iVh5 1 8 l:te l .i.xh3 1 9 gxh3 'iVxh3 20 l:te5 ? l:lae8 2 1 :g5 :f6 2 2 l:tg2 l:Ih6 2 3 liJh2 f3 0- 1 Emst­ Kling, COIT. 1 987) 17 . . . .i.xh3 ! ? 1 8 'ii'f3 .i.xg2 19 'iVxg2 'ii'h5 20 liJf3 :f6 was an interesting, though speculative, approach in de Firmian­ I.Sokolov, Bie1 1 989. a2) 16 'iVb3 .i.b7 1 7 liJd2 f4 1 8 liJf3 'ii'h5 1 9 l:te l ( 1 9 :e6? l:Iad8 20 l::te l .i.c8 2 1 liJe5 .i.xe5 22 l:txe5 'iVg6 23 �h2 f3 gives Black good at­ tacking chances) 1 9 . . .l:tf6 20 a4 l:tg6 2 1 �h l l:tf8 22 axb5 (22 'iVd 1 ? .i.c8 -+) 22 . . . .i.c8 23 l:te5 .i.xe5 24 'iVxd5+ .i.e6 25 'iVxe5 l:tf5 26 'iVb8+ ! (else . . . .i.d5) 26 . . . l:tf8 27 'iVe5 repeats. a3) 1 6 'iVf3 ! .i.b7 1 7 liJd2 g5 1 8 'iVe2 ! f4 ( l 8 . . . g4 also looks insufficient after 1 9 l:te6 :ad8 20 liJn gxh3 2 1 g3 ! 'ii'g4 22 .i.h6 'ii'xe2 •••

130

UNDERSTANDING THE MARSHAll A TTACK

23 %lxe2 :fe8 24 .l:tae 1 ±) 1 9 �f3 'iWh5 20 �xg5 ! 'iWg6 (20 . . . 'iWxg5 2 1 .l:tg3) 2 1 .l:te6 'iWxg5 22 ':xd6 %lae8 23 .l:te6 rJ;fl 24 l:.e5 +- Hubner­ Nunn, World Cup, Skelleftea 1 989. b) 15 ... g5 seems cruder, but it is playable. 1 6 'iWf3 (after 1 6 �d2 g4 1 7 �fl both 1 7 . . . rJ;h8 1 8 hxg4 i.xg4 and 1 7 . . .gxh3 1 8 g 3 'iWg4 1 9 'iWxg4+ i.xg4, as in Stolz-Van den Doe1, Bundesliga 20001 1 , give Black enough play) 1 6 . . . i.e6 1 7 'iWf6 (D) and now:

hxg6 with compensation, Hubner-Nunn, Euro­ pean Team Ch, Haifa 1 989. We now return to 15 ... 'iWf4 (D) :

w

B

b 1 ) 1 7 . . . 'iWh5 ? ! 1 8 �d2 g4 1 9 %1xe6 ! fxe6 20 'iWxe6+ 'iWfl 2 1 'iWxd6 'iWxf2+ 22 �h2 %lae8 23 'iWg3 was given as ;l; by Tal, but this looks more like ±. In any case, Black should steer clear of this line. b2) 17 .. Jlae8? ! 1 8 �a3 'it'h5 1 9 i.d2 h6 ( 1 9 . . .i.f4 20 .l:.d3 i.xd2 2 1 .l:.xd2 i.xh3 ? 22 gxh3 'iWxh3 23 'iWxg5+ �h8 24 'it'f6+ �g8 25 %ld3 'it'xd3 26 �h2 +- wins for White, but this does not work with the rook on a8 instead of f8 ! ) 20 �c2 i.f4 2 1 .l:.d3 'it'e2 22 i.xf4 gxf4 23 �e 1 'it'xb2 24 .l:.ad 1 �h7 25 'iWxf4 'iWxa2 26 %le3 %lc8 27 �d3 'ii'c 2 28 l:a1 'it'xc3 29 .l:.xa6 ± Hubner-Pinter, Solingen 1 989. b3) 17 ... .:fe8 ! 18 �a3 ( 1 8 �d2 'it'f4 1 9 'iWxf4 i.xf4 20 .l:te l i.xh3 2 1 ':'xe8+ %lxe8 22 �f3 i.xc 1 23 ':xc 1 { this actually came about from 1 7 . . . :ae8 } 23 . . . i.f5 24 %le I %lxe l + 25 �xe l f6 = Brkic-Naiditsch, European Ch, Kusa­ dasi 2006) 1 8 . . . 'it'h5 1 9 i.d2 i.e7 20 'iff3 (or 20 'it'e5 .l:.ad8 2 1 f4 'it'g6 22 fxg5 i.f5 23 'it'g3 i.e4 24 .l:.fl i.d6 25 'it'h4 .l:.e6 26 .l:.e2 .l:.de8 27 .l:.fe l f6 28 .l:.f2 i.e7 29 gxf6 .l:.xf6 30 .l:.ee2 l:txf2 3 1 'iWxf2 .l:.f8 with compensation, Wolff­ Hellers, New York Open 1 990) 20 . . .'it'g6 2 1 .l:.ae l g4 22 'iWg3 gxh3 2 3 gxh3 i.d6 ! 24 'iWxg6+

16 .l:.eS The time-wasting 1 6 .l:.g3 ? ! is bad: 1 6 .. .'.f5 (also good is 1 6 . . .'it'f6 1 7 .l:.f3 'it'g6 1 8 'it'd3 .l:.e8 19 i.d2 .l:.e4 20 �a3 i.f5 21 'it'fl .l:.h4 ! + Boudy-Am.Rodriguez, Capablanca Memorial, Cienfuegos 1 983) 17 .l:.f3 'it'h5 1 8 'it'd3 i.g4 ! 1 9 .l:.e3 :ae8 20 �a3 i.f5 2 1 'iWfl 'it'g6 + Barcenilla-Van Gisbergen, World Junior Ch, Mamaia 1 99 1 . 1 6 'iWf6 17 ':e1 (D) 1 7 %lxd5 ? ! i.b7 1 8 .l:.g5 ':fe8 gives Black a strong attack, while White can offer a repeti­ tion with 1 7 %le3 'iWf4 1 8 %le5 'iWf6 1 9 .l:te3 . Black can decline to repeat, but there are risks : 1 9 . . . 'iWg6 20 �d2 f5 2 1 'it'f3 f4 22 l:te l (22 'it'xd5+ �h8 23 .l:.e5 ! i.xe5 24 'it'xa8 looks critical) 22 . . . �h8 23 �b3 i.f5 24 i.d2 h5 ! with compensation, Gufeld-Blatny, Honolulu 1 996. •••

EARLY DEVIA TIONS

17 Ji'g6 IS 'ii'f3 1 8
•••

131

'ii'd 3 with compensation, J.Polgar-Z.Almasi, FIDE Knockout, Groningen 1 997. 20 'ii'e2 'ii'g6 21
Section 7 .4 1 e4 eS 2 lLlf3 lLlc6 3 J.bS a6 4 J.a4 lLlf6 S 0-0 J.e7 6 ':'eI bS 7 J.b3 0-0 S c3 dS 9 exdS lLlxdS 10 lLlxeS lLlxeS 11 ':'xeS c6 12 d4 J.d6 13 ':'el 'ii'h4 14 g3 'ii'h3 IS 'ii'e2 (D)

B

This is a relatively new try which has been popular lately. It is amazing that there can be such fertile ground at such an early stage in a well-analysed opening like the Marshall At­ tack ! White hopes to exchange or drive the black queen away as quickly as possible by playing 'ii'f l . We examine: 131 A : IS J.g4 1 32 B : IS J.d7 .••

•••

19 'ii'c2 This is certainly good enough, but Black can also play 1 9 . . . ':'ac8 20 lLld2 b4 2 1 cxb4 J.xb4 22 a3 J.d6 23 ':ac 1 ':c2 24 b4 ':xc 1 25 l:txc 1

A)

••.

IS J.g4 Black continues in the normal fashion. 16 'ii'n 'ilVhS •••

UNDERSTANDING THE MARSHAll A TTA CK

132

This is sharper than 1 6 . . . l:tae8 1 7 l:txe8 'it'xfl + 1 8 �xfl l:txe8 19 .td2 .tf5 20 a4 (cer­ tainly better than 20 lDa3 ? ! , when 20 . . . .td3+ 21 �g2 l:te2 22 l:td l .txa3 23 bxa3 .te4+ 24 �fl .td3 was soon drawn in C.Balogh-Z.AI­ masi, Hungarian Ch, Nyiregyhaza 2008, and Mikhalevski's suggestion 20 . . . .th3+ 2 1 �gl :e2 22 %:td l lDf6 23 lDc2 .i.g4 is probably even better) 20 . . . h6 2 1 axb5 axb5 22 .i.d l ! g5 23 lDa3 , which left White a little better in Volo­ kitin-Inarkiev, European Ch, Budva 2009. 17 lDd2 1 7 .te3 l:tae8 1 8 lDd2 l:te6 1 9 f3 is Kram­ nik's line in Section 3 .4. 17 l:tae8 18 f3 (D) After 1 8 ':xe8 l:txe8 1 9 f3, 1 9 . . . .tf5 ? ! 20 lDe4 ! .tc7 (20 . . . .txe4 2 1 fxe4 l:txe4 loses to 22 .td l ! 'it'g6 23 .tc2) 2 1 .td2 'it'g6 22 l:te l .txe4 23 l:txe4 l:txe4 24 .tc2 f5 25 fxe4 fxe4 26 'it'g2 was better for White in Svidler-Leko, World Ch, Mexico City 2007 . Black should prefer 1 9 . . . .th3 ! 20 'it'f2 .tf5, because 2 1 lDe4? can simply be met by 2 1 . . ..txe4 22 fxe4 ':xe4 be­ cause 23 .td l is not possible.

an improvement in Volokitin-Kumosov, Euro­ pean Ch, Dagomys 2008) 22 'it'd l 'it'f5 23 .tc2 'iVh3 24 'it'f3 ;I; Caruana-Negi, Wijk aan Zee 2008. White ' s bishop-pair gives him a certain initiative. 19 fxg4 fxg4 20 l:Ixe8 'it'xe8 21 'it'g2 'it'e3+ 22 �hl :f2 (D)

•••

23 'ilfgl Mikhalevski suggests 23 .txd5+ ! cxd5 24 lDfl 'it'e2 25 'ii'x d5+ �h8 26 .te3 ! l:txfl + 27 l:txfl 'it'xfl + 28 .tgl ;1;. 23 .txg3! 24 'it'xg3 'it'e2! 2S 'it'gl �h8! 26 .tdl l:txh2+ 27 'it'xh2 'it'el+ 112_ 112 So-Gupta, Wijk aan Zee 2009. •••

B) IS .td7 (D) •••

18 fS! ? A sharp try. Black has also played: a) 1 8 . . . .th3 1 9 'it'f2 l:txe l + (Black may be better off with 1 9 . . . f5 20 ':'xe8 l:txe8 2 1 c4 lDf4 ! , which was unclear in Kapengut-Malan­ iuk, Minsk 1 985) 20 'it'xe l lDf6 2 1 'it'f2 l:te8 22 lDfl 'it'g6 23 .i.d2 h5 24 l:te 1 l:txe 1 25 .txe 1 was Motylev-Vajda, Romanian Team Ch, Pre­ deal 2007 . White has managed to exchange all of the rooks and still has an extra pawn. b) 1 8 . . . ':xe l 19 'it'xe l .txf3 20 lDxf3 'it'xf3 2 1 .td2 h5 (2 1 . . . 'it'g4 22 'it'd 1 was not much of •••

Instead of attacking the white queen, which intended to move anyway, Black focuses on seizing the e-file.

EARLY DEVIATIONS

133

16 'ii'n 'ii'fS! ? This i s Black's latest try, and we shall take it as the main line because it was Aronian' s choice. Other moves have been seen more frequently: a) 1 6 . . . 'ii'h S is popular. 17 lLld2 l:tae8 (D) and here:

a l ) 1 8 f3 lLlf4 1 9 1he8 1he8 20 lLle4 lLlh3+ 21 rJilg2 (not 21 h 1 ? .i.xg3 22 hxg3 lLlgS+ 23 rJilg1 lLlxf3+ 24 f2 ':xe4 2S .i.d 1 .i.g4 and Black wins) 2 l . . .lLlf4+ 22 gl lLlh3+ 23 rJilg2 112- 112 Kasimdzhanov-Bacrot, French Team Ch, Noyon 2008. a2) 1 8 lLle4 .i.h3 1 9 .i.d 1 (not 19 'ii'e2 'ii'xe2 20 l:txe2 fS) 1 9 .. :iffS 20 'ii'd 3 'ikg6 (this is better than 20 .. :ife6 2 1 l:te2 .i.fS 22 f3 hS 23 .i.gS 'ii'g 6 24 'ii'd2 .i.c7 2S .i.c2 f6 26 .i.h4 l:te6 27 ':ae 1 :lfe8? 28 lLlxf6+ ! , as in Volokitin­ Gustafsson, Bundesliga 2008/9) 2 1 .i.d2 .i.fS 22 .i.f3 .i.f4 ! 23 .i.xf4 lLlxf4 24 'ii'd2 lLlh3+ 2S rJilg2 1he4 26 .i.xe4 .i.xe4+ 27 l:txe4 'ii'xe4+ 28 rJilxh3 l:te8 gave Black enough compensation in Volokitin-Fressinet, European Team Ch, Kher­ sonisos 2007 and the game was soon drawn. b) 16 . . . l:tae8 heads for an endgame. 1 7 ltxe8 'ii'x fl + 1 8 rJilxfl l:txe8 1 9 .i.d2 lLlb6 20 .i.e3 lLldS 2 1 .i.xdS cxdS 22 lLld2 as 23 lLlb3 a4 24 lLlcs (D) and now: b 1 ) 24 . . . .i.h3+ 2S rJile2 l:ta8? ! 26 .i.f4 .i.e7 27 rJild2 gS 28 .i.c7 ltc8 29 .i.b6 ':c6 30 l:te 1 ! .i.f8 3 1 .i.d8 .i.xcs 32 dxcS ':xcS 33 .i.e7 l:tc4 34 .i.xgS d4 3S .i.h6 f6 36 lte7 and White kept pressing in Kamsky-Bacrot, FIDE Grand Prix, Nalchik 2009. b2) 24 . . . .i.fS 2S rJile2 f6 26 d2 gS 27 b3 a3 28 b4 rJilf7 29 lLld3 rJilg6 30 ltb1 hS with com­ pensation, Caruana-Sargisian, Merida 2008.

17 .i.e3 Black holds the balance after 1 7 lLld2 :ae8 1 8 f3 lLlf6 1 9 a4 cS 20 l:txe8 ':xe8 2 1 lLle4 ':xe4 ! ; e.g., 22 fxe4 (22 .i.c2 cxd4 ! 23 fxe4 'ii'c S ! gave Black enough play in Svidler-Gri­ shchuk, Russian Ch, Moscow 2009) 22 . . . 'ii'xe4 23 .i.f4 ! ? .i.c6 ! 24 dS ! c4 ! 2S .i.xd6 'ii'x dS 26 .i.d 1 'ii'h 1 + 27 f2 'ii'xh2+ 28 rJile 1 'ii'x b2 ! 29 l:ta3 lLle4 30 'ii'f4 'ii'g 2 3 1 .i.e2 'ii'g l + 32 .i.fl gS 33 'ii'f3 lLld2 ! ? 34 'ili'xc6 'ii'x fl + 3S rJilxd2 'ii'f2+ 36 rJilc 1 'ii'e 1 + 112- 112 L.Dorninguez-Leko, Wijk aan Zee 20 1 0. 17 l:taeS Is lLld2 hS Here is one point of Black's l Sth move - he can bring another unit into play. This looks good enough, but Black could also try 1 8 . . . :le7 . 19 'ii'g2 h4 (D) •••

w

20 'ii'fJ Other moves: a) 20 .i.xdS cxdS 21 'ii'f3 hxg3 22 hxg3 'ii'xf3 23 lLlxf3 f6 24 .i.f4 .i.e7 2S .i.c7 rJilf7 and Black held the endgame without much difficulty

UNDERSTANDING THE MARSHALL A TTA CK

134

in Balogh-Hracek, European Team Ch, Novi Sad 2009. b) 20 .:tac 1 100ks rather odd. Then 20 . . . 'ii'h5 2 1 ttJe4 i.b8 22 i.d2 i.g4 23 ttJg5 h3 24 'ifn .:txe l 25 .:txe l f6 26 ttJe6 .:te8 27 ttJc5 .:te2 28 i.c 1 ? ! i.xg3 ! was good for Black in Banikas­ Melkumyan, Kavala 2009, but 20 . . . hxg3 2 1 hxg3 ttJxe3 22 l:txe3 (22 fxe3 'ifg6) 22 . . .l:he3 23 fxe3 'ifg5 24 ttJn c5, with compensation, looks simpler. 20 ttJxe3 21 l:txe3 (D) 2 1 fxe3 hxg3 22 hxg3 'ifg6 ! ? gives Black good play for the pawn.

22 'ifg6 23 'it>g2 i.g4 24 'iff2 e5 25 .:tel After 25 e4 cxd4 26 cxd4 i.b4 Black gets se­ rious counterplay. 25 i.f5 (D) •.•

•••

w

•••

B

Black has excellent play so White hurries to simplify the position. 26 e4 i.xe4+ 27 ttJxe4 'iWxe4+ 28 'ii'fJ 'iWxf3+ 29 �xf3 hxg3 30 hxg3 ':e8 = Alekseev-Aronian, FIDE Grand Prix, Nal­ chik 2009.

Section 7 . 5 21.. .:txe3 It would be simpler to play 2 1 . . .'ifxf3 22 ttJxf3 hxg3 23 hxg3 ':'xe3 24 fxe3 i.xg3. 22 fxe3 (D) 22 'ifxe3 .:te8 23 'iff3 and now the natural 23 . . . 'ifxf3 24 ttJxf3 hxg3 25 hxg3 .:te2 26 ttJg5 i.e8? (26 . . . ':xb2 ! is necessary) runs into 27 'it>n .:txb2 28 .:te l ! . Instead Black could try 23 . . . c5 with counterplay. •

B

1 e4 e5 2 ttJf3 ttJe6 3 i.b5 a6 4 i.a4 ttJf6 5 0-0 i.e7 6 .:tel b5 7 i.b3 0-0 8 e3 d5 9 exd5 ttJxd5 10 ttJxe5 ttJxe5 11 .:txe5 e6 12 d4 i.d6 13 .:tel 'ifh4 l4 g3 'iWh3 15 i.xd5 As we have seen, White may play this cap­ ture at almost any time, but here it is played im­ mediately because White wants to attack the d5-pawn. Instead 1 5 i.e3 is the Main Line of the first three chapters, while 1 5 ':e4 is the Modern Variation of Chapter 4. Here we shall consider other 1 5th moves as well: a) 1 5 ttJd2? is bad: 1 5 . . . i.g4 1 6 ttJf3 (the al­ ternative 1 6 f3 i.xg3 1 7 hxg3 'ifxg3+ 1 8 'it>h l ttJf4 also wins for Black) 1 6 . . . 'ifh5 1 7 'it>g2 .:tae8 1 8 .:txe8 ':xe8 19 i.d2 .:te6 is winning for Black. b) 1 5 'iff3 ? ! i.g4 16 'ifg2 'ifh5 17 i.e3 (even worse are 17 ttJd2 ? ! .:tae8 =+= and 1 7 i.xd5 ? ! cxd5 1 8 i.e3 i.f3 1 9 'ifn f5 20 ttJd2 f4 =+=) 1 7 . . . i.f3 1 8 'ifn f5 1 9 ttJd2 f4 20 ttJxf3 fxe3 2 1 i.d l and here Nunn's suggestion 2 1 . . . .:tae8 ! looks stronger than 2 1 . . . .:txf3 22 i.xf3 'ifxf3 23

EARLY DEVIA TIONS

135

fxe3, which was unclear in Normoyle-Hebden, Irish Ch, Dublin 2007 . c) 1 5 'ifd3 and then: c l ) 15 . . . i.g4 16 'iffl transposes to Section 7 .4. c2) 15 ... lLlf6 16 'iffl 'iff5 17 i.e3 i.b7 intending . . . c5 looks like a reasonable try. c3) 1 5 . . . .i.f5 1 6 'iffl 'ifh5 1 7 i.e3 i.h3 ( 1 7 . . . :ae8 1 8 lLld2 transposes to 1 8 'iffl from Chapter 3) 1 8 .i.d l ( 1 8 'ife2 i.g4) 1 8 . . . 'it'f5 1 9 'it'e2 :ae8 20 lLld2 c 5 could use some tests. Play is similar to the Spassky Variation. 15 cxd5 (D) •••

l:.de8 gives B lack some compensation, but White can fight back with 25 a4. a2) 20 . . . l:.de8 2 1 'it'e2 and now 2 l . . .'ifd5 22 i.e3 (22 'iffl ? i.h3) 22 . . . i.h3 23 f3 f5 is given by Nunn, while the surprising 2 l . . .i.d5 ! ? is also possible; for example, 22 i.e3 i.f3 23 'it'fl f5 with good play. b) 1 7 lLld2 l:.ae8 1 8 l:.e3 l:.e6 ( 1 8 . . . .:txe3 ! ? is also possible, with the idea 19 fxe3 :e8 or 1 9 'it'xe3 i.d7) 1 9 'ifxd5 (D) and here:

B

This position is important because, as men­ tioned before, it may also arise via the move­ order 1 2 i.xd5 cxd5 1 3 d4 i.d6 14 l:.e 1 'it'h4 1 5 g 3 'it'h3 . 16 'it'f3 i.f5 ! (D) This is considered much better than 1 6 . . . i.g4 1 7 'ifxd5, when after 1 7 . . . l:.ae8 ? ! 1 8 i.e3 Black will have trouble mustering enough play for the two pawns. Black could try 1 7 . . . l:.ad8, with the idea 1 8 'it'g2 (this is the most natural move, but may not be best) 1 8 . . . 'it'h5 1 9 .i.e3 .i.f3 20 'iffl f5 2 1 lLld2 f4 with serious counterplay. 17 'it'xd5 This greedy move forces Black to play vig­ orously. Other moves give him more choice: a) 17 'it'g2 'ifh5 18 'ifxd5 (after 1 8 f3 l:.ae8 1 9 i.e3 i.h3 20 'iff2 Black could try 20 . . . f5 or 20 . . :ii'g6 ! ?) 1 8 . . . l:.ad8 ( 1 8 . . . ':'fe8, as tried once by Topalov, is also interesting) 1 9 'it'c6 ( 1 9 'it'g2 l:fe8 gives Black a strong initiative) 1 9 . . . i.e6 20 'it'e4 and now: a l ) 20 . . . i.d5 2 1 'it'e2 i.f3 22 'iffl f5 (Black can consider 22 . . . :de8 ! ?) 23 lLld2 i.d5 24 f3

b l ) 1 9 . . . i.d3 is Black's traditional move. 20 l:.xe6 (20 'ifg2 ! ? is an idea, intending 20 . . . 'it'h5 2 1 'iff3) 20 . . .fxe6 2 1 'it'xd6 ':'xf2 22 �xf2 'ifxh2+ 23 �f3 'it'e2+ 24 �f4 h6 ! 25 'ifd8+ �h7 26 'ife7 �h8, with a forced repetition, is given by Nunn. b2) 1 9 . . . i.xg3 ! ? looks worth a try. After 20 hxg3 ':'xe3 2 1 fxe3 'it'xg3+ 22 'ifg2 'it'xe3+ Black has counterplay: 23 �h2 (or 23 �h l 'it'h6+ 24 'it'h2 'it'g6) 23 . . . 'ifh6+ 24 �g3 ':'e8 25 lLlf3 'it'g6+ 26 �h2 'it'h5+ 27 �g l i.e4 28 �f2 l:.e6 ! '

UNDERSTANDING THE MARSHAU A TTA CK

136

17 .l:.aeS 1S i.d2 (D) After 1 8 .l:.e3 ? ! 'it'hS ! Black has good pros­ pects; for example. 1 9 'it'f3 (worse are 19 'it'xd6? i.h3 -+ and 1 9 f3 ? 'it'g6 ! -+ attacking the b l ­ knight and threatening . . . i.f4) 1 9 . . . 'it'xf3 20 .l:.xf3 .l:.e l + 2 1 'it>g2 i.e4 22 lLld2 i.dS and Black is much better.

20 i.d7 21 'it'g2 'it'hS 22 f3 i.c6 (D)

•••

•••

w

B

IS .i.f4! This sharp move is Nunn's recommendation. Instead both 18 ... .l:.e6?! 1 9 lLla3 ! and 18 ... 'it'g4? ! 1 9 lLla3 (not 1 9 'it'xd6? .l:.xe l + 20 i.xe l 'it'd l 2 1 'it'eS i.d7 ! ) 1 9 . . .i.xa3 20 f3 ! look better for White. However. 1 8 . . . i.d3 is quite solid. After 1 9 lLla3 i.xa3 20 bxa3 i.c4 2 1 'it'f3 'it'd7 White' s winning chances look minimal despite the two extra pawns because White's pawns are bad and the opposite-coloured bishops help Black. Chandler-Nikolic. Leningrad 1 987. 19 ':xeS Alternatives: a) 1 9 gxf4 ? ! l:lxe l + (this is even better than 1 9 . . . 'it'g4+ 20 'it>h l i.e4+. when Black has at the very least a draw) 20 i.xe l 'it'g4+ 2 1 �h l i.e6 ! 22 'it'e4 'it'd l ! was given by Nunn. If 23 fS . then 23 . . . i.d7 ! with the idea 24 . . . .l:.e8 . b ) 1 9 'it'g2? .l:.xe l + 2 0 i.xe l i.c l ! 2 1 'it'xh3 i.xh3 22 lLld2 i.xb2 23 .l:.bl i.xc3 recovers both pawns and leaves Black clearly better. Sakhalkar-Barczay. corr. 1 9S9-60. 19 .l:.xeS 20 'it'c6 20 'it'g2? 'it'hS 2 1 f3 (2 1 gxf4 .i.h3 -+) 2 l . . ..i.d3 ! 22 gxf4 .l:.e6 ! 23 'it'f2 .i.e2 ! =+= gave Black a strong attack in Femmel-Sakai. corr. 200 1 . •••

•••

Now: a) 23 iLxf4? is bad due to 23 . . . .i.xf3 and Black is much better. b) 23 gxf4? l:.e6 ! 24 �f2 ':g6 2S 'it'h 1 'it'h4+ 26 �e3 and now: b l ) Nunn originally gave 26 . . . .l:.e6+ 27 �d3 'it'f2. After 28 �c2 i.xf3 29 'it'c 1 .l:.e2 30 a4 i.e4+ 3 1 �b3 g6. matters are still not so clear although Black can hardly be worse. b2) When this book was at proof stage. Nunn pointed out 26 . . . 'it'h3 ! . when White is in deep trouble. After 27 lLla3 .l:.g2. 28 'it'd 1 'it'e6+ 29 �d3 'it'fS+ 30 �e3 gS ! is winning for Black. while after 28 'it'n 'it'e6+ 29 �d3 'it'fS+ 30 'it>e3 .l:.xh2 ! White will not solve his king's problems without shedding too much material. c) After 23 g4 'ii'h4 24 'it'f2 Black probably has nothing better than forcing a draw with 24 . . . .i.xh2+ 2S 'iltxh2 :e l +.

Concl usions Although 1 2 .l:.e l and 1 2 g3 are often used as transpositional devices. there are some inde­ pendent lines. The old tries 1 2 d4 i.d6 1 3 .l:.e2. 12 i.xdS cxdS 1 3 d4 i.d6 14 .l:.e3 and IS i.xdS cxdS 1 6 'it'f3 have not seen many developments over the last couple of decades. The most topi­ cal try is I S 'ilte2. Black can reply with the stan­ dard I S . . . .i.g4. but currently I S . . . .i.d7 looks more solid. This is a line we shall surely see more of in the future.

Pa rt 3 : Anti - M a rsha l l

8 Anti - M a rsha l l : 8 a4 1 e4 eS 2 lLlf3 lLlc6 3 J.bS a6 4 J.a4 lLlf6 5 0-0 J.e7 6 .l:.e1 bS 7 J.b3 0-0 8 a4 (D)

Section 8 . 1 1 e4 eS 2 lLlf3 lLlc6 3 .i.bS a6 4 J.a4 lLlf6 5 0-0 J.e7 6 %let bS 7 .i.b3 0-0 8 a4 b4 (D)

w

Very often White does not want to delve into the Marshall Attack. Although White wins a pawn and has several ways to test both Black's resources and his memory, many players prefer not to be defending with the white pieces at such an early stage. For a long time, the most common way for White to dodge the Marshall Attack was with 8 a4. White makes a probing move on the queenside without committing his centre at all. This continuation was brought back into prominence by Garry Kasparov in his World Championship against Nigel Short. Black's most common answer is to develop with 8 . . . J.b7 . After 9 d3 d6, Kasparov honed 1 0 lLlbd2 (instead o f the traditional 1 0 lLlc3) into White's preferred weapon. Recently 8 a4 has taken a backseat to the modem 8 h3 of Chapter 9. One reason for this is the resurrection of 8 . . . b4 as a viable system. Section 8.1 : Section 8.2:

8 b4 8 J.b7 •••

•••

1 37 1 47

For a long time this move was considered rather second-best because it does not develop a piece and the c4-square falls into White's hands. Matters are not so simple, however, and many of the world's top players, including Anand, Aronian, and especially Grishchuk, now favour 8 . . . b4. Usually White continues with 9 d3 , when a positional battle ensues, but 9 d4 must also be considered. Thus we have : 138 A: 9 d4 B: 9 d3 1 39 9 a5 d6 1 0 d3 transposes to Line B2. After 9 c3 Black can try to play in Marshall style with 9 . . . d5, although I suspect this will fa­ vour White; or he may play more along the lines of the Anti-Marshall. A couple of examples: a) 9 . . . d6 1 0 h3 .l:.b8 1 1 d4 bxc3 ! 12 bxc3 exd4 1 3 cxd4 d5 14 e5 lLle4 with good play for Black, Zapata-I.Zaitsev, Moscow 1 989.

UNDERSTANDING THE MARSHALL A rrACK

138

b) 9 . . J:tb8 1 0 d4 d6 1 1 h3 bxc3 1 2 bxc3 exd4 1 3 lLlxd4 lLlxd4 14 cxd4 d5 15 e5 lLle4 1 6 lLld2 �b4 1 7 l::te 3 .i.xd2 ( 1 7 . . . lLlxf2 ! ? 1 8 �xf2 'ifh4+ 1 9 �e2 'ifxd4 20 %la2 �c5 is interest­ ing) 1 8 �xd2 was Lahno-Gustafsson, Pulver­ miihle 2006. Now the simplest is 1 8 . . . c5 =.

A) 9 d4 (D)

B

9 ... d6 9 . . . exd4 ? ! 10 e5 lLle8 1 1 �d5 favours White, while 9 . . . lLlxd4 1 0 lLlxd4 (White could also try 10 .i.xf7+ here) 1 O . . . exd4 1 1 e5 lLle8 1 2 'ifxd4 �b7 1 3 �f4 c5 14 'ife3 d5 1 5 exd6 �xd6 1 6 lLld2 'ifc7 1 7 �e5 �xe5 1 8 'ifxe5 'ifc6 1 9 f3 lLld6 20 lLlc4 lLlxc4 2 1 �xc4 ;j;; was Grishchuk­ Beliavsky, FIDE Knockout, Tripoli 2004. 10 dxe5 10 h3 exd4 1 1 lLlxd4 lLla5 ( l l . . .lLlxd4 1 2 'ifxd4 c5 1 3 'ifd3 �b7 1 4 c4? ! lLld7 1 5 f4 l:.e8 1 6 lLld2 .i.f8 1 7 .i.c2 g6 1 8 %lbl �g7 1 9 b3 'ifh4 with the initiative, Kobaliya-Khalifman, St Petersburg 2003) 1 2 c3 :b8 1 3 �c2 c5 1 4 cxb4 l:.xb4 1 5 lLlf3 �e6 1 6 lLlbd2 'ifb8 112- 112 Gashimov-Grishchuk, European Team Ch, Kher­ sonisos 2007 . 10 lLlxe5 It is understandable that Black wants to ex­ change knights, but 1 O . . . dxe5 leaves White without f3 for his queen and is worth consider­ ing. After 1 1 'ifxd8 l::t x d8 White has : a) 1 2 lLlbd2 and then: a l ) 12 . . . �d6 1 3 a5 h6 14 �c4 �f8 15 b3 �e 7 ? ! 16 �d5 lLlxd5 17 exd5 lLla7 1 8 lLlxe5 �f8 1 9 lLldc4 �xe5 20 l::txe5 lLlc6 2 1 �g5 ! •••

hxg5 22 dxc6 f6 23 l::tc 5 .i.e6 24 lLlb6 ! ± Karia­ kin-Aronian, FIDE Grand Prix, Nalchik 2009. a2) 1 2 . . . h6 1 3 a5 �d6 1 4 .i.c4 �f8 1 5 lLlb3 �e7 1 6 lLlfd2 �e6 1 7 �fl lLld7 1 8 lLlc4 lLlc5 1 9 lLlxc5 �xc5 20 c3 f6 2 1 �e3 112- 112 Nisi­ peanu-Inarkiev, French Team Ch, Evry 2008. b) 12 �g5 and now : b l ) 1 2 . . . �f8 1 3 lLlbd2 ( 1 3 �xf6 gxf6 { I do not see anything wrong with 1 3 . . . �xf6 } 1 4 �d5 �b7 1 5 lLlbd2 lLla5 1 6 �xb7 lLlxb7 1 7 b3 looked a little better for White in Svidler-Aro­ nian, Internet 2004) 1 3 . . . lLld7 14 �d5 .i.b7 1 5 lLlc4 f6 1 6 �e3 �c5 = has been played a couple of times with Black by Zviagintsev. b2) 1 2 . . . l::t b 8 1 3 lLlbd2 h6 14 �xf6 �xf6 1 5 �d5 lLla5 ! (Black contests the light squares on the queens ide ) 1 6 lLlb3 (or 1 6 %lad 1 l:.b6 1 7 lLlb3 112- 112 Svidler-Adams, Dortmund 2006) 1 6 . . . lLlxb3 1 7 cxb3 c6 1 8 �c4 ( 1 8 �xc6 �e6) 1 8 . . . a5 19 h3 �b7 20 l::t ad l �f8 = UlybinNegi, New Delhi 2009 . l l lLlxe5 Other moves do not look dangerous: a) 1 1 �f4 lLlg6 12 �g3 �b7 1 3 lLlbd2 lLlh5 1 4 lLlc4 lLlxg3 1 5 hxg3 a5 1 6 e5 .i.xf3 17 'i!kxf3 d5 1 8 lLle3 c6 1 9 'ifh5 �c5 20 .:tad 1 'ifb6 is much better for Black, Alekseev-Yakovenko, Russian Team Ch, Dagomys 2009. b) 1 1 lLlbd2 avoids the fixed centre, but it is rather harmless. l l . . .lLlxf3+ 12 lLlxf3 �b7 1 3 e5 lLld7 ! 1 4 e6 (after 1 4 �f4 both 1 4. . .�xf3 1 5 'ifxf3 lLlxe5 and Beliavsky's suggested 14 . . . lLlc5 1 5 exd6 �xd6 1 6 �xd6 'ifxd6 17 'ifxd6 cxd6 look close to equal) 14 . . . fxe6 1 5 �xe6+ �h8 16 �d5 c6 17 �a2 c5 1 8 �d5 �xd5 19 'ifxd5 lLlb6 20 'ife6 �f6 2 1 a5 lLld7 22 'ifd5 lLle5 23 lLld2 lLlg4 24 lLlf3 lLle5 25 lLld2 lLlg4 112- 112 Anand-Naiditsch, Dortmund 2004. 1l dxe5 12 'iff3 Other moves lead nowhere: a) 1 2 .i.g5 .i.c5 (after 1 2 . . . �b7 1 3 lLld2 lLld7 14 �xe7 'ifxe7 1 5 'ife2 a5 1 6 �c4 White had a small edge in lansa-I.Sokolov, Stary Smokovec 1 99 1 ) 1 3 'ii'f3 �g4 ! (this is justi­ fied because of the weakness of the f2-pawn) 14 �xf6 'ifxf6 1 5 'ifxf6 gxf6 =. b) 12 'ife2 �c5 (instead 12 . . .'ifd4 is proba­ bly too active, and after 1 3 lLld2 �b7 14 'iith l a5 1 5 f3 l::tfd8 1 6 lLlc4 .i.a6 1 7 �e3 .i.xc4 1 8 �xc4 'ifd7 1 9 l:.ed l White's bishop-pair gave him a comfortable edge in Palac-Aronian, •••

ANTI-MARSHALL:

Lausanne 2003) 1 3 i.g5 ( 1 3 i.e3 i.d4 is even, and 1 3 liJd2 i.b7 1 4 liJn ? would drop the e4pawn) 1 3 . . . i.g4 14 'it'c4 (again 14 i.xf6? 'it'xf6 1 5 'it'xg4 fails to 1 5 . . . 'it'xf2+) 1 4 . . . 'it'd6 was at least equal for Black in Tomer Planell-Moreno Carretero, corr. 2005 . We now return to 1 2 'ikf3 (D) :

8 a4

139

White invites complications. 15 �h8 16 liJg3 liJxe4 And Black obliges ! However, at this point there was little choice, as White was intending liJf5 with a strong initiative. 17 liJxe4 f5 (D) •••

w

B

12 i. b7 Also sufficient is 1 2 . . . i.e6 1 3 liJd2 i.c5 1 4 h3 liJd7 1 5 'it'e2 'it'e7 1 6 i.c4 a5 1 7 liJb3 i.b6 = L.Dominguez-Navara, Turin Olympiad 2006. This looks pretty safe. 13 liJd2 i.c5 (D) The immediate 1 3 . . 5Jth8 worked out for Black after 1 4 'ikf5 i.c5 ! 1 5 h3 (not 1 5 'ikxe5 ? i.xf2+) 1 5 . . . 'ikd4 in Akopian-Pashikian, Euro­ pean Ch, Budva 2009, but 14 liJc4 would be more challenging; e.g., 1 4 . . . i.xe4 ( 1 4 . . . liJxe4? 1 5 liJa5 ! ) 1 5 'ikg3 . White will follow with liJxe5 and keep some initiative. •••

18 liJxc5! i.xf3 19 gxf3 f4 20 �h2 'iWf5 21 i.d2 The position was very unclear in Ivanchuk­ Aronian, Russian Team Ch, Sochi 2006, al­ though White went on to win.

B) 9 d3 (D)

B

w

14 liJn 'ikc8 15 h3

This is the most common: White prefers a manoeuvring game. Often White will fix the a6- and b4-pawns with the advance a5 and bring a piece to c4. It is still possible to play for the d4 advance, with or without c3. Black will usually play . . . i.e6 followed by both .. J::tb 8 and

UNDERSTANDING THE MARSHALL A TTACK

140

. . . 'ii'c 8. These moves secure the queenside and give both major pieces more scope. 9 d6 White now has to make a decision. He can simply develop and allow Black to gain space with . . . lLla5 and . . . cS or he can try to clamp down on the queenside. We examine: 1 40 B l : 10 lLlbd2 1 42 B2: 10 a5

B11) 10 ... i.e6 (D)

•••

After 1 0 c3 l::tb 8 1 1 lLlbd2 lLlaS , 1 2 i.c2 cS 1 3 d4 'ii'c 7 gives Black a good 'Chigorin' , while 1 2 i.a2 cS ( 1 2 . . . i.g4 ! ?) transposes to the note to Black's 1 2th move in Line B 1 2, which looks a little better for White. The advance 1 0 h3 seems rather unnecessary here. Black can continue 1O . . . lLlaS 1 1 i.a2 cS 1 2 lLlbd2 and now both 12 . . . i.e6 and 12 . . . l:r.b8 look good.

B1) 10 lLlbd2 (D) White eschews the chance to fix Black's queenside with 1 0 as , preferring to develop.

w

This was tried by Grishchuk against Leko one round after his game with Anand in Line B 1 2. 1 1 i.xe6 The only real way to fight for the initiative. 1 1 lLlc4 can be met by several moves, such as 1 1 . . .h6, 1 1 . . .lLld7, 1 1 . . .a5 or 1 1 . . .l::tb 8. 11 fxe6 Now Black is intending to play . . . 'ii'e 8, . . . lLlhS and . . . 'ii'g 6 with counterp1ay. 12 lLln 1 2 c3 is less good because Black will be ready to use the b-file: 1 2 . . . l::tb 8 (or 1 2 . . . bxc3 1 3 bxc3 ':'b8, Hou Yifan-Bacrot, Wijk aan Zee 2008) 1 3 lLln (after 1 3 d4 bxc3 14 bxc3, even 14 . . . dS ! ? is possible, as in Papadopoulos-Gupta, Differdange 2008) 1 3 . . . bxc3 14 bxc3 lLla5 I S l:.a2 c S i s level. After 1 6 d4? ! exd4 1 7 cxd4 lLlb3 + Black already had the initiative in Car1sen-Bacrot, European Team Ch, Khersonisos 2007 . 12 'ii'd7 Black should avoid 1 2 . . . lLlhS ? 1 3 lLlxeS , but 1 2 . . . 'ife8 was suggested by Leko in New in Chess magazine. After 1 3 lLlg3 Black has : a) 1 3 . . . lLlhS and now 14 lLlxeS was Leko's in­ tention, but he points out the resource 1 4 . . . dxeS I S lLlxhS lLld4 ! 1 6 lLlg3 ( 1 6 i.e3 lLlf3+ 1 7 gxf3 'ii'xhS gives Black excellent play) 1 6 . . . 'ii'c 6! 1 7 c 3 bxc3 1 8 bxc3 'ii'x c3, which he gave a s un­ clear. Black certainly looks OK here. b) 1 3 .. .'ith8 14 c3 bxc3 I S bxc3 lLlaS (Black could consider lS .. Jlb8 ! ?) 16 d4 lLld7 17 'ifd3 •••

B

•••

Now Black can either continue his develop­ ment or he can try to punish White's omission. We have: 1 40 B l l : 10 i.e6 B12: 10... lLlaS 141 .•.

1 O . . . .:tb8 i s also possible. Then 1 1 as i.e6 transposes to the note to Black's 1 1 th move in Line B2, while 1 1 lLlc4 i.e6 12 a5 is line ' a2' of that note. 1 1 c3 was mentioned above via the move-order 10 c3 .:tb8 1 1 lLlbd2.

ANTI-MARSHALL:

cS was level in Chuprov-Nayer, Novokuznetsk 2008. c) 13 ... 'ilkg6 14 c3 bxc3 IS bxc3 .:tab8 16 .i.e3 ( 1 6 h3 lbhS) 16 ... lbg4 17 .i.d2 hS ! was suggested by Pavlovic, and this looks fine too. 13 lbg3 Stopping any . . . lbhS ideas. 13 lba5 (D) •••

8 a4

141

After 1 7 h3 Leko suggests 1 7 . . . lbe8 ! intend­ ing . . . lbc7 =. 17 bxc3?! Black should maintain the tension on the queenside and play 1 7 . . . lbg4 ! 1 8 eS dS 1 9 h3 c4 20 'il'e2 l:.xf3 ! intending . . . lbgxeS with com­ pensation. 18 bxc3 ..ti>h8 19 h3! ;!;; Leko-Grishchuk, World Ch, Mexico City 2007. Black has some difficulties arranging his pieces effectively. •••

w

812) 10 lba5 (D) •••

w

14 d4! White strikes immediately because 14 c3 bxc3 I S bxc3 l:.ab8 (or IS . . . cS 16 d4 exd4 1 7 cxd4 c4) 1 6 d4 lbb3 1 7 l:.a2 exd4 1 8 cxd4 lbxc 1 1 9 'ilkxc 1 l:.b4 gives Black good counter­ play according to Leko. 14 exd4 15 'il'xd4 c5 The desirable I s . . . lbc6 1 6 'ilkd3 lbg4 is well met by 1 7 eS ! because of the pin on the d-file; e.g., 17 . . . dS 1 8 h3 lbh6 19 .i.xh6 gxh6 20 'ilke3 ±. 16 'il'd3 lbc6 (D) ••.

w

17 c3

Black tries to take advantage of the fact that White did not bother playing 1 0 as . l l .i.a2 c5 Black has also tried the immediate 1 1 . . . ..te6 ! ? 1 2 ..txe6 fxe6 1 3 d4 ( 1 3 c 3 bxc3 1 4 bxc3 l:.b8 I S 'ilke2 lbhS 1 6 g 3 'ilke8 1 7 ..ta3 lbb3 1 8 :a2 lbxd2 1/2- 1/2 Adams-Leko, Linares 200S) 1 3 . . . exd4 1 4 lbxd4 'il'd7 I s lbn , and here: a) I s . . . lbc6 worked out alright for Black af­ ter 1 6 lbb3 'il'e8 1 7 lbg3 hS ! ? 1 8 h3 'ilkg6 1 9 'ilke2 h4 20 lbn lbhS 1/2 - 1/2 in Kariakin-Alekseev, FIDE World Cup, Khanty-Mansiisk 2007. b) I S . . . cS looks more natural. 1 6 lbf3 lbc6 1 7 lbg3 lbg4 1 8 eS is the same idea as we saw in the note to Black's 1 7th move in Line B 1 1 . Black has: b l ) 1 8 . . . .:ad8? ! just allows White to saddle Black with an inferior pawn-structure: 1 9 exd6 ..txd6 20 ..tgS ..txg3 2 1 hxg3 'ii'x d l 22 l:.axd l l:.xd l 23 l:.xd l l:.fS 24 l:.d6 ;!;; Svidler-Navara, European Team Ch, Khersonisos 2007.

142

UNDERSTANDING THE MARSHALL A TTACK

b2) 1 8 . . . d5 1 9 'ife2 ( 1 9 h3 ttlxf2 ! 20 �xf2 .Yl.h4 2 1 l:le4 'ife7 22 l:lf4 g5 ! 23 l:lxh4 gxh4 24 ttlh5 ttlxe5 is good for Black) 1 9 . . . c4 20 h3 l:lxf3 followed by 2 1 . . . ttlgxe5 gives Black good play. 12 c3 White creates some tension in the centre. Too dry is 1 2 ttlc4 and now either 1 2 . . . ttlc6 1 3 c3 l:lb8, as in Morozevich-Leko, Amber Rapid, Monte Carlo 2006, or simply 1 2 . . . ttlxc4 1 3 .Yl.xc4 when both 1 3 . . ..Yl.e6 and 1 3 . . ..Yl.b7 look fine for Black. 12 ttlc6 1 2 . . . bxc3 1 3 bxc3 .Yl.e6 1 4 .Yl.xe6 fxe6 1 5 'ifc2 intending 'ifa2 is an idea given by Carlsen, while after 1 2 . . . l:lb8 1 3 d4 Carlsen implies White has an edge in his notes to Anand-Grishchuk in New in Chess. 1 3 . . .'ifc7 (D) and now: ••.

Again, 1 3 ttlc4 is too slow. After 1 3 . . . :b8 1 4 h 3 both 1 4 . . . .Yl.e6 and 1 4 . . . h6 100k fine for Black. 13 bxc3 Carlsen suggests that Black keep his pawn on b4 and try 1 3 . . . cxd4 14 cxd4 .Yl.g4 1 5 h3 .Yl.h5 16 d5 ttla5 with some queenside play. 14 bxc3 exd4 15 cxd4 ttlb4 Even the world's best players disagree on how to handle these strategic positions. Carlsen also queried this move, preferring 1 5 . . . cxd4 1 6 ttlb3 with only a small edge for White. 16 .Yl.bl .Yl.g4 17 h3 .Yl.h5 18 g4 .Yl.g6 19 d5 (D) •••

B

w

The g6-bishop is in danger of remaining out of play. 19 ttld7?! This was queried by Anand and Marin, but not by Carlsen ! Anand gives 1 9 . . . l:.b8 20 ttlh4 (20 ttlc4 ttlbxd5 ! manages to take advantage of the g6-bishop's position) 20 . . . ttld7 (20 . . . .Yl.xe4 ! looks even stronger) 2 1 ttlxg6 hxg6 22 l:la3 .Yl.f6 with counterplay. 20 ttlc4 Now, there is harmony in the consensus of the world's great players. White is clearly better, because the g6-bishop is out of play and d6 is weak, Anand-Grishchuk, World Ch, Mexico City 2007. •••

a) 1 4 ttlfl exd4 (a more cautious option is to open the b-file with 1 4 . . . bxc3) 1 5 cxd4 c4 1 6 .Yl.g5 d5? ! ( 1 6 . . .b 3 i s better since 1 7 e 5 can be met by 1 7 . . . dxe5 1 8 dxe5 l:ld8) 1 7 exdS ttlxd5 1 8 .Yl.xe7 ttlxe7 1 9 l:lc 1 b3 20 .Yl.bl ;!;; Megaranto­ Z.Almasi, Calvia Olympiad 2004. White's cen­ tral control was more important than Black's queenside demonstration and White eventually brought down his famous opponent. b) 14 dxe5 dxe5 1 5 ttlc4 .l:[d8 1 6 'ife2 b3 1 7 ttlxa5 'ifxaS 1 8 .Yl.bl c4 ! ? (Black sacrifices a pawn for some activity rather than have a rather sterile disadvantage after 1 8 . . . 'ifc7 1 9 .Yl.d3) 1 9 'ifxc4 .Yl.c5 20 h 3 was Cheparinov-Aronian, So­ fia 2008. Black has a little activity, but with a healthy extra pawn, White must still be better here. 13 d4!

82) 10 as (D) This is the most common continuation. White fixes both Black's a- and b-pawns and prevents . . . ttlaS . Now Black's typical scheme of develop­ ment will be . . . .Yl.e6, . . . 'ifc8 and . . . l:lb8, which

ANTI-MARSHALL:

8 a4

143

initiative in Anand-I.Sokolov, Wijk aan Zee 1 996. 11 ttJbd2 (D)

B

B

will protect the b-pawn and also create the possi­ bility of playing . . . .l:.bS, when White may have to care for his advanced as-pawn. lO ii.e6 1 0 . . . l:tb8 1 1 ttJbd2 ii.e6 just transposes to line 'a' of the next note, but 1 O . . . ii.g4 (D) in­ tending . . . ttJd4 is interesting. Some examples : •.•

a) 1 1 c3 bxc3 1 2 bxc3 .t'tb8 1 3 ttJbd2 �c8 1 4 ii.c4 ttJa7 I S d4 ttJbS 1 6 'i:!Vc2 ttJd7 1 7 h 3 ii.hS was unclear in Kosteniuk-Moradiabadi, Paris 2008. b) 1 1 ttJbd2 is simple. 1 l . . .ttJd4 12 ii.c4 c6 1 3 h3 ttJxf3+ 14 ttJxf3 ii.hS IS g4 ii.g6 1 6 ttJh4 ;j;; Hracek-I.Sokolov, Pamu 1 996. c) 1 1 ii.e3 dS (after 1 l . . .ttJd4 12 ii.xd4 exd4 both 1 3 ttJbd2 and 1 3 h3 give White an edge) 1 2 ttJbd2 h6 ! 1 3 h3 d4 1 4 hxg4 dxe3 I S fxe3 ttJxg4 1 6 ii.dS ( 1 6 ttJc4 ;j;; is given by Anand because his bishop is stronger than Black's) 1 6 . . .i6d7 1 7 ttJc4 ii.f6 1 8 ttJfd2 hS 1 9 :tn (this leads to an interesting queen sacrifice) 1 9 . . . .t'tad8 20 .t'tfS g6 2 1 �xg4 ! hxg4 22 l:txf6 gave White the

1 1 �c8 This is consistent with Black's masterplan. 1 l . . ..l:.b8 generally amounts to the same thing if Black plays . . . 'i:!Vc8 next move. Another idea is simplifying with 1 l . . .ii.xb3, although this seems a bit cooperative. A couple of independent ideas: a) 1 1 . . . .t'tb8 and here: a 1 ) 1 2 ii.c4 ii.xc4 (transposing to Line B 2 1 with 1 2 . . . �c8 i s probably better) 1 3 dxc4 (af­ ter 1 3 ttJxc4 Black has scored well with both 1 3 . . . .t'te8 and 1 3 . . . ltbS) 1 3 . . . h6 1 4 ttJn �c8 I S ttJe3 �e6 1 6 ttJdS ii.d8 1 7 i6d3 ttJe7 1 8 ii.d2 ;j;; Z.Almasi-Peng Xiaomin, FIDE Knockout, Groningen 1 997. White has a grip on dS and the b4-pawn is weak. Also in this structure White never has to worry about any . . . ltbS ideas. a2) After 1 2 ttJc4 Black can of course play 1 2 . . . 'iVc8 (Line B22), but there are other moves: a2 1 ) 12 ... ii.g4 13 ii.e3 ttJd7 14 h3 ii.xf3 (not 14 . . . ii.hS ? I S g4 ii.g6 1 6 ii.a4 +- as the c6-knight has nowhere to go) I S �xf3 ii.gS 1 6 ii.a4 ii.xe3 1 7 'iVxe3 ttJd4 1 8 c 3 bxc3 1 9 bxc3 ttJbS 20 d4 was better for White in Kasparov­ Grishchuk, Moscow rapid 2002. a22) 1 2 . . . ttJd7 1 3 ii.e3 ii.f6 1 4 c3 bxc3 I S bxc3 ltb7 1 6 �c2 ttJa7 1 7 d4 'iVb8 1 8 dS ii.g4 1 9 ttJfd2 h6 20 h3 ii.hS 2 1 ii.a4 ttJbS 22 ttJa3 gave White a big advantage on the queens ide in de Firmian-Beliavsky, Politiken Cup, Copen­ hagen 2004. a23) 1 2 . . . h6 1 3 c3 (after 1 3 h3 Black should just play 1 3 . . . 'iVc8) 1 3 . . . bxc3 14 bxc3 ttJd7 I S ii.a4 ttJa7 1 6 ii.e3 ttJbS 1 7 �c2 fS was played .•.

144

UNDERSTANDING THE MARSHALL A TTACK

in Shirov-Z.Almasi, Calvia Olympiad 2004. Black has some play but White ' s pieces look better coordinated. b) 1 l . . .iLxb3 is playable, but not very dy­ namic. 12 tDxb3 (D) and now :

B

B21 : 12 iLc4 B22: 12 tDc4

b I ) 1 2 . . . d5 1 3 iLg5 ! (this forces Black to fix the centre) 1 3 . . . d4 1 4 ..txf6 ..txf6 1 5 tDfd2 g6 1 6 tDc4 :b8 and now Sutovsky played the sharp 1 7 f4 exf4 1 8 e5 ..th4 1 9 :e4 :b5 20 'it'f3 iLg5 2 1 .:tae 1 with compensation in Sutovsky­ Hracek, European Team Ch, Gothenburg 2005 . The e6 advance is coming and White has a strong initiative on the light squares. b2) 12 . . ..:te8 is solid. White then has: b2 1 ) 1 3 iLg5 tDd7 1 4 iLe3 iLf8 1 5 c3 bxc3 1 6 bxc3 l:.b8 1 7 d4 exd4 1 8 cxd4 tDb4 1 9 iLg5 'it'c8 ! 20 :e3 'it'b7 21 tDfd2 c5 22 :c 1 cxd4 23 tDxd4 d5 24 tDf5 dxe4 25 tDxe4 :xe4 26 :xe4 'it'xe4 27 'it'xd7 tDd3 was already a little better for Black in Stellwagen-Beliavsky, Wijk aan Zee 2006. b22) 1 3 h3 h6 ( 1 3 . . . 'it'd7 1 4 iLg5 h6 1 5 iLh4 tDh7 1 6 ..txe7 :xe7 1 7 d4 exd4 1 8 tDfxd4 tDf6 1 9 f3 tDxd4 20 'it'xd4 'it'b5 2 1 'it'd3 'it'g5 22 'it'd2 'it'xd2 23 tDxd2 l:tae8 24 tDb3 c5 25 l:ted 1 .:te6 26 .:td2 'it.'f8 27 l:lad 1 'it.'e7 28 tDc 1 gave White a lasting edge in Anand-Ivanchuk, Am­ ber Rapid, Monaco 200 1 ) 1 4 d4 exd4 1 5 tDfxd4 'it'd7 1 6 f3 tDxd4 1 7 tDxd4 c5 1 8 tDb3 'it'c6 (Black is comfortable) 1 9 iLe3 d5 20 e5 tDd7 2 1 iLf2 f6 22 exf6 iLxf6 23 l:lb1 'it'd6 1/2- 1/2 Anand­ Adams, Wijk aan Zee 2006. We return to 1 l . . .'it'c8 (D). White must now decide how to complete his development. Black is likely to respond 1 2 . . . :b8 in any case. We have:

1 44 1 45

1 2 h3 :b8 1 3 tDc4 transposes to Line B22 and 12 iLxe6 is harmless: 12 ... fxe6 13 tDc4 l:lb8 1 4 ..tg5 :b5 1 5 tDfd2 'it'e8 = Ivanchuk­ Kramnik, Amber Blindfold, Monte Carlo 2007.

82 1 ) 1 2 iLc4 White eyes the a6-pawn and places his bishop on a square where Black is less likely to ex­ change it because of the possibility of dxc4. This will allow him to manoeuvre with the d2knight and complete his development. Another new possibility is to develop with b3 and iLb2. 12 :b8 This is the typical building-block move, but other ideas are also possible: a) 1 2 . . . h6 1 3 h3 :e8 14 b3 iLf8 1 5 iLb2 'it'd7 16 'it'e2 iLxc4 ! ? 1 7 tDxc4 ( 1 7 dxc4 tDh5 1 8 tDfl tDf4 1 9 'it'd 1 l:te6 100ks OK for Black) 1 7 . . . g6 1 8 tDh2 iLg7 1 9 tDg4 tDh7 20 tDge3 f5 2 1 tDd5 ;!;; Akopian-Svidler, Wijk aan Zee 2004. White's knights are very well placed . b) 1 2 . . . :d8 1 3 ..txe6 (this move appears strange after playing iLc4 the move before) 1 3 . . . 'it'xe6 1 4 tDc4 :ab8 1 5 b3 :b5 gives Black counterplay. After 1 6 iLb2 tDd7 1 7 tDe3 l:le8 1 8 d4 exd4 1 9 tDd5 iLd8 20 tDxd4 tDxd4 2 1 'it'xd4 'it'e5 22 'ii'xe5 ? ! tDxe5 23 .:ted ! tDc6 24 f3 f5 ! 25 exf5 .:te2 Black had a clear advantage in Sutovsky-Ponomariov, Russian Team Ch, Sochi 2006. c) 1 2 . . . :e8 1 3 tDfl ( 1 3 b3 iLf8 1 4 iLb2 h6 1 5 h3 :b8 1 6 tDfl g6 1 7 tDe3 iLg7 1 8 tDd2 ...

ANTI-MARSHAll:

iLxc4 1 9 dxc4 �d4 20 �dS �xd5 2 1 exdS cS 22 dxc6 'ii'x c6 23 iLxd4 exd4 24 ':xe8+ :xe8 2S :a4 'ii'c s 26 �f3 was perhaps a tiny bit better for White in Kasimdzhanov-Grishchuk, FIDE Knockout rapid, Tripoli 2004) 1 3 . . . h6 1 4 �e3 .i.f8 (here i s another point to these rook moves: Black can fianchetto his king ' s bishop and may also consider . . . �e7 and . . . �g6 to shore up the kingside) I S b3 g6 1 6 iLb2 iLg7 1 7 'ii'd2 %:tb8 1 8 h 3 �d8 1 9 �h2 .i.xc4 20 bxc4 �e6 2 1 �hg4 �d7 22 �dS (White looks ac­ tive but his pieces will be pushed back) 22 . . . c6 23 �b6 ? ! �xb6 24 axb6 hS 2S �e3 %:txb6 26 cS (this was White ' s idea, but it backfires) 26 . . . �xcS 27 �c4 'fic7 28 �xb6 'fixb6 29 d4 �e6 30 %:ta4 as and with two pawns and a good position for the exchange, Black ' s play was much easier in Hracek-B acrot, B undesliga 2003/4. We now return to 1 2 .. .l:tb8 (D) :

8 a4

145

'fixf6 22 iLe3 iLg7 23 %:ta4 'ii'e6 24 .i.d2 :eb8 2S �e3 �e7 = Leko-Grishchuk, Dubai rapid 2002) 14 . . . :e8 I S �dS iLd8 1 6 iLd2 .i.xdS 1 7 iLxdS �xdS 1 8 exdS �e7 1 9 d4 �xd5 20 dxeS dxeS 2 1 �xeS iLf6 22 �c6 %:ta8 23 �xb4 �xb4 24 iLxb4 iLxb2 2S :b 1 1h- 1h Leko-Svid­ ler, Russia-ROW rapid, Moscow 2002. c) 1 3 . . . .i.xc4 14 dxc4 �d8 ! was suggested by Kasparov, although it has not seen a high­ level over-the-board test. Black's knight will be well placed on e6, where it defends the kingside and can jump to cS to attack the e4-pawn. After I S �e3 g6 ( l S . . . �e6 1 6 �fS .i.d8 with the idea . . . �cS also looks fine) 1 6 �dS �xdS 17 cxdS c6 ! ? 1 8 dxc6 'fixc6 1 9 iLh6 l:e8 20 'fid3 �e6 2 1 %:te2 %:tec8 Black had sufficient play in this Sicilian structure in Glaser-Lambert, corr. 1 994. 13 �d7 14 �n iL f6 15 �e3 g6 Black can also play l s . . . iLxc4 ! ? immedi­ ately. After 1 6 �xc4 ( 1 6 dxc4 is possible as well) 1 6 . . . �cS 1 7 iLb2 �e6 1 8 �e3 g6 1 9 c3 bxc3 20 iLxc3 iLg7 2 1 �dS 'fid7 Black had equalized in Yakoveoko-Zhang Zhong, Taiyuan 2006. 16 .i.b2 .i.xc4 Black makes this capture now that White has spent a few moves manoeuvring with his knight. 17 �xc4 After 1 7 dxc4 �cS 1 8 �dS iLg7 Black should be safe enough. The cS-square is a great square for Black's knight and Black can prepare . . . fS . 17 l:te8 18 d4? ! This i s mistimed for tactical reasons. After 1 8 h3 or 1 8 'fid2 the position is about level. 18 exd4 19 �xd4 �ce5 This is OK, but Marin points out the tactical opportunity 1 9 . . . iLxd4 ! ? 20 .i.xd4 dS . 20 �xe5 iLxe5 21 l:bl 'ii'b7 22 f3 �f6 23 'fid2 This was Ponomariov-Bacrot, Sofia 2006. Here Marin suggests 23 . . . �hS 24 g3 cS 2S �e2 iLxb2 26 :xb2 dS ! ? to try to take advantage of White's uncoordinated pieces. •••

•••

13 b3 1 3 �f1 is quite common. Black then has: a) 1 3 .. J:te8 ? ! 14 �e3 �d4?! IS �xd4 exd4 1 6 �dS �xdS 1 7 exdS .i.d7 1 8 .i.d2 .i.f6 1 9 %:txe8+ .i.xe8 20 'fie2 iLbS 2 1 %:te l .i.xc4 22 dxc4 h6 23 b3 cS was Kasparov-Short, World Ch match (game 1 ), London 1 993. I think White 's best here is simply 24 dxc6 'fixc6 2S 'fio 'fic7 26 iLxh6 (26 iLxb4 is also strong be­ cause 26 . . . %:txb4 27 :e8+ 'it>h7 28 'fifS+ mates) 26 . . . d3 27 cxd3 iLc3, which was given as good for Black by various commentators, but White has 28 'fig3 ! ±. b) 1 3 . . . h6 14 �e3 (or 14 h3 :e8 I S �e3 iLfS 1 6 iLd2 iLxc4 1 7 �xc4 'fie6 1 8 �h2?! l:tbS ? ! { 1 8 . . . �xe4 ! } 1 9 b 3 g 6 20 �g4 'it>h7 2 1 �xf6+

•••

822) 12 �c4 (D) White avoids exchanging pieces for the time being and puts pressure on eS to discourage the . . . dS advance. This is just as common as 1 2

146

UNDERSTANDING THE MARSHALL A TTACK

B

b3) 1 5 .i.d2 f5 1 6 lbe3 is considered the most principled try by Aronian. White wants to control the central light squares. He gives 1 6 . . . .i.xb3 1 7 cxb3 fxe4 I S dxe4 .i.dS as being a little better for White. 13 h6 14 .i.e3 (D) •••

B

.i.c4, and here too Black has been holding his own. 12 ... ':b8 As always, this is a useful move for Black. It protects the b4-pawn and may even come out to b5 to put pressure on White's as-pawn. 13 h3 1 3 .i.g5 has also been played at a very high level, although frankly I do not really under­ stand this move. Black has: a) 1 3 . . . h6 is natural and good: a l ) 14 .i.e3 just gives Black . . . h6 for free. 1 4 ... ':dS 15 h3 .i.fS 16 'ii'd2 (after 16 d4 exd4 1 7 lbxd4 lbxd4 I S .i.xd4 lbxe4 ! ? 1 9 :xe4 .i.xc4 { 1 9 . . . d5 ! ? 20 :xe6 'fi'xe6 2 1 lbd2 c5 22 .i.e3 c4 is unclear } 20 .i.xc4 d5 2 1 .i.xd5 :xd5 22 'ii'e2 _f5 the position is equal) 16 . . . lbe7 1 7 'ith2 lbg6 112- 112 Esserman-Vigorito, Somerville 2009. a2) 1 4 .i.h4 :b5 was suggested by Pono­ mariov. One point is that after 1 5 d4? ! exd4 1 6 lbxd4 lbxd4 1 7 'ii'xd4 d5 ! Black's pieces all spring to life. b) 1 3 . . . 'ithS ! ? intends . . . lbgS and . . . f5 with counterplay. 14 h3 (after 14 d4 Black can play 14 . . . exd4 15 lbxd4 lbxd4 1 6 'ii'xd4 lbgS 1 7 .i.f4 f5 ! , as i n Leko-Anand, Amber Rapid, Mo­ naco 2006, or Aronian's suggestion 14 . . . .i.g4) 14 . . . lbgS ! and here: b I ) 15 c3 ? ! allowed Black to execute his plan with 15 . . . bxc3 16 bxc3 f5 in Leko-Aro­ nian, Morelia/Linares 2006. Black already has the initiative. b2) 1 5 .i.xe7 lbgxe7 1 6 lbg5 lbd4 1 7 .i.a2 b3 I S .i.xb3 lbxb3 19 cxb3 lbc6 gives Black good compensation for the pawn according to Aronian.

14 :d8 This is a healthy move. With Black's rook lined up against White's queen, . . . d5 is a dis­ tinct possibility and Black can also manoeuvre with . . . .i.f8 and maybe . . . lbe7 . Harikrishna's 1 4 . . . lbh7 ! ? is also very interesting. Against an indifferent response, Black will play . . .f5 as in Leko-Aronian above, while . . . lbg5 is also a possibility. 1 5 d4 exd4 1 6 lbxd4 lbxd4 1 7 .i.xd4 :b5 ! I S f4 .i.f6 1 9 f5 ? ! ( 1 9 'fi'd3 is safer) 1 9 . . . .i.xc4 20 .i.xc4 .i.xd4+ 2 1 'fi'xd4 :e5 22 'ii'd 3 lbf6 23 .i.xa6 'ii'a S 24 .i.b5 ':bS 25 .i.c4 ':xa5 26 1lxa5 'ii'xa5 27 'ii'b 3 :fS 2S 'ith 1 'ii'c 5 29 c3 bxc3 30 'ii'x c3 :eS 3 1 .i.d3 'ii'x c3 32 bxc3 ':'e5 =+= Hou Yifan-Harikrishna, Paks 2007 . Of course White can improve somewhere here, but Black still looks pretty comfortable. 15 'fi'e2 (D) The exotic 1 5 'ii'b l lbd7 1 6 'ii'a2 does not trouble Black after 1 6 . . . :b5 1 7 c3 d5 ! I S exd5 .i.xd5 1 9 lbd6 ( 1 9 lbfd2 lbf6 is also better for Black because his pieces coordinate so well) 19 . . . .i.xb3 20 lbxcs .i.xa2 2 1 lbxe7+ lbxe7 22 :xa2 bxc3 23 bxc3 lbc6 gave Black an edge in Stellwagen-Onishchuk, Wijk aan Zee 2005 be­ cause the a5-pawn is weak. 15 .i.f8 1 5 . . . 'ii'b 7, as in Caruana-Werle, Wijk aan Zee 2009, also looks quite reasonable. 16 lbfd2 .••

•••

ANTI-MARSHALL:

B

This move protects the e4-pawn and the c4knight so that both .i.a4 and the d4 advance be­ come possibilities. Grishchuk has twice faced 1 6 lLlcd2, which seems rather passive. After 1 6 . . . .i.xb3 1 7 lLlxb3 .l:r.e8 ( 1 7 . . . d5 ! ?) 1 8 lLlfd2 d5 19 exd5 lLlxd5 20 �f3 l:lb5 2 1 lLlc4 �d7 there is: a) 22 l:tad l f5 23 .i.c 1 g6 24 l:.e2 h7 25 .l:r.de l .i.g7 26 'ifig3 �f7 27 �f3 �d7 28 �g3 �f7 29 �f3 was agreed drawn here in the game Smirin-Grishchuk, Russia-ROW rapid, Moscow 2002 although there is still plenty of play left. b) 22 .i.d2 g6 23 .l:r.e2 .i.g7 24 l:lae l f5 25 �g3 g5 26 f3 �f7 27 �f2 lLlf4, 28 .i.xf4 gxf4 was unclear in A.Kovacevic-Grishchuk, Euro­ pean Clubs Cup, Kallithea 2002. 16 lLle7 1 6 . . . d5 ? ! runs into 1 7 .i.a4 ! dxc4 1 8 .i.xc6 cxd3 19 cxd3 .i.d7 20 .l:r.ec I with an edge for White (Gershon). With the text-move, the knight heads to the secure g6-square in anticipation of White's d4 advance. I like this idea, but another possibility that has been played by a couple of rather de­ cent players is 1 6 . . . �b7 ! ? This is a sensible move that protects the c6-knight and connects the rooks. White has tried: a) 1 7 �f3 h7 1 8 lLlfl lLle7 1 9 lLlg3 lLlg6 20 lLlh5 (20 lLlf5 d5 ! ) 1/2- 1/2 Svidler-Aronian, Morelia/Linares 2006. b) 1 7 .i.a4?! looks like a shot in the dark. 1 7 . . . lLle7 1 8 'ili'f3 lLlg6 ! 1 9 .ltb3 ( 1 9 .i.xh6 d5 ! ) 1 9 . . . .i.e7 2 0 'ii'g 3 h8 2 l lLlf3 lLlh5 2 2 'ili'h2 c5 23 lLlcd2 'ific8 24 .i.xe6 'ili'xe6 25 lLlfl f5 gave Black the initiative in Lutz-Svidler, Bundesliga 2005/6. Clearly Svidler was impressed with •••

8 a4

147

Aronian ' s idea, as this game was played a cou­ ple of months after Svidler-Aronian. 17 d4 lLlg6 This is solid. Black chose to give up the cen­ tre with 1 7 . . . exd4 1 8 .i.xd4 in Volokitin-Gus­ tafsson, Bundesliga 2005/6, and now I prefer the solid 1 8 . . . lLld7 to the game's 1 8 . . . lLlh7. 18 d5 .ltd7 19 .i.a4 White cannot play 1 9 c3 because 1 9 . . . bxc3 20 bxc3 lLlxe4 ! exploits the loose bishop on b3 . Black's position is solid enough after the tempo­ rizing 1 9 .l:r.ec 1 lte8 20 c3 lLlf4 2 1 �d 1 �d8 ! ? (the queen i s not needed at c 8 any more so Black places it on its home square, from where it can access either side of the board) 22 .lta4 �e7 23 b3 g6 24 lLlb2 ':ec8 25 .i.xd7 'ili'xd7 26 c4 �e8 27 �fl 1/2- 1/2 Kariakin-Aronian, Wijk aan Zee 2006. 19 .i.b5! Now if White wants to exchange bishops he wili have to prepare it because after 20 .i.xb5 ? axb5 he loses the knight on c4 . 20 b3 .i.e7 21 .l:r.ecl c6! This is a typical method for Black in the Ruy Lopez, especially when White cannot play c4. Black fights in the centre and gains room for his pieces. 22 .i.xb5 ':xb5 23 lLlb6 'ili'b7 24 dxc6 'ili'xc6 25 �c4 1/2- 1/2 Adams-Anand, FIDE World Ch, San Luis 2005 . Already Black could consider play­ ing for an advantage. After 25 . . . �b7 with the idea . . . d5 Black has good chances. •••

Section 8 . 2 1 e4 e5 2 lLlf3 lLlc6 3 .i.b5 a6 4 i.a4 lLlf6 5 0-0 .i.e7 6 ':el b5 7 .i.b3 0-0 8 a4 .i.b7 (D) This is the traditional main line. Black devel­ ops a piece and avoids committing his queen­ side pawn-structure for the time being. 9 d3 Exchanging immediately with 9 axb5 axb5 1 0 .l:r.xa8 is harmless and both 1 0 . . . .i.xa8 and 1 O . . . 'iIi'xa8 are fine for Black. 9 c3 d5 10 exd5 lLlxd5 is covered via the move-order 8 c3 d5 9 exd5 lLlxd5 1 0 a4 .i.b7 in Line A of Section 1 0.5. The only other move to consider is 9 lLlc3, but this is well met by 9 . . . lLld4 ! :

148

UNDERSTANDING THE MARSHALL A TTACK

A: 10 c3 B: 10 lLlc3 C: 10 lLlbd2

w

a) 1 0 lLlxe5 ? ! lLlxb3 1 1 cxb3 b4 wins back the pawn with advantage. b) 10 lLlxd4 exd4 1 1 lLld5 lLlxd5 1 2 exd5 d3 ! disrupts White's development. c) 10 �a2 b4 1 1 lLld5 lLlxf3+ (the immedi­ ate l l . . .lLlxd5 is also equal) 1 2 'tlVxf3 lLlxd5 1 3 exd5 �d6 i s level. d) 10 axb5 lLlxb3 1 1 cxb3 axb5 12 Iha8 'tlVxa8 1 3 d3 b4 (or 1 3 . . . d6 14 lLlxb5 , when both 14 .. .l1b8 and 14 . . .'tlVa5 give Black compensa­ tion for the pawn) 14 lLle2 d5 and Black has the initiative. 9 d6 (D) Several other moves are playable here, with the principal alternative being 9 . . .l:1e8, but we shall stick with the main line.

1 48 1 50 1 52

White can also play the unnatural 1 0 �d2, which is directed against the traditional . . . lLla5 plan. However, after 1O . . . b4 (there are other moves too, but this is consistent with the lines examined here) White does not have a square for his queen's knight. 1 1 c3 (after 1 1 a5 , the typical I l . . J�b8 is a good answer, while 1 1 �g5 can be met by 1 1 . . . lLld7, 1 1 . . . lLla5 or even 1 l . . .'it>h8 ! ? intending . . . lLlg8 and . . . f5) l l . . .bxc3 1 2 bxc3 ( 1 2 lLlxc3 lLla5 1 3 �a2 c5 = ) and here 1 2 . . . lLla5 , intending . . . c5, is very solid, while Black can also try the speculative 1 2 . . . d5 with the idea 1 3 exd5 lLlxd5 14 lLlxe5 lLlxe5 1 5 �xe5 �f6 followed by . . . 'tlVd6, when Black's devel­ opment lead and pressure on c3 give him some compensation for the pawn.

A} 10 c3 (D)

...

B

w

White continues in ' slow Ruy Lopez' fash­ ion. 10 lLla5 Black can also play 1 O . . . lLlb8 or 1 O . . . lLld7, but there is no reason not to continue in 'Chigorin' style and grab space. 1 l �a2 This gives the game a little original flavour. Instead 1 1 �c2 c5 1 2 lLlbd2 is a Ruy Lopez with d3 . Black can play 1 2 . . . 'tlVc7, 1 2 . . . lLlc6 or 1 2 . . . �e8. 1l c5 12 lLla3 •.•

Now White has three main ways to play. The first looks like a normal Ruy Lopez with an early d3, which should not be dangerous to Black. The second is the traditional main line, while the third is the most flexible. We exam­ ine:

•••

ANTI-MARSHALL:

White takes aim at the b5-pawn. 1 2 lLlbd2 looks natural, but with the bishop on a2 the e4pawn is a bit tender, and 1 2 . . . 'ii'c 7 1 3 lLlfl c4 ! undermines White's centre. Black has a devel­ opment lead and plenty of space and already stands well. Blowing up the position with a quick . . . d5 is even a possibility. 12 'ii'd7 13 axb5 axb5 14 .i.d2 (D) •••

B

8 a4

The queenside pawns have disappeared and the structure is symmetrical. Black is very close to equality, but White can still press a little bit. 19 ..tc4! This trick allows White to keep some initia­ tive. 19 'ii'xb4 After 1 9 . . . 'ii'd7 20 'ii'b 3 White maintains an edge because of the pressure on f7, as he also does in the case of 1 9 . . . .l:.xal 20 'ii'x al 'ii'd7 2 1 'ii'a3 .l:.a8 2 2 'ii'b 3 ;1;. 20 l:tbl 'ii'a5! ? 2 0. . . 'ii'c 5 2 1 .l:.xb7 .l:.a7 2 2 .l:.b5 (Black grad­ ually equalized after 22 ':'xa7 'ii'xa7 23 lLlh4 'it?h7 24 lLlf5 g6 25 lLle3 'it?g7 = in Leko-Adams, Wijk aan Zee 1 996) 22 . . . 'ii'c7 and now instead of 23 'ii'b 3 lLld7 (Kholmov-Geller, Moscow 1 995), 23 lLlh4 gives White a nagging little edge. 21 .l:.xb7 .l:.a7 22 'ii'b l Another possibility is 22 ':b5 'iia3 23 lLlh4 ;1;. 22 ..td8 Perhaps 22 . . . lLld7 is a better try. 23 .l:.b5 'ii'aJ 24 g3 lLld7 25 '.tg2 g6 26 .l:.e2 'ii'a l 27 'ii'x al White is unambitious. After 27 'ii'b4 or 27 'ii'c 2 Black should still equalize, but he will have to take some care because White's bishop is more active. 27 ':'xal The game is equal, Hiibner-Kamsky, peA Qualifier, Groningen 1 993. •••

.••

White intends to play b4 to drive back the black knight, after which the b5-pawn will be loose. Black can ignore the threat or prevent it. 1 49 A I : 14 h6 A2: 14 c4 1 49 •••

•••

AI)

•••

14 h6 Black is not concerned with White's idea and simply makes a useful move. 15 b4 cxb4 16 cxb4 lLlc6 17 lLlxb5 lLlxb4 This resource regains the pawn. 18 .i.xb4 'ii'xb5 (D) •••

w

149

A2) 14 c4 (D) •••

w

UNDERSTANDING THE MARSHALL A TTACK

150

This is a more active continuation. Black dis­ courages b4 and sets out to undermine White's e4-pawn. 15 .i.g5 White intends dxc4, when tDxe5 will be a possibility because of the latent pressure on the d-file and the h4-d8 diagonal. 15 JWc6! (D) This has been seen a few times in practice. 1 5 . . . l:tfc8?! allowed White to carry out his threat with 1 6 dxc4 tDxc4 1 7 .i.xc4 bxc4 1 8 tDxe5 in Balashov-Solozhenkin, Katowice 1 99 1 . ••

w

1 8 dxe5 dxe5 1 9 tDxe5 ? fails to 1 9 . . . .i.xa3 , winning a piece. 18 tDc6 19 d5 tDb8 20 l:tadl tDa6 The knight is heading to c5. Already Black's position looks more promising. Black could avoid the following trick with 20 . . . tDbd7 , al­ though there is really nothing to fear from White's tactics. Note that 2 1 tDxb5 would be impossible because of the loose bishop on a2. 21 tDxe5 dxe5 22 d6 The threat of .i.xf6 and d7 wins back the piece. 22 ... tDc5 23 dxe7 'iVxe7 + Palac-Pujos, Cannes 2002. Black's space advantage and active pieces are more important than White's bishop-pair. •.•

B) 10 tDc3 (D)

B

Black has not experienced any real problems from this position. 16 'iVe2 A couple of other moves have also been tried, but White has not achieved anything: a) 16 tDc2 h6 17 tDb4 'ii'c5 18 .i.h4 l:ta7 ( 1 8 . . . l:.fe8 looks more natural) 1 9 dxc4 bxc4 20 .i.xf6 .i.xf6 2 1 'ili'e2 and now 2 l . . .':'c8 was equal in Smirin-Gyimesi, Croatian Team Ch, S ibenik 2005, while 2 l . . .l:tfa8 intending . . . .i.c8-e6 even gives Black the chance to play for an edge. b) 1 6 d4 is natural, but with White's pieces jumbled on the queens ide it is difficult for him to fight for the initiative. The a3-knight is es­ pecially awkward. After 16 . . . l:tfd8 17 tDh4 g6 1 8 'iVf3 rt;g7 1 9 l:tad l 'iVe8 20 .i.bl tDb3 Black was fine in Milos-Nguyen Anh Dung, Istanbul Olympiad 2000. 16 l:.fc8 17 d4 Again, this is not terribly dangerous. Usually White gets to play this move very early in the Ruy Lopez. Here Black has already made in­ roads on the queenside. 17 'iVe8! ? 18 'iVc2 .••

•.•

This is the traditional main line. 10 ... tDa5 Following the standard recipe. Another idea is to force things with 1 O . . . b4, because 1 1 tDd5 ( 1 1 tDe2 tDa5 1 2 .i.a2 will transpose to the main line) is not too dangerous. After l l . . .tDxd5 1 2 .i.xd5 l:lb8 1 3 �d2 'iVd7 1 4 a5 tDd8 1 5 .i.c4 c5 1 6 c3 bxc3 17 bxc3 tDe6 1 8 'iVc2 .i.f6 19 'iVa2 tDc7 20 l:.abl tDb5 2 1 h3 .i.c6 Black had equal­ ized in Kotronias-Radjabov, European Team Ch, Gothenburg 2005 . l l .i.a2 The greedy 1 1 axb5 tDxb3 1 2 cxb3 axb5 1 3 l:txa8 'iVxa8 1 4 tDxb5 gives Black compensa­ tion after 14 . . . l:tb8 or 14 . . . 'iVa5 . 1 l b4 12 tDe2 c5 (D) •••

ANTI-MARSHAll:

8 a4

151

14 . . . tLJxb3 I S cxb3 as 1 6 i.d2 tLJd7 1 7 i.c3 tLJbS IS tLJd2 tLJc6 19 tLJc4 i.cs 20 :fl i.e6 2 1 f4 i n Anand-Bryson, Thessaloniki Olympiad 1 9S4. Another thematic idea is 1 3 . . . :bS (D).

w

w

13 tLJg3 The knight was not well placed on e2 be­ cause it disturbed the coordination of White's pieces and interfered with White's protection of the e4-pawn, so this is the most logical continu­ ation. Others : a) 1 3 i.d2 :bS 1 4 tLJg3 i.cs I S h3 tLJeS ! ? ( I S . . . i.e6 i s also possible) 1 6 tLJh2 i.gS 1 7 tLJf3 i.f6 I S tLJh2 g6 1 9 tLJg4 i.gS 20 tLJe3 �hS was Caruana-Adams, Gibraltar 2007 . Black will soon play . . . fS with counterplay. b) 1 3 c3 used to be considered somewhat critical. 1 3 . . . bxc3 14 bxc3 and now 1 4 . . . c4 ! is a typical move undermining the e-pawn. After I S tLJg3 Black has: b l ) I S . . . cxd3 1 6 'fixd3 i.cs (Black intends to regroup with . . . i.e6 and . . . tLJb7-cS) 1 7 i.gS ! tLJb7 I S tLJd2 tLJcs 1 9 'fic2 was Ljubojevic-Kar­ pov, London 1 9S2, and now 1 9 . . . i.e6 100ks the most consistent. b2) I S . . . g6 was suggested by Nunn. This re­ mains untried, but looks reasonable. b3) IS . . . tLJd7 was recommended by Nunn and Harding. Then 1 6 i.a3 ( 1 6 tLJfS tLJcs gives Black good play) 1 6 . . . g6 1 7 d4 'fic7 I S l:[bl (the odd-looking I S l:[c l was met with I S . . . :feS 1 9 i.b4 l:[acS 20 tLJd2 dS ! ?, blowing up the position, in Kupreichik-Am.Rodriguez, Minsk 1 9S2) I S . . . :abS looks fairly level. 13 i.c8 This is a typical manoeuvre. The bishop heads to e6 to neutralize White's a2-bishop. The pawn sacrifice 1 3 . . . b3 looks too ambitious in view of 14 i.xb3 ( 1 4 cxb3 tLJc6 leaves the a2-bishop buried and justifies Black's play), when White's knight will head to c4. Black did not really have anything for the pawn after •••

This is a bit of a waiting move but it antici­ pates White's natural plan to play in the centre with c3 by occupying the b-file. White has : a) 14 tLJd2 i.cs ( 1 4 . . . tLJeS I S tLJc4 tLJxc4 1 6 i.xc4 tLJc7 1 7 'ifg4 gives White some initia­ tive) I S tLJc4 ( 1 S h3 brings us back to the main line) I S . . . tLJxc4 (another idea is IS . . . i.g4 ! ? 1 6 f3 i.e6) 1 6 i.xc4 tLJeS (now Black cannot play 16 . . . i.e6 because of 1 7 i.xa6) 1 7 f4 ( 1 7 tLJfS i.gS =) 1 7 . . . exf4 1 S i.xf4 i.f6 1 9 :bl tLJc7 20 tLJhS i.d4+ 2 1 i.e3 i.eS 22 tLJf4 tLJe6 23 tLJxe6 i.xe6 24 i.xe6 fxe6 was pleasant for Black in A.Sokolov-Balashov, Moscow TV rapid 1 9S7. b) 1 4 tLJfS i.cs and here: b l ) I S tLJxe7+ 'ifxe7 gives Black no prob­ lems because . . . i.e6 will exchange off one of White's bishops. 1 6 tLJd2 i.e6 1 7 tLJc4 tLJxc4 I S i.xc4 as = Short-Kholmov, Erevan 1 9S4. b2) IS tLJ3h4 tLJeS 16 tLJxe7+ 'fixe7 17 g3 i.e6 = Naiditsch-Ehlvest, Aeroflot Open, Mos­ cow 2007 is similar. b3) I S tLJe3 i.e6 1 6 i.d2 tLJeS 1 7 i.xe6 fxe6 I S c3 tLJc6 (Black could consider I S . . . bxc3 1 9 bxc3 tLJc6) 1 9 tLJc4 tLJc7 20 i.e3 i.f6 was pretty solid for Black in Anand-Short, Amster­ dam 1 992. We now return to 13 . . . i.cS (D): 14 h3 White maintains the tension. 14 tLJd2 i.g4 I S f3 i.e6 1 6 tLJc4 tLJxc4 (or 1 6 . . . tLJc6) 1 7 i.xc4 'fid7 i s fine for Black, while 1 4 c 3 l:[bS

UNDERSTANDING THE MARSHALL A TTACK

152

w

w

1 5 d4 gives Black a choice between 1 5 . . . bxc3 16 bxc3 'ikc7 and 1 5 . . . b3 1 6 �b 1 'ikc7 in both cases with an unclear position. 14 J:tb8 1 4 . . . .i.e6 1 5 �xe6 fxe6 1 6 c3 bxc3 1 7 bxc3 J:.b8 1 8 �e3 'ikd7 19 'ilVc2 is supposed to be a little better for White based on the game Chan­ dler-Tal, London 1 984, but the position is prob­ ably not so bad for Black; it is very similar to some in Chapter 9. One possibility here is 1 9 . . . c4 ! ? Is lLld2 White can also play 1 5 �d2, which trans­ poses to note 'a' to White's 1 3th move. IS �e6 This is the most common, but Black has also tried 1 5 . . . lLle8 1 6 lLlc4 lLlxc4 1 7 iLxc4 and now both 1 7 . . . .i.g5 and 1 7 . . . lLlc7 have been played, while 17 . . . g6, intending . . . lLlg7, is also possible. 16 lLlc4 lLlc6 (D) 1 6 . . . lLlxc4 17 �xc4 iLxc4 1 8 dxc4 has also been played with some success (drawing suc­ cess, that is), but I think that Black should avoid this structure unless he gets something (like the bishop-pair) in return. 17 f4 This does not lead to much, but it is not easy for White to develop any kind of initiative. 17 exf4 18 iLxf4 lLle8 1 8 . . . :c8 1 9 lLlf5 ( 1 9 lLle3 ! ?) 1 9 . . . �xf5 20 exf5 d5 2 1 lLle3 h6 22 �g3 :e8 = Yudasin­ Kuporosov, USSR 1 985. Black should be care­ ful not to get overly ambitious: 1 8 . . . lIb7 1 9 lLle3 iLxa2 20 lha2 b 3 2 1 cxb3 lLld4? (2 1 . . .g6) 22 lLlgf5 lLlxb3 23 lLlc4 ± Zapata-Gutierrez Carmona, Aguascalientes 2008 is one such cau­ tionary tale.

19 lLlfS �xfS This move was not mentioned in Nunn & Harding's book, but soon after it came out Nunn played it himself! Instead 1 9 . . . �f6 still looks satisfactory; for example, 20 :f1 (White gets nowhere with 20 lLlcxd6 �xb2 or 20 lLlfxd6? lLlxd6 2 1 lLlxd6 { 2 1 �xd6 �xc4 =+= } 2 l . . .iLxb2 =+=) 20 . . . �xf5 2 1 exf5 d5 ! ? 22 iLxb8 'ilVxb8 ! gives Black good compensation for the exchange. 20 exfS �f6 (D) 20 . . . lLld4 is also possible.

••

•.•

•••

21 'iii>h l 'ilVd7 22 �h2 lLld4 23 lLle3 b3! ? 24 �xb3 lLlxb3 25 cxb3 iLxb2 26 :bl �c3 27 :n iLd4 28 lLlc2 iLeS = Kindermann-Nunn, Bundesliga 1 988/9.

C) 10 lLlbd2 (D) It is this move that brought 8 a4 back into fashion when Kasparov adopted it in his 1 993 World Championship match with Short.

ANTI-MARSHALL:

8 a4

1 53

b4 1 5 �g5 ':'f8 1 6 �g4 �xg4 1 7 'ili'xg4 i.cS I S 'i!i'h4 h5 1 9 'ili'g3
W

w

1 2 �n After 1 2 c3 'ili'c7 1 3 �fl Black has the famil­ iar idea 1 3 . . . c4 ! , undermining White's centre. 12 ... b4 This is Black's usual choice. If White is go­ ing to play in the centre with an eventual c3 and d4, this ensures that Black can open some lines on the queenside. A couple of illustrations of the dangers Black can face in other lines: a) 1 2 . . .l:teS 1 3 �e3 g6 ( l 3 . . . h6 1 4 i.d2 { White threatens 15 .i.xa5 'ili'xa5 16 axb5 with the idea 1 6 . . . axb5 ?? 1 7 i.xf7 + } 14 . . . c4 1 5 i.c3 'ili'b6 1 6 �d2 gave White a comfortable plus in Kasparov-Leko, Linares 200 1 because his pieces are much better coordinated) 14 i.d2

13 �e3 Or: a) 1 3 c3 bxc3 (illogical is 1 3 . . . l::tb S? ! 1 4 cxb4 cxb4 1 5 i.d2 � ) 1 4 bxc3 c4 ! 1 5 �g3 transposes to note 'b' to White's 1 3th move in Line B . Here White's knight has gone d2-fl -g3 instead of c3-e2-g3. b) 13 �g3 transposes to Line B . c) The other alternative, 1 3 i.d2, does not look too dangerous. After 1 3 .. JlbS 14 �e3 i.cs 1 5 �c4 �c6 1 6 h3 �eS 1 7 �e3 �c7 I S �d5 i.e6 1 9 �xc7 ( 1 9 c 3 i.f6 was also fine for Black in Lutz-Beliavsky, Bundesliga 200213) 19 .. :iWxc7 20 i.c4 'ii'c s 2 1 b3 h6 22 .i.c 1 f5 23 'ili'e2 f4 Black was doing well in Hracek-Beliav­ sky, Bundesliga 2002/3 . 13 ... i.c8 14 �d2 1 4 �c4 transposes to Section S. l , with both sides losing two tempi ! Here Black has lost time playing . . . i.b7-cS and White has lost time playing �d2-fl -e3-c4 instead of �b1 -d2-c4. White may also try 14 c3. After 14 .. JlbS ( 1 4 . . . bxc3 15 bxc3 ':'bS is also possible) 1 5

UNDERSTANDING THE MARSHAll A TTACK

154

cxb4 ( I S ltJd2 .lte6 1 6 ltJdS ltJxdS 1 7 exdS .ltd7 1 8 ltJc4 ltJxc4 1 9 .ltxc4 as 20 cxb4 lhb4 2 1 b3 .ltgS was also very comfortable for Black in Geller-Xie Jun, Vienna 1 993) I S . . . l:txb4 1 6 .ltd2 l:tb8 1 7 .ltc3 ltJc6 1 8 ltJd2 .lte6 1 9 .ltc4 'it'c8 ! (better than 19 . . . a5 ? ! 20 ltJb l ! intending ltJa3-bS ;1;;) 20 as i.d8 2 1 'it'a4 ltJb4 22 l:tec 1 ltJhS ! 23 i.xe6 'it'xe6 24 i.xb4 l:txb4 2S 'it'd l ltJf4 gave Black the initiative in Ivkov-Xie Jun, Vienna 1 993. 14 l:tb8 (D) A sensible alternative is 14 . . . .lte6 I S ltJdc4 ltJc6.

17 ::tbS We have seen this idea in Section 8. 1 . Here the as-pawn just drops off, but White has suffi­ cient play to draw. 18 .ltb3 ltJxaS 19 ltJxaS l:txaS 20 i.xe6 fxe6 21 ltJc4 l:txa1 22 'it'xa1 'it'c8 (D) •••

w

•••

w

23 'it'a2 White sets his sights on the e6-pawn and is able to force a repetition. 23 'it'c6 24 ltJaS 'it'd7 2S ltJc4 'it'c6 Both 2S . . . l:ta8? and 2S . . . 'it'c8 ? fail because of 26 ltJb6. 26 ltJaS 'ili'd7 27 ltJc4 112- 112 Milos-Adams, Istanbul Olympiad 2000. •••

lS ltJdc4 I S i.c4 ltJxc4 1 6 dxc4 saddles Black with a backward d-pawn, but the bishop-pair should compensate for this. 1 6 . . . i.e6 17 b3 l:tb7 gives the rook some scope along the second rank and prepares a knight manoeuvre. 1 8 l:tn ltJd7 1 9 'it'e2 ltJb8 2 0 ltJdS ltJc6 2 1 .ltb2 .ltgS 2 2 l:tad l as 23 f4 exf4 24 ltJxf4 i.xf4 2S l:txf4 f6 was equal in Svidler-Adams, FIDE Knockout, Mos­ cow 200 1 . lS ltJc6 16 i.d2 i.e6 17 as This leads to nothing, but White has trouble coming up with a plan. •••

Concl usions The Anti-Marshall with 8 a4 has lost a lot o f its lustre because of the modem 8 . . . b4, which is fa­ voured by most top players. Black has just about perfected a healthy scheme of development with . . . d6, . . . .lte6, . . . 'it'c8 and . . . l:tb8 which leaves White struggling to develop any initiative at all. The older 8 . . . i.b7 is not seen as much nowa­ days, but it remains perfectly viable as well.

9 Anti - M a rsha l l : 8 h3

1 e4 e5 2 lZ'lf3 lZ'lc6 3 i.b5 a6 4 i.a4 lZ'lf6 5 0-0 i.e7 6 :tel b5 7 i.b3 0-0 8 h3 (D)

c3) 1 1 d3 is relatively best and will be con­ sidered in note ' c ' to White's 1 1 th move in Line A. We now return to 9 d3 (D) :

B B

Since Black has come to terms with 8 a4, this has been a very popular Anti-Marshall line for the past decade. White makes a relatively use­ ful move and is willing to transpose back into the main lines of the Ruy Lopez. 8 i.b7 This keeps things in the realm of the Anti­ Marshall. Instead 8 . . . d6 is certainly playable, but after 9 c3 we are back in the main lines of the Ruy Lopez. This position would normally come about via the move-order 7 . . . d6 8 c3 0-0 9 h3. 9 d3 After 9 c3 ? ! Black gets an improved Mar­ shall with 9 . . . dS ! . White has: a) 1 0 d3 is passive. Black can exchange queens with 1 O . . . dxe4 or keep the tension with 1O .. :it'd6. In both cases Black has easy equality. b) 10 d4 gives Black a pleasant choice be­ tween 1 O . . . lZ'lxe4 1 1 dxeS lZ'laS 1 2 i.c2 i.cs and 1 O . . . dxe4 I I lZ'lxeS lZ'laS 1 2 i.c2 cS. c) 1 0 exdS lZ'lxdS and now: c 1 ) 1 1 lZ'lxeS ? lZ'lxeS 1 2 :'xeS lZ'lf4 gives Black a strong attack. For example, 1 3 d4 lZ'lxg2 14 �g4 lZ'lh4 :t. c2) 1 1 d4 exd4 1 2 cxd4 ( 1 2 lZ'lxd4 lZ'lxd4 1 3 'ii'xd4 c S + ) 1 2 . . . i.f6 1 3 lZ'lc3 lZ'lb6 ! ? gives Black good play. •••

Now that White has committed to 9 d3, the main line has always been 9 . . . d6, but recently Black has also investigated the 'Marshall move' 9 . . . dS ! ? We shall look at both. I SS A : 9...d5! ? B: 9... d6 I S9 Black can also play 9 . . . :te8 or 9 . . . h6, hoping either to relocate the e7-bishop before playing . . . d6 or to get in . . . dS in one move, but we shall not examine those moves here.

A) 9 d5! ? (D) This gambit had always been considered suspect, but thanks to the efforts of I.Sokolov, Aronian, and especially Bacrot, this bold ad­ vance has become an active option for Black. 10 exd5 lZ'lxd5 l l lZ'lxe5 Taking the pawn is critical, but sometimes White chooses to decline: a) I I lZ'lc3 lZ'lxc3 1 2 bxc3 'it'd6 1 3 a4 lZ'laS 1 4 i.a3 c S I S lZ'lxeS lZ'lxb3 1 6 cxb3 i.f6 1 7 d4 b4 1 8 cxb4 cxb4 1 9 i.b2 �dS 20 f3 :tfe8 with •••

1 56

UNDERSTANDING THE MARSHALL A TTACK

compensation, Shomoev-Pashikian, European Ch, Budva 2009. b) 1 1 lt:)bd2 f6 ( l l . . .lt:)f4 1 2 It:)e4 It:)a5 1 3 �xf4 exf4 1 4 d4 'it'd7? ! 1 5 It:)e5 'it'f5 1 6 'it'd3 l:.ad8 1 7 lt:)c5 gave White the initiative in Topa­ lov-Leko, Nanj ing 2009) 1 2 c4 (or 1 2 lt:)e4 'ifilh8 1 3 d4 exd4 14 It:)xd4 It:)xd4 1 5 'it'xd4 .l:.e8 1 6 �d2 lt:)b6 = Shirov-Bacrot, Moscow blitz 2007) 1 2 . . . bxc4 1 3 It:)xc4 �c5 14 It:)e3 It:)a5 1 5 d4 exd4 16 It:)xd4 It:)xb3 (or 16 . . . .l:.e8 ! ? =) 1 7 'it'xb3 �xd4 1 8 'it'xb7 �xe3 1 9 �xe3 'it'd7 20 .l:.ad 1 .l:.fd8 2 1 'it'b3 .l:.ab8 22 'it'a3 'it'b5 = Adams­ Bacrot, FIDE Grand Prix, Baku 2008. c) 1 1 c3 and here: c l ) 1 l . . .�f6 1 2 lt:)bd2 ! ? It:)f4 1 3 lt:)e4 lt:)xd3 1 4 lt:)xf6+ gxf6 1 5 .l:.e4 lt:)xc 1 1 6 .l:.g4+ 'ifilh8 1 7 'it'xc 1 gave White good compensation for the pawn in Grishchuk-Aronian, World Ch, Mex­ ico City 2007 . c2) 1 l . . .'it'd7 offers up the pawn on e5 again. 1 2 It:)xe5 It:)xe5 1 3 ':xe5 %lad8 14 'it'g4 f5 1 5 'it'g3 (after 1 5 'fi'h5 �f6 1 6 l:.e 1 Wh8 1 7 �g5 , Zagrebe1ny-Georgadze, Erevan Olympiad 1 996, Georgadze suggests that 17 .. :iVc6 1 8 f3 'it'b6+ is good for Black as 19 d4 c5 gives him the ini­ tiative) 15 . . .�f6 16 .l:.e 1 'ifilh8 1 7 lt:)d2 b4 gave Black counterp1ay in the game Topa1ov-Leko, More1ialLinares 2008. c3) 1 l . . .'it'd6 is the most solid. 12 It:)bd2 .l:.ad8 1 3 lt:)e4 'it'g6 (or 1 3 . . .'it'd7 14 a4 'ifilh8 1 5 axb5 axb5 1 6 d4 and now instead of 1 6 . . . exd4 1 7 cxd4 ;!;; Kamsky-Bacrot, E1ista 2007, Black can choose 16 . . . f5 1 7 It:)c5 �xc5 1 8 dxc5 e4 with counterplay) 1 4 lt:)g3 It:)f4 (not 14 . . . �c5 ? 1 5 d4 ! exd4 1 6 It:)h4 'ii'f6 1 7 It:)e4 ! 'it'xh4 1 8 �g5 trapping the queen, but making a little room for the queen with 14 . . . h6 ! ? is possible)

15 .i.xf4 exf4 1 6 It:)e4 It:)a5 1 7 �c2 c5 gave B lack good play in Mrdja-Djuric, Montecatini Terme 2005 . d) 1 1 a4 has been White's most recent at­ tempt to cause trouble, but Black has had the last word: 1 l . . .lt:)d4 ! 1 2 lt:)xd4 exd4 1 3 axb5 axb5 14 .l:.xa8 �xa8 1 5 It:)a3 �b4 ! (Shipov gives both 1 5 . . . 'it'd7 1 6 'it'g4 ! and 1 5 . . . �xa3 1 6 bxa3 c5 1 7 'it'g4 with the initiative for White) 1 6 .l:.e5 ( 1 6 �d2 will transpose) 1 6 . . . �d6 1 7 .l:.e 1 � b4 1 8 �d2 �xd2 1 9 'it'xd2 'it'f6 ! ( l 9 . . . �c6? 20 .l:.e5 b4? 2 1 lt:)b5 ! +- Leko-Yakovenko, FIDE Grand Prix, Elista 2008) and now: d l ) 20 It:)xb5 It:)f4 gives Black sufficient play ; for example, 2 1 f3 'it'h6 ! (2 l . . .�xf3 22 gxf3 'it'h4 also leads to a draw) 22 'ifilh2 (22 'it'f2? 'it'g5 + hits the knight on b5) 22 . . . �xf3 ! 23 gxf3 'it'xh3+ with a draw. d2) After 20 �xd5 a draw was agreed in the game Kariakin-Aronian, Wijk aan Zee 2009 . 20 . . . �xd5 2 1 It:)xb5 �xg2 ! 22 'ifilxg2 'it'c6+ is level. 1l lt:)d4! (D) This is the key move in Black's concept. 1 l . . .lt:)xe5 ? ! 1 2 .l:.xe5 �f6 1 3 .l:.e 1 c5 14 It:)d2 intending It:)e4 ± leaves Black struggling. •••

w

Black has a lead in development and can grab the bishop-pair with . . . It:)xb3 . Chasing the bishop with . . . a5 and . . . a4 is another idea. Black may also be able to harass the e5-knight be­ cause retreating to f3 may be difficult because of the pressure of the d4-knight and b7-bishop. This becomes clearer when Black plays a quick . . . It:)b4, which not only unleashes the b7-bishop but also puts pressure on c2. 12 lt:)c3

ANTI-MARSHAll:

This is the most natural and testing move, but there are other options: a) 1 2 lLlf3 lLlxb3 1 3 axb3 lLlb4 1 4 lLlbd2 .:te8 I S d4 cS 1 6 c3 lLld3 1 7 l:te3 lLlf4 1 8 lLln lLle6 gave Black enough for the pawn in Brkic­ Ivanisevic, Bosnian Team Ch, Vogosca 2007. b) 1 2 'it'g4 lLlxb3 1 3 axb3 fS I 4 'ii'f3 l:tb8 1 S i.f4 ? ! i.b4 1 6 .l:.n ( 1 6 i.d2 i.xd2 1 7 lLlxd2 'it'gS ! =+ hits the d2-knight and threatens both 1 8 . . . lLlb4 and 1 8 . . . lLle3) 1 6 . . . gS 1 7 i.d2 i.xd2 1 8 lLlxd2 lLlf4 1 9 'it'e3 lLlxg2 ( 1 9 . . . i.xg2 ! ? 20 ::tfe l .l:.b6 2 1 h4 .l:.h6 was recommended in New in Chess) 20 'it'g3 lLlf4 2 1 'iti>h2 (2 1 .:tfe l is a better try) 2 1 . . .lLle2 22 'it'e3 lLld4 =+ Spoelman­ I.Sokolov, Dutch Ch, Hilversum 2007 . c) 1 2 lLld2 and now: c 1 ) 1 2 . . . cS 1 3 a4 ( 1 3 lLldf3 lLlxb3 1 4 axb3 f6 I S lLlg4 .:te8 1 6 �d2 'it'd7 1 7 c4 lLlc7 1 8 lLlh4 b4 1 9 l:te3 i.d6 20 l:ta5 fS 2 1 lLlh2 lLle6 gave Black decent play for the pawn in Mamedov­ Citak, B aku 2009) 1 3 . . .lLlf4 and now White should try 14 i.xf7+ ! ? l:txf7 I S lLlxf7 Wxf7 1 6 lLle4 intending c3, as 1 4 lLlef3 ? (Kariakin-Bac­ rot, FIDE World Cup, Khanty-Mansiisk 2007) 1 4 . . . 'it'd7 ! ! , threatening I s . . . lLlxh3+, favours Black, who can meet I s lLlxd4 by I S . . . i.xg2 ! . c2) 1 2 . . . .l:.e8 1 3 lLldf3 (after 1 3 c3 lLlxb3 1 4 lLlxb3 c S I S 'it'hS g 6 1 6 'it'f3 f6 1 7 lLlxg6 hxg6 1 8 c4 'ii'd7 19 cxdS i.xdS 20 'iWg3 gS, Black was doing well in Hou Yifan-Kosteniuk, Women's World Ch match (game 3), Nalchik 2008) 13 ... lLlxb3 14 axb3 f6 I S lLlg4 'it'd7 16 �d2 cS 17 'ii'e2 hS 18 lLlgh2 i.d6 19 'ii'n 112-112 Ivan­ chuk-Svidler, Foros 2008. White is very pas­ sive so Black has excellent compensation for the pawn. d) 12 i.d2 (D) looks passive, but Yakovenko used this move to beat Aronian, so it cannot be taken too lightly. Black has : d l ) 1 2 . . . cS ? ! 1 3 lLlc3 lLlxb3 1 4 axb3 lLlb4 I S .l:.c 1 (this is a typical consolidating move: White protects the c2-pawn) I S . . . f6 1 6 lLlf3 (now this square is safe for White's knight be­ cause he can recapture with the queen in the event of any . . . �xf3 ideas) 1 6 . . . 'it'c7 (no better is 1 6 . . . fS 1 7 'it'e2 .i.f6 1 8 'it'e6+ Wh8 1 9 lLleS i.xeS 20 'it'xeS ± Vuckovic-Pavlovic, Serbian Ch, Mataruska Banja 2008) 1 7 lLle4 .l:.fe8 1 8 i.xb4 cxb4 1 9 c4 bxc3 20 bxc3 �a3 2 1 .l:.c2 l:tad8 22 'ii'a l .i.f8 23 l:tce2 l:te7 24 lLld4 .l:.de8 2S b4 ± Yakovenko-Aronian, FIDE World Cup,

8 h3

157

B

Khanty-Mansiisk 2007 . White has consoli­ dated his extra pawn and his knights have good squares in the centre. d2) 1 2 . . . 'it'd6 ! ? 1 3 lLlc3 lLlxb3 1 4 axb3 lLlb4 I S �f4 .l:.ae8 1 6 lLle4 'it'dS gave Black reason­ able compensation for the pawn in Parligras­ Ivanisevic, European Ch, Plovdiv 2008. d3) 12 ... aS ! ? 13 lLlc3 lLlf6 intending ... a4 was suggested in New in Chess Yearbook 86. d4) 1 2 . . . .l:.e8 100ks good. 1 3 lLlc3 lLlxb3 1 4 axb3 �d6 I S lLlf3 .l:.xe l + 1 6 lLlxe l lLlf6 1 7 i.gS h6 1 8 i.h4 g S 1 9 i.g3 �xg3 20 fxg3 'it'd6 2 1 �h2 l:te8 gave Black excellent play in Zhigalko-Pashikian, Martuni 2008. We now return to 12 lLlc3 (D) :

12 lLlh4 This finesse has been popularized by Bacrot. Black refrains from capturing on b3 and main­ tains the tension. By retaining the d4-knight, Black keeps pressure on the f3-square, but al­ lowing the b3-bishop to live means that Black must keep an eye on f7 too. •••

158

UNDERSTANDING THE MARSHALL A TTACK

The alternative is to take right away, and sometimes the two moves will transpose. After the immediate 1 2 . . . tt'lxb3 1 3 axb3 tt'lb4 (D), Black hopes that the bishop-pair and pressure on c2 will grant him sufficient compensation for the pawn.

White has: a) 14 i.e3 is covered in note 'a' to White's 1 3th move below. b) 14 d4 :e8 15 i.e3 i.d6 ( 1 5 . . . 'ii'c 8 was suggested by Quezada) 16 tt'ld3 :e6 ( 1 6 . . . tt'lxd3 1 7 'ii'x d3 b4 1 8 tt'la4 i.e4 1 9 'ii'e 2 'ii'f6, with compensation, is given by Quezada) 1 7 d5 ( 1 7 f3 ! i s a little better for White according to Quezada) 1 7 . . . tt'lxd5 1 8 tt'lxd5 i.xd5 1 9 tt'lf4 i.xf4 20 i.xf4 l:he l + 2 1 'ii'xe l c6 was drawn in Quezada-P.H.Nielsen, Capablanca Memo­ rial, Havana 2007. c) 14 i.f4 'ii'd4 1 5 i.g3 i.d6 ( 1 5 . . . :ae8 ! ?) 1 6 tt'le2 'ii'c 5 ! ? (after 1 6 . . . 'ii'x b2 1 7 c3 :fe8 1 8 ':b l White has a repetition but I do not see any­ thing more than that) 1 7 c3 i.xe5 1 8 i.xe5 'ii'xe5 1 9 cxb4 :ad8 20 'ii'd2 (despite White's ' Irish pawns' he maintains some initiative be­ cause he can use the c-file) 20 . . . :d7 2 1 %:tac l h6 22 l:tc5 'iVd6 23 tt'lf4 g5 24 tt'lh5 'iVg6 25 tt'lg3 11xd3 26 'iVc 1 i.d5 27 l:txc7 :xb3 28 'ilkc5 ;t Gormally-Beliavsky, European Union Ch, Liverpool 2008 . d) 14 i.d2 intends to cover c2 with :c 1 to allow tt'lf3 , so Black should act immediately: 14 . . .f6 15 tt'lg4 (Black's compensation is obvi­ ous after 15 tt'lf3 i.xf3 16 gxf3 'ii'd 7) and now 15 . . . f5 is also a typical advance. Black relin­ quishes the e5-square but the f-pawn may play an important role. 16 tt'le5 i.d6 and now:

d l ) 1 7 tt'la2 i.xe5 ! (Black gains time with this capture) 1 8 :xe5 tt'lc6 19 .:te l f4 20 i.c3 (Sokolov gives the line 20 f3 ? ! tt'ld4 21 tt'lc3 i.xf3 22 gxf3 'ii'g 5+ 23
13 tt'le4 It is not so easy to shake off Black's pres­ sure : a) 1 3 i.e3 tt'lxb3 1 4 axb3 i.d6 ( 1 4 . . . f6 1 5 tt'lf3 i.xf3 1 6 gxf3 f5 { 1 6 . . .'i!kd7, with com­ pensation, has also been played a couple of times } 1 7 tt'le2 i.d6 1 8 f4 'iVh4 gave Black the initiative in Prasad-Negi, New Delhi 2007) 1 5 i.d2 :e8 1 6 d4? ! (this allows a typical trick, so 1 6 tt'le4 ! ? is a better try) 1 6 . . . 'ii' h4 ! 17 tt'le2 'ii'e4 ! 1 8 tt'lf4 'ii'x c2 1 9 i.xb4 'ii'x d l 20 .:texd l i.xb4 '+ Kasimdzhanov-Bacrot, Mainz rapid 2007 . Black has recovered the pawn and has the bishop-pair.

8 h3

ANTI-MARSHALL:

b) 1 3 i.d2 ? ! i.d6 14 tilg4 (Black has com­ pensation after either 14 tile4 tilbxc2 ! 15 tilxf7 1::txf7 16 i.xc2 i.e5 or 14 l:Ic 1 tilxb3 15 axb3 l:Ie8) 14 . . . tilxb3 1 5 axb3 f5 ! 1 6 tilh2 (after 1 6 tile5 l:te8 1 7 d4 both 1 7 . . . 'it'h4 and 1 7 . . .c 5 give Black good play) 1 6 . . .'it'f6 1 7 l:Ic 1 'it'g6 1 8 tilf3 ':ae8 1 9 tilh4 'it'f6 20 tilf3 l:Ixe l + 2 1 i.xe l 'ifh6 + Lutz-Bacrot, Bundesliga 2007/8. c) 13 i.f4 ! ? c5 ? ! (it is probably better to play 1 3 . . . tilxb3 1 4 axb3, transposing to note 'b' to Black's 1 2th move) 1 4 ':c l .i.d6 15 tile4 tilxb3 1 6 axb3 i.e7 (if 1 6 . . . .i.xe5 ? 1 7 .i.xe5 'it'd5 ?, then 1 8 tilf6+ ! wins instantly) 1 7 tilg3 was Kasimdzhanov-Bacrot, Mainz rapid 2007. White has developed reasonably well, so it is not clear what Black has for the pawn. 13 tilxb3 14 axb3 f5 1 4 . . . 'it'd5 ? ! 1 5 tilf3 'it'd7 allows White to consolidate by 1 6 .i.d2 ! f5 1 7 tileg5 l:Ife8 1 8 tile5 'it'd5 1 9 tilgf3 (taking over the centre) 1 9 . . . 'it'd6 20 d4 i.e4 2 1 i.xb4 'it'xb4 22 tild3 'it'd6 23 tilc5 i.d5 24 'it'd3 ± Nijboer-Gupta, Wijk aan Zee 2009. White has an extra pawn and the better position. 15 tild2 (D) •••

1 59

a) 1 6 . . . 'it'd6 1 7 d4 i.e4 1 8 l:Ie2 l:Iad8 (Black could try 1 8 . . . tilc6 or 1 8 . . . c5) 1 9 c3 i.xe5 ? ! 20 tilxe5 tilc2 2 1 ':xe4 tilxal 22 :te l 'it'd5 23 i.g5 tilxb3 24 i.xd8 ':xd8 25 tild3 was much better for White in B .Vuckovic-M.Vujic, European Ch, Budva 2009 because of White's central grip and Black's poorly placed knight. b) 1 6 . . . c5 ! ? is also possible. The point is to protect the knight on b4, so that after 17 .i.d2 Black can play 17 . . . i.xf3 1 8 tilxf3 ( 1 8 'it'xf3 ? leaves c2 hanging, while 1 8 gxf3 'it'd5 gives Black obvious compensation for the pawn) 1 8 . . . i.xb2 1 9 l:Ibl i.f6, when he is not too badly off. c) Another idea is the direct 16 . . . i.xe5 ! ? 1 7 tilxe5 ( 1 7 ':xe5 i.xf3 again forces the ugly 1 8 gxf3) 1 7 . . . 'it'd5 1 8 tilf3 ( 1 8 f3 'it'c5+ wins the c2-pawn) 1 8 . . . 'iIi'd7 intending . . . .i.xf3 , which is not so easy for White to meet. 17 i.d2 17 c3 ? ! i.xe5 1 8 ':xe5 ( 1 8 tilxe5 'it'd5 1 9 tilf3 tilxd3 + ) 1 8 . . . 'it'xd3 was certainly no prob­ lem for Black in Matsenko-Kumosov, Satka 2008. 17 c5 18 d4 i.xe5 Perhaps Black should bring back the knight to fight for the centre with 1 8 . . . tilc6 ! ? 1 9 tilxe5? ! After this move, Black just takes back the pawn and has no problems at all. Mikhalevski suggests that White should try 1 9 ':xe5 ! . After 1 9 . . . i.xf3 20 gxf3 the pressure on c5 induces Black to play 20 . . . l:Ixe5, and after 2 1 dxe5 White can follow up with f4 and .i.e3 . 19 'it'xd4 20 i.xb4 'it'xdl 21 l:Iaxdl cxb4 22 tild3 l:Ixel + 23 l:Ixel ':c8 = Timofeev-Yakovenko, Russian Ch, Mos­ cow 2008. •••

•••

15 i.f6 After 1 5 . . . i.d6 1 6 tildf3 Black should prob­ ably transpose to the next note with 1 6 . . . i.xe5 , because 1 6 . . . c5 ? ! 1 7 i.d2 tild5 (Matsenko­ Khrushchov, Cheliabinsk 2008) can be met by 1 8 tilg5 ! 'it'c8 1 9 'it'h5 tilf6 20 'it'h4 h6 2 1 tilgf3 ±. 16 tildfJ l:Ie8?! This has been chosen by some strong players, but I believe that this is the time to look for an improvement in Black's play. A couple of ideas:

B)

•••

9 d6 (D) In playing 9 . . . d6, Black indicates that he is willing to play along more typical Ruy Lopez lines now that White has committed to 9 d3. This makes 8 h3 look rather non-critical - in normal Ruy Lopez lines this move is played to avoid the pin . . . i.g4 in preparation for advanc­ ing by d4. Now that Black has secured his e5pawn, he is threatening to grab the bishop-pair with . . . tila5 . Thus White's next move. •••

UNDERSTANDING THE MARSHALL A TTACK

160

10 a3 (D) 10 c3 transposes to the Pilnik Variation with White having played h3 rather early, which makes the line even more harmless. As men­ tioned previously, this is well covered in The

Ruy Lopez: A Guide for Black.

B

and . . . lLlbd7, or in 'Zaitsev' style with .. J:te8 and . . . i.f8. The alternative is to regroup with . . .'iWd7 and . . . lLld8, which enables Black to play . . . c6 and . . . lLle6. All of these plans are quite via­ ble and the choice is largely a matter of taste (or experience, if Black has played any of these other defences). For a few reasons, we shall focus on the first of these options. It is the most classical method and fits well with the best defences to other slow systems such as the Pilnik Variation men­ tioned above. I also think that is the easiest sys­ tem to conceptualize, because Black's piece configurations are natural in the sense that they are designed to fight for the centre. Lastly, I have chosen this system because it fights for space. In the main lines of the Ruy Lopez men­ tioned above (Chigorin, Breyer, Zaitsev) White quickly plays d4 and has a space advantage in the centre. In the Anti-Marshall this is not the case, and I think Black should take advantage of this by fighting for space himself, and 1 O . . . lLla5 best suits this purpose. 11 i.a2 c5 (D)

w

White's play is very flexible, but it is also very slow. I find it very difficult to believe that moves such as 8 h3, 9 d3, and 1 0 a3 can consti­ tute any kind of threat to the 7 . . . 0-0 move-order. White now has two plans of development, de­ pending on how he develops his queen' s knight. It most often develops to d2, from where it will usually go to fl and then either e3 or g3. The al­ ternative is to develop to c3 with an eye towards d5, which often leads to exchanges. The knight may also go to g3 via e2, but this is less flexible than the lLld2-fl route. 10 lLla5 Black has four main ways to arrange his pieces. He can develop in 'Chigorin' style with . . . lLla5 and . . . c5, in 'Breyer' style with . . . lLlb8 •••

Black has gained space but somewhat weak­ ened the d5-square. Now White must decide how to develop his pieces. Practice has shown that is not so easy for White to win the d5square. Because Black has the same pieces that White does, he can eventually focus all of them (except the e7-bishop, of course) to fight for this square. Some lines have some similarity to the Anti-Sveshnikov line I e4 c5 2 lLlf3 lLlc6 3 lLlc3 e5. Despite the apparently drastic weaken­ ing of d5, Black has resources to fight for the centre. Because this one-sided approach does

ANTI-MARSHAll:

not give Black difficult problems to solve, White often prefers a manoeuvring game where there is more scope to outplay the opponent. We examine: 161 B l : 12 lLlc3 1 64 B2: 12 lLlbd2

B1) 12 lLlc3 White takes aim at the dS-square, but this does not trouble Black too much, even after his next move. 12 lLlc6 (D) The knight heads back to the centre and eyes the d4-square. This loosens Black's control of dS for the moment, but Black has sufficient re­ sources to fight for the centre. •••

w

8 h3

1 61

b) 13 l:.bl is an odd move. White may play for a b4 advance, but this is not likely to terrify Black. 1 3 . . . l:.c8 1 4 jLd2 lLld4 1 5 b4 and now I s . . . lLlxf3+? ! 1 6 'ili'xf3 c4? ! 1 7 dxc4 bxc4 1 8 'ili'e2 'ili'c7 1 9 �gS left Black with problems with the c4-pawn and the dS-square in Kamsky­ Bacrot, Sofia 2006. Black can improve with I S . . . 'iIi'c7 ! . This overprotects the eS-pawn and puts pressure on the c-file and gives Black a good game. c) 1 3 lLle2 heads for g3, but this is less flex­ ible than manoeuvring with lLlbd2-fl -g3 be­ cause the knight is in White's way on e2 and the e3-square is not an option. 1 3 . . . jLc8 1 4 lLlg3 jLe6 and now: c 1 ) I s lLlfS jLxfS ! 1 6 exfS 'ili'd7 1 7 g4 h6 1 8 c3 l:.fe8 1 9 b4 cxb4 20 cxb4 jLd8 ! 2 1 jLb2 as gave Black the initiative in Gelfand-Adams, Wijk aan Zee 2002. c2) 1 5 c3 'ili'd7 ( 1 S . . . 'ii'c 7, anticipating the opening of the c-file, is also possible) 1 6 d4 jLxa2 1 7 :xa2 l:.fe8 1 8 b4 exd4 1 9 cxd4 c4 ! ? 20 jLd2 and now instead o f 20 . . . aS 2 1 dS lLleS 22 lLlxeS dxeS 23 bxaS jLd8 24 a4 �xaS 25 axbS 'ili'xbS 26 jLxaS %:txaS 27 l:.c2 ;j;; Mega­ ranto-Nguyen Anh Dung, Doha 2006, Black could try 20 . . . dS with the idea 2 1 eS lLle4 22 lLlxe4 dxe4 23 l:.xe4 fS ! with good counter­ play.

B11) 13 �gS (D)

Now White has a few continuations to choose from: 161 B 1 1 : 13 jL gS B12: 13 lLlh2 1 62 B13: 13 lLldS 1 63 There are some alternatives, none particu­ larly dangerous: a) 13 �d2 does not look very threatening. 1 3 . . . lLld4 1 4 lLlxd4 cxd4 1 5 lLle2 dS 1 6 lLlg3 (White's initiative was short-lived after 16 exdS jLxdS 1 7 jLxdS 'ili'xdS 1 8 lLlf4 1!i'cs 1 9 jLb4 'ili'c7 = in Holzke-Shomoev, Cappelle la Grande 2003) 1 6 . . . dxe4 1 7 dxe4 jLc8 ( 1 7 . . . g6 1 8 jLh6 l:.e8 also looks fine) 1 8 c3 jLcs 19 'ili'f3 jLe6 20 lLlfS l:.c8 2 1 l:.ed l l:.e8 22 cxd4 jLxd4 23 jLe3 jLxfS 24 'ili'xfS g6 25 'ii'f3 'ili'b6 = Topalov­ Carlsen, Wijk aan Zee 2008.

B

This intends to fight for control of dS . We shall see the same idea again in Line B2. 13 'iIi'd7! •••

UNDERSTANDING THE MARSHALL A TTACK

1 62

This looks better than 1 3 . . . lLld7 1 4 .i.d2 lLlb6 1 5 lLld5 ( 1 5 lLle2 .i.c8 1 6 b4 and now 1 6 . . . c4? ! 1 7 .i.e3 .i.e6 1 8 lLlc3 l::tc 8 1 9 .i.xb6 'ifxb6 20 dxc4 bxc4 2 1 lLld5 gave White a nice initiative in Gashimov-Beliavsky, Caleta 2009, but Black could try 16 . . . i.e6) 1 5 . . . lLlxd5 16 i.xd5 'ifc7 and now both 1 7 b4 and 17 c3 gave White some advantage in two games from the mini-match Bologan-Bacrot, Odessa rapid 2007 . 14 lLlhl 14 i.xf6 i.xf6 1 5 lLld5 i.d8 intending . . . lLle7 is fine for Black, who can successfully fight for the d5-square. 14 lLle8 This is a typical manoeuvre. The knight heads to c7 where it controls d5 . 14 . . . 'iti>h8 ! ? intending . . . lLlg8 has also been tried. 15 i.d2 lLlc7 16 lLln 'iti>h8 17 lLlg3 1 7 lLle3 i.g5 is fine for Black. 17 lLld4 18 lLlce2 lLlde6 19 b4 d5! (D) •••

•••

w

Black's pieces are all well placed, so striking in the centre is natural. 20 bxc5 i.xc5 2 1 .i.b4 :fe8 22 i.xc5 lLlxc5 23 lLlc3 :ad8 24 'ifh5 f6 gave Black an excellent position in Kasparov­ Adams, Linares 1 999.

13 lLld4 14 lLlg4 (D) Or: a) 1 4 lLlfl a5 ! ? 15 a4 b4 1 6 lLlbl ? ! ( 1 6 lLle2 and 1 6 lLlb5 ! ? both look like better tries) 16 . . . d5 1 7 i.g5 c4 ! ? ( 1 7 . . . dxe4 1 8 i.xf6 i.xf6 1 9 dxe4 .i.c6 +) 1 8 i.xf6 .i.xf6 1 9 dxc4 dxe4 20 lLlbd2 was murky in Bologan-Adams, French Team Ch, Bordeaux 2003. b) 1 4 i.g5 is another attempt to fight for the d5-square. White would like to capture on f6 and then manoeuvre his knights. 1 4 . . . lLld7 ! 1 5 i.xe7 'ifxe7 1 6 lLld5 'iVg5 (Black does not mind losing a little time because he can still fight for d5, but 1 6 . . . i.xd5 ! ? 1 7 i.xd5 .l:tab8 1 8 c 3 lLle6 1 9 lLlf3 lLlf4 also looked fine for Black in Wahls-G.Ginsburg, Wiesbaden 1 996) 17 c3 lLle6 1 8 lLlf3 'iVd8 19 a4 lLlf6 with equality, Ponomariov-Aronian, Russian Team Ch, Sochi 2006. •••

B

812) 13 lLlh2 (D) This is a common idea in many lines of the Closed Ruy Lopez - the knight can go to g4, where it may either eliminate the f6-knight or head to e3 and d5 . However, here White does not have the space in the centre that he usually enjoys. Because Black can play in the centre, this manoeuvre looks rather slow.

14 lLlxg4 Usually Black avoids making this exchange, but here White is unlikely to be able to use the •••

ANTI-MARSHALL:

h-file or g4-pawn to create a kingside attack. Also possible is 14 . . . 'ii?h 8, as played in Inar­ kiev-Lastin, Krasnodar 2002. 15 hxg4 .i.g5 ! (D)

Black trades off his inferior bishop and blocks White's play on the queenside. The chances are balanced; for example, 1 6 It:ld5 .i.c8 ! ? 1 7 c3 It:lc6 1 8 g3 .i.xc 1 1 9 l:.xc 1 .i.e6 20 'ii? g 2 It:le7 2 1 d4 ':c8 was equal in Leko-Grishchuk, Cap d' Agde rapid 2003 or 1 6 .i.d5 l:.b8 1 7 .i.xb7 l:.xb7 1 8 It:ld5 g6 19 c3 .i.xc 1 20 l:.xc 1 It:le6 2 1 g 3 'ii'g 5 2 2 It:le3 'ii? g 7 2 3 'ii? g 2 h 5 24 l:.h l (24 gxh5 l:.h8 25 hxg6? It:lf4+ is much better for Black) 24 . . . hxg4 25 'ii'xg4 'ii'xg4 26 It:lxg4 f5 , Stefansson-Beliavsky, European Ch, Warsaw 2005 .

813) 13 lt:ld5 This is the most direct move. 13 lt:lxd5 14 .i.xd5 'ii'c7 15 c3 (D) •••

8 h3

White has conquered the d5-square for the moment, but Black has pieces that can control the light squares too. 15 lt:lb8! Black exchanges off White's strong bishop. His knight may return to c6 or even head to d7, from where it can control d5 from f6 or b6. White's remaining knight is a long way from d5 and the position is relatively closed, so Black can take the time to do this. 1 5 . . . lt:la5 is also possible, although after 16 b4 ! .i.xd5 1 7 exd5 It:lb7 the knight i s out of play. It i s also difficult to understand 15 . . . 'ii? h 8 16 b4 l:.ac8 1 7 .i.e3 It:ld8 1 8 a4 ;t Navara-Berezjuk, Czech Ch, Luhacovice 2003 . 16 .i.xb7 'ii'xb7 Black has not experienced much trouble from this position. He has enough space and the d5square is under control. White's bishop looks somewhat better than Black's, but the pawn­ structure is not completely fixed yet and Black has a safe, flexible position. 17 d4 White can also manoeuvre his knight with 1 7 It:lh2, and here: a) 1 7 . . . lt:lc6 allows White to realize his idea, but 1 8 lt:lg4 f5 1 9 lt:le3 f4 20 It:ld5 .i.d8 in­ tending . . .It:le7 was fine for Black in Bologan­ Zviagintsev, Poikovsky 2004. b) 17 . . . lt:ld7 1 8 It:lg4 ( 1 8 It:ln looks a little more flexible, although 1 8 . . . l:lfe8 { after the natural 1 8 . . . lt:lf6 1 9 'ii'f3 ! intending .i.g5 and It:le3, White has a pull } 1 9 lt:le3 .i.f8 gave Black a solid position in Djukic-Antic, Serbian Ch, Kopaonik 2005) 1 8 . . .lt:lf6 19 It:le3 g6 20 b3 l:.ab8 (mysterious) 2 1 .i.b2 l:.fd8 22 'ii'f3 d5 , Shirov-Aronian, Wijk aan Zee 2007. Black has more space and it is now White who should be a little careful. 17 lt:lc6 Also playable is 1 7 . . . lt:ld7 ! ? 1 8 dxe5 dxe5 1 9 c4 (else . . . c4 and . . . It:lc5, but now Black will get play on the b-file and his own knight can head for d4) 1 9 . . . bxc4 20 'ii'e 2 l:.fd8 2 1 a4 (2 1 'ii'xc4 'ii'b 5) 2 l . . .l:.ab8 22 It:ld2 It:lf8 23 It:lxc4 lt:le6 ! ? 24 It:lxe5 It:ld4 2 5 'ii'c4 .i.d6 26 It:lf3 'ii'b 3 27 'ii'x b3 J:hb3 , when Black had compensation for the pawn in Nijboer-Sargisian, European Ch, Istanbul 2003 . 18 .i.e3 (D) 18 l:.fd8 •••

.••

B

1 63

•••

UNDERSTANDING THE MARSHALL ATTACK

1 64

B

and . . . i.e6. From e6 Black will cover the d5and f5-squares and challenge White ' s strong a2-bishop. This costs Black some time, but this need not be fatal - White is spending a lot of time on lengthy manoeuvres as well. If White exchanges bishops on e6, B lack ' s doubled pawns will control important central squares and Black will have an open f-file. 12 lLlc6 Black can also play 1 2 . . . i.c8 immediately, but this requires more delicate handling be­ cause for the moment both of Black's queen­ side minor pieces are away from the centre. 1 3 c3 and now: a) 1 3 . . . i.e6 1 4 i.xe6 fxe6 1 5 b4 ! ( 1 5 d4 lLlc6 1 6 dxc5 dxc5 was very solid for Black in Timoshenko-Aronian, Aeroflot Open, Moscow 2004) 1 5 . . . lLlc6 ( 1 5 . . . cxb4 1 6 axb4 lLlc6 has also been played, but there is no real reason to exchange these pawns) 1 6 'ii'b 3 'ii'c 8 1 7 bxc5 dxc5 1 8 a4 .:tb8 1 9 axb5 axb5, Akopian-Beliav­ sky, Calvia Olympiad 2004. Now Beliavsky gives 20 i.b2 intending c4, when the e5-pawn is rather weak. b) 1 3 . . . lLlc6 14 d4 cxd4 (also of interest are 1 4 . . . c4, Shirov-Anand, Amber Rapid, Monte Carlo 2004, and 14 . . . i.d7, Lutz-Anand, B un­ desliga 2003/4) 15 cxd4 exd4 1 6 lLlb3 d3 ! ( 1 6 . . . i.e6 1 7 lLlbxd4 lLlxd4 1 8 lLlxd4 i.xa2 1 9 .:txa2 ;l; Topalov-Shirov, Bastia rapid 2003 gives White a pleasant advantage) 1 7 i.f4 (Black's point is that 1 7 'ii'x d3 can be met with 17 ... i.e6 1 8 lLlbd4 lLle5 ! =) 1 7 . . . i.b7 1 8 'ii'xd3 lLle5 1 9 lLlxe5 dxe5 20 'ii'x d8 .:taxd8 2 1 i.xe5 lLlxe4 was seen in both Kasparov-Bacrot, Armenia­ RoW, Moscow 2004 and Cheparinov-Navara, Wijk aan Zee 2006. White has a tiny edge at most perhaps, but Black has no real prospects of his own. 13 lLln After 1 3 c3, 1 3 . . . i.c8 transposes to line 'b' above, while 13 ... d5 was suggested by Khalif­ man, reasoning that, compared to a regular Mar­ shall, White has played the slow moves c3, a3, and i.a2, while Black has managed to play . . . c5 . After 1 4 exd5 lLlxd5 1 5 lLlxe5 lLlxe5 1 6 ':'xe5 i.f6 Black has an active position. Whether this is worth a pawn is another matter, and it has not yet been tried in grandmaster practice, possibly be­ cause Leko's idea 1 3 . . . 'ii'd7 ! (D) is highly attrac­ tive. •••

1 8 . . . exd4 1 9 cxd4 i.f6 ! gave Black counter­ play in Lutz-Kasimdzhanov, Bundesliga 2003/4; e.g., 20 dxc5 i.xb2 2 1 .:ta2 i.c3 22 .:tee2 dxc5 23 i.xc5 .:tfd8 24 'ii'b 3 i.d4 (Kasimdzhanov). 19 'ii'e2 After 1 9 d5 both 1 9 . . . lLla5 and 1 9 . . . lLlb8 100k OK for Black. With two sets of minor pieces exchanged, Black has enough space and can create play with . . . c4 intending . . . lLld7-c5 . 19 lLla5 20 lLld2 cxd4 21 cxd4 .:tac8 22 .:tac1 ':'xc1 23 ':'xc1 ':'c8 = Iordachescu-Aronian, European Ch, Istan­ buI 2003 . •••

82) 12 lLlbd2 (D)

B

This is much more flexible than 1 2 lLlc3 . The knight heads for fl , from where it can go to ei­ ther g3 or e3, and White keeps the option of playing a quick c3 and d4. Black's best plan is to play . . . lLlc6 and the rather surprising . . . i.c8

ANTI-MARSHALL:

This connects the rooks and allows the pos­ sibility of . . . .i.d8 and even . . . 0.e7 . Black may also contemplate . . . dS , even at the cost of a pawn because he has a lead in development. 1 4 0. n (after 1 4 a4 Black should probably not abandon the a-file with 1 4 . . . .:ae8? ! 1 5 axb5 axb5 1 6 'ifb3 ! 0.a5 17 'ilkc2 0.c6 1 8 'fib3 0.a5 1 9 'ilkc2 0.c6 20 0.n d5 2 1 0.e3 ;t Inarkiev­ Gustafsson, Greek Team Ch, Ermioni Argolidas 2006 but instead prefer something like 14 . . . l:tfe8 or even 1 4 . . . i.d8 ! ?) 1 4 . . . d5 ! ? 1 5 exd5 ( 1 5 i.g5 dxe4 16 dxe4 c4 1 7 0.e3 l:tfd8 1 8 0.f5 'ilke6 1 9 'ilke2 i.f8 20 i.b1 h6 =+= 1/2- 1/2 Kramnik-Leko, World Ch match (game 6), Brissago 2004) 1 5 . . . 0.xd5 and now : a) Declining the pawn by 1 6 0.g3 should not be dangerous. 1 6 . . . l:tad8? ! (this is a bit careless; 16 ... f6 looks fine for Black) 17 0.xe5 0.xe5 1 8 l:txe5 i.d6?! (now bringing the rook to d8 100ks nonsensical) 1 9 l:te 1 i.xg3 20 fxg3 0.f6 2 1 i.e3 :fe8 22 'ilkc2 'ilkc6 23 :te2 :e7 24 ':ae 1 l:tde8 25 g4 h6 26 c4 ± Kariakin-Grishchuk, European Clubs Cup, Kerner 2007. b) 1 6 0.xe5 0.xe5 17 l:txe5 i.f6 1 8 l:te l l:.ae8 1 9 0.g3 ( 1 9 i.d2 and 1 9 l:txe8 are also possible) 19 . . . 0.xc3 ! 20 bxc3 i.xc3 21 i.e3 (af­ ter 2 1 ':xe8 l:txe8 22 i.e3, instead of 22 ... i.xa 1 ? ! 2 3 'ilkxa 1 'ilkxd3 2 4 .i.xc5 ;t Smirnov-Ovechkin, Internet rapid 2004, Black should play 22 . . . 'ilkc6 23 0.e4 i.xal 24 'ilkxal c4 25 f3 'ilkg6 ! ?, with unclear play, as given by Onishchuk) 2 1 . . .i.xe 1 22 'ilkxel 'iVxd3 23 'iVn was Akopian-Onishchuk, Calvia Olympiad 2004. Now Onishchuk gives 23 . . . l:txe3 24 fxe3 'iVxe3+ 25 'ilkf2 'ilkxf2+ 26 'itxf2 l:td8 27 l:te 1 'itfS with a slight advantage for Black. 13 i. c8! (D) .••

8 h3

1 65

Black continues with his plan: the bishop will come to e6. White has three important moves: B21 : 14 c3 1 65 B22: 14 i. g5 1 66 B23: 14 0.e3 1 67

82 1 ) 14 c3 This is the most direct. White begins actions in the centre. 14 ... i.e6 15 i.xe6 fxe6 16 b4 This is the best way to fight for the initiative. The alternative is 1 6 0.g3, and here: a) 1 6 . . . 0.d7 1 7 i.e3 dS 1 8 exdS exdS (Black's centre looks very strong, but it could become vulnerable) 1 9 a4 ( 1 9 b4 'ith8 20 'ilkb3 c4 2 1 dxc4 bxc4 22 'ilka4 'iVc8 was unclear in Shirov­ Onishchuk, FIDE Knockout, New Delhi 2000) 19 . . . l:tb8 20 axb5 axb5 2 1 b3 .:ta8 1/2- 1/2 Kaspa­ rov-Topalov, Linares 2005 . b) 1 6 . . . 'ilke8 (making . . . 0.h5 possible) 1 7 a4 ( 1 7 i.e3 0.h5 1 8 0.xh5 'ilkxh5 1 9 0.g5 1/2- 1/2 Svidler-Anand, Dortmund 2004) 1 7 . . . 'ilkd7 1 8 d4 exd4 1 9 cxd4 cxd4 (this i s solid enough, but I would prefer the dynamic 1 9 . . . c4) 20 0.xd4 0.xd4 2 1 'ilkxd4 bxa4 22 ':xa4 ':fb8 was pretty level in Shirov-Kasimdzhanov, Sarajevo 2003 . We now return to 1 6 b4 (D). This move creates tension in the centre. Both sides must consider multiple pawn advances and captures at each tum. 16. 'ilkd7 This is a solid, healthy move which protects e6 and connects the rooks. A couple of alterna­ tives: ..

1 66

UNDERSTANDING THE MARSHALL A TTACK

17 l:ttb8 (D) 1 7 . . . a5 I S bxc5 dxc5 1 9 'iVb3 (or 1 9 a4 b4, as in Svidler-Bacrot, Sofia 2006, when 20 'ili'b3 is likely to transpose) 19 . . ..:tabS 20 a4 b4 (with the a-file closed, Black's rooks are better placed here than they are on as and bS) 2 1 ':'d 1 bxc3 22 'iVxc3 lL'ld4 23 lL'lxd4 exd4 24 'iVc2 i.d6 was fine for Black in Bacrot-Aronian, Turin Olym­ piad 2006. .••

B

w

a) 1 6 . . . lL'lh5 1 7 lL'l l h2 lL'lf4 1 S i.xf4 l:txf4 1 9 'iVb3 'iVd7 20 a4 ! gave White a slight initiative in Adams-Kasimdzhanov, Linares 2005 . b) 1 6 . . . d5 1 7 lL'lg3 ( 1 7 exd5 'ilxd5 ! ? I S 'iVe2 l:tfdS 1 9 lL'lxe5 'iVxe5 20 'iVxe5 lL'lxe5 2 1 l:txe5 l:txd3 22 l:txe6
Black prepares for the opening of the queen­ side. If White sits still, l S . . . a5 will come next. 1 8 d4 exd4 19 exd4 exd4 Black could consider 1 9 . . . c4 with a double­ edged position. 20 lL'lxd4 lL'lxd4 21 'iVxd4 :le8 22 i.b2 l:te2 23 lladl 'iVa7! 24 e5 dxe5 25 'iVxa7 l:txa7 26 i.xe5 l:td7 Black has sufficient counterplay, Leko-Kram­ nik, Rapid match (game 4), Miskolc 2007.

822) 14 i.g5 (D) White looks to exchange on f6 to help con­ trol the d5-square. 14 lL'le8 Black would rather exchange bishops than allow White to capture on f6. 14 . . . lL'ld7 bottles up his queenside pieces a bit, but Black can also ignore White's 'threat' and play 14 . . . i.e6. White then has : a) 1 5 lL'le3 is harmless. Black can displace the white rook and avoid the exchange of his f6-knight: 15 . . . i.xa2 16 l:txa2 lL'ld7 17 i.. xe7 lL'lxe7 I S c3 'fIc7 1 9 a4 lL'lb6 20 axb5 112- 112 Ivanchuk-Grishchuk, Foros 2006. •••

ANTI-MARSHALL:

8 h3

167

to exchange bishops avoids strengthening his central control, but White will waste time with his rook on a2. 16 .txa2 17 ':xa2 tLlc7 From here the knight controls d5 and it may go to e6. 18 b3 Or 1 8 c4 lIb8 1 9 cxb5 axb5 20 b4 'ilVd7 = Vescovi-Almasi, Wijk aan Zee 2006. 18 g6 19 c4 lIb8 20 .tc3 'ilVd7 21 tLld5 tLlxd5 22 exd5 tLld4 = Bruzon-Aronian, Turin Olympiad 2006. •••

•••

b ) 1 5 �xe6 i s illogical. The bishop has no business being on g5 now. After 1 5 . . . fxe6 1 6 c3 tLld7 1 7 �d2 'it'e8 Black had a very healthy posi­ tion in McShane-Aronian, Bundesliga 2006n. c) 1 5 �xf6 is the only consistent move. 1 5 . . . .txf6 1 6 tLle3 .tg5 1 7 tLld5 .th6 1 8 b4 lIc8 19 bxc5 dxc5 20 a4 'ilVd6 21 axb5 axb5 22 tLlc3 b4 23 tLld5 tLle7 (23 . . . lIa8 24 .tb3 J:ta3 , with counterplay, was suggested by Wei Ming) 24 .tc4 tLlxd5 25 exd5 .txd5 26 tLlxe5 .txc4 27 tLlxc4 ;!;; Gashimov-Bacrot, FIDE Grand Prix, Baku 2008. White's knight is better than Black's bishop. 15 .td2 15 .txe7 is not as harmless at it may look, but Black should be OK: 1 5 . . . tLlxe7 1 6 tLle3 tLlf6 ( 1 6 . . . .te6 is also possible) 1 7 c3 'it'c7 1 8 tLlh2 .te6 1 9 tLlhg4 tLld7 ( 1 9 . . .tLlxg4 20 hxg4 is not bad for Black in such a simplified position) 20 'ilVf3 c4 ! ? 2 1 tLlf5 was Anand-Kasimdzha­ nov, Leon 2005 . Now Black should play a solid move like 2 1 . . .lIfe8 or 2 1 . . .lIae8 with fair pros­ pects. 15 �e6 Worse is 1 5 . . . tLlc7? ! 1 6 tLle3 .te6 1 7 tLld5 and White is slightly better, Gashimov-Pono­ mariov, Russian Team Ch, Dagomys 2008. Black cannot take on d5 with the knight and if he takes with the bishop he will give White the bishop-pair and the knight on c7 will lack pros­ pects. 16 tLle3 Black experienced no difficulties after 1 6 .txe6 fxe6 1 7 b4 a5 1 8 c 3 axb4 1 9 axb4 lIxal 20 'ilVxal tLlc7 2 1 tLle3 'ilVd7 22 'ilVb2 lIa8 23 lIal in Anand-Aronian, MorelialLinares 2007 and the game was drawn here. Allowing Black

823) 14 tLle3 This is the most testing continuation. 14 .te6 (D) •••

w

Here 1 5 tLld5 is the obvious move, but Sutov­ sky's idea 1 5 .td5 is also interesting. 1 67 B23 1 : 15 .td5 1 68 B232: 15 tLld5

•••

823 1 ) 15 .td5 White hopes to disturb Black's piece coordi­ nation with this move. 15 .td7 ! ? Black would like to capture on d5 with the knight, but of course to do so immediately would lose a piece. This move was Aronian's choice against Anand, but there are some de­ cent alternatives: a) 15 ... �xd5 is rather obvious, and it is not so clearly bad either. 1 6 exd5 tLlb8 1 7 b4 ! •••

UNDERSTANDING THE MARSHALL A TTACK

1 68

liJbd7 1 8 bxcS liJxcs 1 9 d4 exd4 20 liJxd4 was Almasi-Navara, Wijk aan Zee 2006. White has some initiative, but the position is still rather unclear because White ' s pawn-structure is a bit disjointed. b) With l S . . . .:.c8, Black wants to develop his rook before retreating his knight to b8, but White can keep some initiative: 1 6 c3 �xdS 1 7 exdS liJb8 1 8 a4 'it'd7 1 9 axbS axbS 20 c4 ;j;; Sutovsky-Beliavsky, European Team Ch, Goth­ enburg 2OOS . c) l S . . . 'it'd7 is a typical and solid response. 16 c3 �d8 17 b4 ( 1 7 a4 was met by 17 . . . �aS ! ? 1 8 axbS axbS 1 9 'fib3 �xdS 20 liJxdS liJxdS 21 'it'xdS liJe7 22 'it'b3 h6 = in Sutovsky-Gri­ shchuk, Turin Olympiad 2006, and 17 . . . �b6 also looks reasonable) 17 . . . cxb4 ? ! (this looks a little cooperative; again, 1 7 . . . �b6 is a reason­ able option) 1 8 axb4 �b6 19 'it'b3 was a little better for White in Shirov-Grishchuk, Foros 2006. 16 .Jtb3 16 .Jta2 �e6 repeats, but White could con­ sider allowing Black to take on dS and get on with it by 1 6 b4 ! ? 1 6 �e6 17 c3 (D)

20 a4! b4 21 .Jtd2 'it'c7 22 liJh2! l:tfd8 23 liJhg4 liJxg4 24 liJxg4! This is better than 24 hxg4 'it'd7, tying White to the defence of the g4-pawn. After 24 liJxg4 White has excellent piece-play and his queen is very strong on a2. White had a nice advantage in Anand-Aronian, Wijk aan Zee 2006.

8232) IS liJdS (D)

B

•••

B

White does not mind allowing the exchange of bishops if he can recapture with his queen on b3 (instead of with his rook on a2). 17...�xb3?! Now White can develop an initiative. Instead 17 . . . 'it'd7 intending . . . �d8 100ks very solid. 18 'it'xb3 g6 19 'it'a2! as This allows White to further his campaign on the light squares. 19 . . . 'it'd7 20 b4 as is a flex­ ible alternative.

It may seem strange to make the journey liJd2-fl -e3-dS instead of the more direct liJc3dS, but Black has spent time playing . . . �b7c8-e6, and now he cannot capture on dS with the f6-knight without losing a piece. So Black must either capture with the e6-bishop, when White will enjoy the bishop-pair, or play around the dS-knight for the time being. IS ...�xdS This is by far the most common move, but Black should consider the alternatives: a) The 'Breyer' idea lS . . . liJb8 avoids the fork on dS and thus prepares to capture with the f6-knight, but it still looks rather slow. Gri­ shchuk thought it was worth a try anyway, in a rapid game at least. 1 6 liJxe7+ 'it'xe7 1 7 �bl liJc6 18 liJgS (Kha1ifman suggests 18 c3 in­ tending d4) 18 . . . �d7 19 f4 exf4 20 .Jtxf4 l:tae8 gave Black a solid position in Anand-Grishchuk, Cap d' Agde rapid 2003 . b) l S . . . liJd4 looks very logical . 1 6 liJxe7+ ( 1 6 liJd2 liJxdS 17 exdS �d7 gives us the same pawn-structure as in the main line but Black has not given up the bishop-pair) 1 6 . . . 'ii'x e7 1 7 c3 liJxf3+ 1 8 'it'xf3 liJd7 was fine for Black

ANTI-MARSHALL:

in Zhigalko-Beliavsky, European Ch, Plovdiv 200S. 16 exd5 lLld4 1 6 . . . lLlbS 1 7 b4 cxb4 I S axb4 lLlbd7 has also been tried, but after 19 iLd2 (stopping . . . a5) White was better in Al Modiahki-Kaplan, Biel 2007. 17 lLld2 White keeps pieces on the board and intends to drive the d4-knight away with c3. This is the most common, but it is not the only move. a) 1 7 b4 lLlxf3+ I S 'fi'xf3 lLld7 is fine for Black according to Khalifman. Black is happy to exchange another set of pieces and may try to trade another with . . . iLgS . b) 1 7 lLlxd4 cxd4 I S c3 dxc3 1 9 bxc3 lLld7 20 a4 'fi'c7 2 1 iLb2 lLlcs 22 iLb3 l:tfcS 23 iLc2 lLld7 was fine for Black in Ehlvest-Vescovi, Cali 2007 . White's bishops are not very threat­ ening. c) 1 7 lLlh2 looks a bit odd here, but it has similar ideas to the main line. 1 7 . . . lLlfS I S c3 lLld7 ( l S ... g6 ! ? Pavlovic) 1 9 d4 iLgS ( l 9 ... cxd4 ! ? 2 0 iL b 1 lLlh4 2 1 cxd4 l:tcs i s another idea) 20 dxeS iLxc 1 2 1 l:txc 1 lLlxeS 22 iLb1 'it'gS 23 l:te4 gave White some initiative in Kasimdzhanov­ Akopian, Russian Team Ch, Sochi 2006. We now return to the position after 1 7 lLld2 (D) :

8 h3

1 69

This was suggested b y Khalifman a s well. 1 9 b 4 is more common. Black then has: a) 1 9 . . . ':cS 20 iLd2 g6 ! ? 2 1 c3 lLlg7 22 c4 lLlfS ! ? (now that the d4-square is open again) 23 a4 cxb4 24 cxbS a5 was unclear in Anand­ Svidler, Amber Rapid, Monaco 2006. b) 1 9 . . . cxb4 ! ? (Black immediately fights for the initiative) 20 axb4 as 2 1 bxaS 'fi'xaS 22 iLd2 'fIIc 7 23 c3 lLlh4 ! 24 g3 lLlg6 2S h4 fS 26 lLlgS iLxgS 27 hxgS lLlcs was excellent for Black in Gopal-Kasimdzhanov, FIDE World Cup, Khanty-Mansiisk 2007 . 19 g6 Khalifman suggests the immediate 1 9 . . . l:tcS. 20 h4 ':cS 20 . . . a5 ! ? 2 1 b4 'fi'b6 (D) 2 l . . .lLlg7 may be better. •••

w

B 22 h5 Leko-Kramnik, Rapid match (game 2), Mis­ kole 2007. Black should probably maintain the tension by 22 . . . 'fllc 7 23 c3 as , with counterplay.

Concl usions

17 lLlf5 1S lLle4 I S b4 'fIIc 7 1 9 iLb2 is given by Khalifman, but this does not look too dangerous. IS lLld7 Here too Pavlovic suggests l S . . . g6 with the idea . . . lLlxe4, . . . lLlg7, . . . JLgS and . . . fS . 19 g3! ? •••

•.•

This modem line has been White's main choice to avoid the Marshall for a while now, but Black is doing just fine. It is hard to believe that Black's set-up can be challenged by moves such as S h3, 9 d3 and 10 a3 . Black has more than one logical way to develop, and the classical 1O . . . lLla5 1 1 iLa2 cS remains sufficient. Black has also re­ cently added the Marshall thrust 9 . . . dS to his ar­ senal, showing that White's quiet play does not promise him a quiet life.

1 0 Other Anti - M a rsha l l Li nes

In this chapter we look at some loose ends where White avoids the main lines. In Section 10. 1 we examine the slow 8 d3, which can also arise if White plays d3 even ear­ lier. This is a line that needs to be prepared by any Ruy Lopez player, along with lines like the Exchange Variation, Worrall Attack, and oth­ ers. It was well covered in The Ruy Lopez: A Guide for Black but we shall take a look at it here too, especially as it is a common reply to the 7 . . . 0-0 move-order at amateur level when White decides to avoid the Marshall and lacks a well-worked-out repertoire. The second part of this chapter features the very slow 8 a3, which, although not as harmless as it seems at first sight, should not be too diffi­ cult to handle. Section 1 0.3 explores the sharp Anti-Mar­ shall 8 d4, which is quite popular, especially as a surprise weapon. Black's most solid response is 8 . . . d6. This is covered in The Ruy Lopez: A Guide for Black, so here we shall only look at 'pure' Anti-Marshall systems for Black, which can become quite tactical. The fourth section of the chapter investigates a line that is also usually employed as a surprise weapon, but if Black knows the positional ideas White's system is not very dangerous. In the last part of this chapter we survey lines where White enters the Marshall but then de­ cides against capturing the pawn on move 1 0. Thus we have: Section 10. 1 : S d3 1 70 1 72 Section 10.2: S a3 1 73 Section 10.3: S d4 1 83 Section 10.4: S c3 d5 9 d4 Section 10.5: S c3 d5 9 exd5 ttJxd5 1 87 10 others

Section 1 0 . 1 1 e4 e5 2 ttJf3 ttJc6 3 iLb5 a6 4 iLa4 ttJf6 5 0-0 iLe7 6 l:te1 b5 7 iLb3 0-0 S d3 (D)

B

This move is often played by those who want to avoid theoretical lines. White will play slowly with d3 and c3. The delay in playing d4, how­ ever, does put less pressure on Black. S d6 Black can also opt for 8 . . . iLb7, but after 9 ttJbd2 (9 a4 d6 is Chapter 8) 9 . . . d6 1 0 a3 ! ? ttJa5 1 1 iLa2 c5 1 2 ttJn the play resembles that of Chapter 9. However, here White has not spent a tempo on the move h3 and that should favour White somewhat. 9 c3 This is sometimes referred to as the 'Pilnik Variation' and this is well covered in The Ruy Lopez: A Guide for Black. 9 a4 leads to Chapter 8 after either 9 . . . h4 or 9 . . . iLb7 . Instead, 9 a3 ttJa5 1 0 iLa2 c5 is once again similar to Chapter 9. Here too, White has not played h3, but this time Black has not committed to . . . iLb7 yet ei­ ther, so Black may actually end up with a better version of those lines because he will not spend time playing . . . iLb7-c8 . 9 ttJa5 10 iLc2 c5 1 1 ttJbd2 ttJc6 1 1 . . . :e8 is given in The Ruy Lopez: A Guide for Black. The text-move may lead to similar play. 12 ttJn l:teS (D) There are other moves here, but we shall do what Aronian does ! Black's set-up is very simi­ lar to that in Chapter 9, but here he has not spent •••

•••

OTHER ANTI-MARSHALL LINES

1 71

15 exd4 16 cxd4 cxd4 17 lLlxd4 lLlxd4 18 'it'xd4 ':'c8 19 s.. b3 (D) •••

B

time playing . . . .i.b7-c8, while White has not played h3 or a3 . 13 h3 So White spends a tempo on this move after all. Instead 1 3 a3 h6 14 b4 .i.e6 I S lLlg3 dS ! 1 6 exdS 'it'xdS already saw Black taking the initia­ tive in Shirov-Aronian, Candidates match (game 6), Elista 2007. The probing 1 3 s.. gS is well met by 1 3 . . . s..e6 14 lLle3 lLlg4, exchanging pieces. After 1 3 a4 s..e6 14 lLle3, which was seen in J.Polgar-Aronian, Hoogeveen 2003, Polgar rec­ ommends 1 4 . . . dS ! . 13 h6 14 lLlg3 1 4 a4 can be met by either 1 4 . . .l::t b 8 or 1 4 . . . i.e6, while the immediate 14 d4 is well met by the liquidation 14 . . . cxd4 I S cxd4 exd4 1 6 lLlxd4 lLlxd4 1 7 'it'xd4. This is a common motif in many lines of the Closed Ruy Lopez. The evaluation of the position revolves around piece activity. Here Black seems to be doing quite well, which is not surprising considering White took two moves to play the d4 advance. A couple of examples: a) 17 . . . s..b7 1 8 lLlg3 dS 19 eS lLle4 20 lLlxe4 dxe4 = Kasparov-Svidler, Linares 1 999. b) 17 . . . s..e6 1 8 s.. f4 ':c8 19 s.. b 3 .i.xb3 20 axb3 dS 2 1 eS lLle4 22 lLle3 s.. c S (Black al­ ready has the initiative and Ivanchuk gradu­ ally outplays his less experienced opponent) 23 'it'd3 'it'h4 24 g3 'it'xh3 2S 'it'xdS s.. x e3 26 'it'xe4 .i.xf4 27 'it'xf4 'iVe6 '+ Lahno-Ivanchuk, Benidorm 2008. 14 i.e6 15 d4 White took an extra move to secure his cen­ tre before making this advance. With the bishop somewhat exposed on e6, Black is forced to give up the centre, but this is good enough.

19 i.xb3 If Black wants a more complicated game, he can play 19 . . . dS 20 eS lLld7 as in Svidler­ Aronian, Amber Blindfold, Monte Carlo 2007 . 20 axb3 d5 21 e5 i.c5 22 'iVf4 lLle4! Black must stay active. A pawn will drop but Black has enough activity to hold the balance. 23 lLlxe4 dxe4 24 ':'xa6 'it'd3! (D) •••

•••

•••

w

25 'iVxe4 s..xf2+! This temporary sacrifice leads to a drawn ending. 26 �2 .:tc2+ 27 �gl 'it'xe4 28 :Xe4 :Xci + 29 �h2 ':'dl This is probably more precise than 29 . . . ':'c2, as in Fressinet-Naiditsch, Bundesliga 2007/8, al­ though Black drew quickly in that game as well. 30 ':'a7 �f8 31 ':'f4 ':'e7 32 ':'a8+ ':'e8 33 ':a7 ':e7 34 :a8+ 1/2- 1/2 Leko-Aronian, MoreliaiLinares 2008.

1 72

UNDERSTANDING THE MARSHALL A TTA CK

Section 1 0 . 2

one go, so I present the text-move as an alterna­ tive.

1 e4 eS 2 liJf3 liJc6 3 i.bS a6 4 i.a4 liJf6 5 0-0 i.e7 6 l:tel bS 7 i.b3 0-0 8 a3 (D) w

B

This is an odd little move. White does not forsake advancing with d4 in one go, but he re­ tains the option of playing d3 too. 8 d6 After 8 . . . i.b7 9 d3 d6, 10 h3 transposes to Chapter 9, but White can try to save time on this move and play 10 liJc3 or 1 0 liJbd2 instead. Black sometimes plays 8 . . . i.cS, reaching a kind of Arkhangelsk Variation where White has played the rather slow advance a3 (usually in that line White plays a quick a4 and c3 and avoids .::te l ), but this leads to positions very dif­ ferent from the ones covered in this book. The text-move is simple and sound, and also gives Black a relatively rare option that is not difficult to learn against the rare move 8 a3 . 9 c3 After 9 h3, Black can choose 9 . . . i.b7 1 0 d3, when we are in Chapter 9, while if he prefers the lines of that chapter without . . . i.b7, 9 . . . liJaS 1 0 i.a2 cS could be played. This position some­ times comes about from the move-order 7 . . . d6 8 c3 0-0 9 a3, which is attributed to Suetin. 9 i.e6 ! ? (D) It would seem more consistent to continue 9 . . . liJaS but after both 1 0 i.c2 cS and 1 0 i.a2 cS White can play 1 1 d4. This is hardly fatal, be­ cause after 1 1 . . :fj'c7 Black has a Chigorin set­ up where White has played the slow move a3 . However, in general I have tried to avoid these Closed positions where White gets to play d4 in ...

.•.

By playing this move, Black essentially es­ says the rare Kholmov Variation which comes about after 7 . . . d6 8 c3 0-0 9 h3 i.e6. Here White has played a3 instead of h3, which somewhat favours Black. In the Kholmov line White's main try for an advantage is 10 d4 i.xb3 1 1 axb3, which clearly will not be possible here. Although the move a3 does cover b4 (the move h3 covers g4, which White is missing in this po­ sition), it also weakens the b3-square, and this can favour Black in some lines. 10 d4 (D) This move is White's only try for an advan­ tage. 1 0 i.c2 dS 1 1 exdS 'ilVxdS is fine for Black, while 1 0 i.xe6 fxe6 1 1 d4 'ilVe8 ! 1 2 dxeS liJxeS l 3 liJxeS dxeS already gave Black the ini­ tiative in Short-Ivanchuk, Dubai (rapid) 2002. The doubled pawns cover important squares and Black is ready to play . . . l:td8 and perhaps even . . . liJhS .

B

OTHER ANTI-MARSHALL LINES

10 .i.xb3 11 'it'xb3 l:te8 (D) This solid move has been played by both Aronian and Ivanchuk, so we shall take it as the main line. Other moves are also possible: a) l l . . .d5 ! ? 12 exd5 iLla5 1 3 'it'c2 exd4 1 4 cxd4 iLlxd5 1 5 iLlc3 l:te8 was Suetin-Lukacs, Leipzig 1 986. Black has good squares for his pieces, but White still has a slight initiative. b) l l . . .iLld7 ! ? 1 2 'it'c2 (Greenfeld points out the trick 1 2 'it'd5 iLla5 1 3 dxe5 c6 14 'ii'd4? iLlb3, exploiting the weakened b3-square) 12 ... iLla5 1 3 iLlbd2 exd4 1 4 cxd4 c 5 1 5 b4 (this leads to a structure where both sides have a dangerous pawn-mass) 1 5 . . . iLlc6 1 6 bxc5 dxc5 17 d5 iLlce5 1 8 iLlxe5 iLlxe5 1 9 .i.b2 .i.f6 20 f4 iLld3 2 1 e5 c4 22 .:tf1 'iWb6+ 23 �h l was Grishchuk-Avrukh, European Clubs Cup, Saint Vincent 2005 . Here Black should probably play 23 . . . .i.d8 ! with a complicated game. This system can be studied in tandem with the proper Kholmov Variation in which White plays 10 d4 .i.xb3 1 1 'it'xb3 (instead of 1 1 axb3 !).

1 73

.••

w

16 b3 c6 17 c4 l:tac8 18 'it'dl cxd5 19 cxd5 iLlh7 20 a4 f5 Black had good counterplay in Topalov­ Ivanchuk, Frankfurt rapid 2000.

Section 1 0 . 3 1 e4 e5 2 iLlfJ iLlc6 3 .i.b5 a6 4 �a4 iLlf6 5 0-0 .i.e7 6 .:tel b5 7 �b3 0-0 8 d4 (D)

B

12 'ii'c2 'ii'd7 After 1 2 . . . .i.f8 1 3 .i.g5 ! h6 14 .i.h4 (it is probably better to play 14 .i.xf6 'ii'xf6 15 d5 iLle7 1 6 iLlbd2, when the attempt to fight in the centre with 1 6 . . . c6 1 7 dxc6 iLlxc6 leaves the d5-square weak) 1 4 . . . g5 ! ? 1 5 .i.g3 g4 1 6 iLlh4 exd4 1 7 iLld2 dxc3 1 8 'it'xc3 iLle5 1 9 iLlf5 gave White some compensation for the pawn, but no more than that, in Ponomariov-Aronian, FIDE World Cup, Khanty-Mansiisk 2005 . 13 iLlbd2 .i.f8 14 d5 iLle7 15 iLln h6! (D) Black wants to keep both of his knights to help fight for the light squares in the centre, so he takes a moment to prevent .i.g5 .

White looks to open the centre immediately. This can lead to sharp play and sometimes it is White, and not Black, who ends up sacrificing a pawn. This is a reasonable option for White if he wants to avoid the main lines, especially if he is well versed in the Yates Variation, 7 . . . d6 8 c3 0-0 9 d4. 8...iLlxd4 This move gives the game an independent flavour. Instead 8 . . . exd4?! runs into the annoy­ ing 9 e5 iLle8 10 .i.d5 ! . However, 8 . . . d6 is a sound choice. Then White's best is probably 9

1 74

UNDERSTANDING THE MARSHALL A rrACK

c3 transposing to the Yates Variation, which normally comes about after 7 . . . d6 8 c3 0-0 9 d4. This is well covered in The Ruy Lopez: A Guide for Black. After the text-move (8 . . . liJxd4), 9 liJxeS would allow Black to grab the bishop-pair with 9 . . . liJxb3, when only Black can be better, so White must choose between the tricky 9 i.. xf7 + and the normal 9 liJxd4. A: 9 i.. xf7+ 1 74 178 B: 9 liJxd4

A) 9 i.. xf7+ (D)

d) 1 O . . . liJe6 ? ! 1 1 liJxf7 ct;xf7 1 2 eS i..b7 ( 1 2 . . . liJe8 1 3 'ii'f3+) 1 3 exf6 .i.xf6 1 4 liJc3 ± . Black does not have enough for the exchange. 11 'ii'xd4 c5 (D) Black is a pawn down, so he must play ag­ gressively. This is the best way to fight for the initiative. Instead 1 1 . . .i..b7 1 2 liJc3 cS gener­ ally leads to the same position after 1 3 'ii'd l , but we shall take 1 1 . . .cS as the main line because it is more popular and it will also allow us to look at other retreats by the white queen. One independent possibility is 1 1 . . .i..b7 1 2 c4. After 1 2 . . .c S 1 3 'ii'd 3 "fIc7 1 4 liJf3 bxc4 I S 'ii'xc4+ d S 1 6 exdS i..x dS 1 7 'ii'e2 i..d6 1 8 liJc3 i.. xf3 19 'ii'xf3 i.. xh2+ 20 �f1 .l:.ad8 2 1 g3 liJhS ! the complications led to a draw in Kam­ sky-Leko, FIDE Grand Prix, Jermuk 2009: 22 'ii'xhS lhf2+ 23 ct;xf2 'ii'x g3+ 24 c;tJe2 'ii'g 2+ 2S ct;e3 'ii'g 3+ 26 ct;e2 'ii'g 2+ 27 ct;e3 'ii'g 3+ 28 ct;e2 'ii'g 2+ 112- 112.

w

This sharp move is rather surprising. For a long time it was thought to favour Black, but af­ ter Ivanchuk used it to defeat Leko, it became clear that Black should not ignore this tactical possibility. 9 .l:.xf7 10 liJxe5 1lf8 Black does best to return the piece immedi­ ately. Attempts to win material run into trou­ ble because of the possibility of White playing the eS advance in combination with 'ii'f3+ or 'ii'd S+. Some examples: a) 1O . . . .i.cS ? fails to l l liJxf7 ct;xf7 12 i..e 3 liJe6 1 3 .i.xcs liJxcs 14 eS and White wins ma­ terial. b) 1O . . . cS? l l liJxf7 �xf7 1 2 eS liJe8 1 3 c3 liJc6 1 4 'ii'd S+ �f8 IS l:.e3 ! (the rook-lift is deadly) I S . . . g6 1 6 1H3+ �g7 1 7 'ii'f7+ �h8 1 8 i.. h6 leads to mate. c) 1 O . . . liJc6 ? ! l l liJxf7 ct;xf7 1 2 eS liJe8 1 3 'ii'd S+ ct;f8 1 4 l:.e3 ± . White again has a strong attack. •••

12 'ii'd l This is the safest move. Other retreats leave the queen more vulnerable to further attack and allow Black to develop his initiative without any great difficulty : a) 1 2 'ii'e 3 ? ! 'ii'c 7 1 3 liJf3 i..b7 1 4 liJbd2 ( 1 4 e S liJg4 + ) 1 4 . . . c4 I S e S liJdS 1 6 "fId4 i.. c s 1 7 'ii'g4 liJb4 + was given by Panczyk and Ilczuk in New in Chess Yearbook 87. b) 1 2 'ii'd 2? ! 'ii'c 7 1 3 liJf3 i..b7 14 liJc3 l:.ae8 I S eS b4 1 6 exf6 bxc3 1 7 bxc3 i.. xf6 gave Black more than enough for the pawn in Vogt­ Goldberg, East German Ch, Nordhausen 1 986. c) 12 'ii'd3 ? ! c4 (or 12 ... i..b 7, when 13 'ii'b 3+ transposes to line 'd2 1 ' below, while 1 2 . . . 'ii'c 7 1 3 liJf3 i..b7 allows White to develop more

OTHER ANTI-MARSHALL LINES

easily with 1 4 .i.gS) 1 3 'it'e2 .i.b7 1 4 lLlc3 'it'c7 I S lLlg4 b4 1 6 lLldS lLlxdS 1 7 exdS l:tae8 1 8 .i.e3 .i.xdS + Belotti-Minasian, European Ch, Ohrid 200 1 . d) 1 2 'it'c3 ? ! and then: d 1 ) 12 . . .'it'c7 13 a4 .i.b7 14 axbS lLlxe4 I S 'iVb3+ c4 ! ? 1 6 'iVxc4+ 'iVxc4 1 7 lLlxc4 axbS 1 8 l:txa8 l:txa8 1 9 lLlcd2 l:ta1 gave Black sufficient counterplay in Kamsky-Anand, Amber Blind­ fold, Nice 2009. d2) 1 2 . . . .i.b7 and now: d2 1 ) If 1 3 'it'b3+ then Black should not play 1 3 . . . c4? 14 lLlxc4 but instead 1 3 . . . dS ! with compensation. d22) 1 3 f3 c4 14 h3 ( 1 4 .i.e3 ? ! d6 I S lLlg4 lLlxg4 1 6 fxg4 'it'd7 +) 14 . . . dS ( 1 4 . . . 'it'c7 also looks good) I S exdS lLlxdS gave Black a strong initiative in Nowak-Kwiatkowski, COIT. 1 998. We now return to 12 'it'd 1 (D):

1 75

1 6 g3 gS 1 7 lLlg2 g4 gives Black a strong initia­ tive . c) 1 3 lLlf3 iLb7 1 4 iLgS ( 1 4 lLlc3 b4 trans­ poses to Line A2) 14 . . . dS ! ? (a sharper idea than 14 . . . l:.ae8 l s lLlc3 or 1 4 . . . b4 I S lLlbd2) I S exdS ( I S eS lLle4 1 6 .i.xe7 'it'xe7 1 7 lLlbd2 lLlxd2 1 8 'it'xd2 d4 with compensation) I s . . . lLlxdS 1 6 iLxe7 lLlxe7 1 7 lLlgS ( 1 7 lLlbd2 lLldS intending . . .lLlf4 looks OK for Black) 1 7 . . . 'iVf4 (after 1 7 . . . l:tf6 1 8 lLld2 l:taf8, as in Ulybin-Timo­ shenko, Cheliabinsk 1 989, White could try 1 9 lLlge4 ! ?) 1 8 'iVd2 lLlg6 1 9 'ii'xf4 lLlxf4 20 lLle4 was Timofeev-Nyback, European Ch, Plovdiv 2008. Here the simplest is 20 . . . iLxe4 2 1 l:txe4 l:tad8 with compensation; for example, 22 lLlc3 l:td2 or 22 l:te l lLle2+ 23 'ii>f l lLld4. 13 lLlxg4 14 'iVxg4 Now the play becomes rather forced. 14 dS IS 'it'hS dxe4 (D) •••

•••

B

Black now has a choice: 1 7S A I : 12 'iVc7 1 76 A2: 12 .i. h7 •••

•••

AI) 12 'ii'c7 Black attacks the eS-knight immediately. This was Leko's choice when Ivanchuk wheeled out the surprising 9 .i.xf7 +. 13 lLlg4 Other moves offer White little: a) 1 3 .i.f4? just loses material to 1 3 . . . .i.d6. b) 1 3 lLld3 ? ! .i.b7 14 f3 c4 l s lLlf4 ( l S lLlf2 .i.d6 1 6 h3 .i.h2+ 1 7 'ii> f l dS is much better for Black, Szyszka-Mroczek, corr. 1 996) I S . . . .i.d6 ! ? (or I S . . . .i.cS+ 1 6 'ii>h l l:.ae8 with compensation) •••

16 'it'dS+ 1 6 lLlc3 gives White nothing after 1 6 . . . :fS 1 7 'iVe8+ l:tfS 1 8 'iVhS l:tfS 1 9 'iVe8+ 112- 112 Smejkal-I.Zaitsev, Polanica Zdroj 1 970. Black could also try 1 6 . . . .i.fS 17 lLlxe4 'iVeS with good play for the pawn. 16 'ii>h8 17 'iVxa8 White can take the pawn instead with 1 7 'it'xe4, but 1 7 . . . l:ta7 followed by either . . . .i.fS or even . . . iLd6, . . . iLb7 and . . . l:taa8, gives Black good compensation. 17 iL h7 (D) 18 'it'a7! White will lose the queen, but he wants to inconvenience Black as much as possible. 1 8 'it'xf8+ iLxfS 1 9 lLld2 .i.d6 20 lLlfl .i.eS 2 1 c3 h6 22 .i.e3 .i.dS 23 h3 .i.c4 24 lLld2 .i.d3 gave •••

•••

UNDERSTANDING THE MARSHALL A TTA CK

1 76

32 11xf8+ (32 :ac8 .te6 is equal) 32 . . .'i¥xf8 33 11xf8+ 'iii>xf8 34 .txcS+ 'iii>f7 and the oppo­ site-coloured bishop ending is drawn. 30 ltJc1 ! ± Ivanchuk-Leko, MorelialLinares 2008. White retains his knight and Black has trouble creating counterplay.

w

A2) 12 ....tb7 (D)

Black sufficient play in Vorobiov-Novik, Sochi 2007. IS .. .l:taS 19 .tf4! Another nuisance move. 1 9 'iVxa8+ .txa8 20 .te3 'i¥eS 2 1 c3 .td6 22 g3 is similar, but White is a tempo behind the actual game. 19 'iVc6 20 'i¥xaS+ .txaS 21 .te3 'i¥f6 22 c3 .td6 23 ltJd2 'i¥e5 Black could consider 23 . . . .tc6 to stop a4. 24 g3 h6 25 a4! .tc6 The unclear 2S . . . 'ti'e8 26 axbS 'i¥xbS was suggested by Panczyk and Ilczuk, but the text­ move is not bad either. 26 axb5 axb5 27 ltJb3 .tfS 27 . . . 'iii>h 7 is unclear - Leon Hoyos. 2S 11edl .td5 29 h4 (D) 29 ltJxcS .txcS 30 .txcS is more obvious, but 30 . . . e3 3 1 .txe3 'i¥e4 32 11xdS 'ilt'xdS 33 .td4 'iii>g 8 34 na7 gS (Leon Hoyos) is not so clear. •••

B

29 ...'iii> gS? Better is 29 . . . .txb3 ! 30 11d8 'iii>g 8 3 1 l:taa8 'i¥fS ! (not 3 L. .'ti'e7? 32 11e8 'i¥d6 33 11ad8 +-)

Black attacks the e4-pawn and delays com­ mitting his queen, if only for a moment. 13 ltJc3 'i¥c7 1 3 . . . b4 1 4 ltJdS ltJxdS I S exdS is not so con­ vincing for Black, but 1 3 . . . 'i¥e8 ! ? is rather in­ !eresting. White has: a) 1 4 f4 d6 I S ltJf3 'ti'c6 ! ? puts pressure on White's e4-pawn, and 1 6 ltJdS can be met by 1 6 . . . 11ae8 or even 1 6 . . . .td8 ! ? b ) 1 4 ltJg4 has been the most popular move. l 4 . . . 'ti'g6 (also interesting is 14 . . . b4 I S ltJdS ltJxdS 1 6 exdS 'iVf7 1 7 f3 11fe8 1 8 ltJeS 'i¥xdS 1 9 'ilt'xdS+ .txdS 20 ltJxd7 11ad8 with compen­ sation) I S ltJxf6+ .txf6 1 6 ltJdS ( 1 6 f3 .txc3 1 7 bxc3 11ad8 intending . . .d S gives Black coun­ terplay) 1 6 . . . 11ae8 1 7 ltJxf6+ 11xf6 1 8 f3 dS 1 9 e S ( 1 9 exdS 11xe l + 20 'i¥xe l .txdS with com­ pensation due to the weakness of c2 and f3) 1 9 . . . 11fS 20 f4 d4 gave Black the initiative in Szczepanski-Malinin, corr. 1 992. c) 1 4 ltJdS looks critical. After 1 4 ... .td6 (or 14 . . . d6 I S ltJxe7+ 'i¥xe7 1 6 ltJg4) I S ltJxf6+ 11xf6 1 6 ltJg4 I:te6 1 7 f3 ;j; Black may have some compensation, but it is not clear that it is enough for a pawn.

OTHER ANTI-MARSHAU LINES

14 tDg4 Other moves give Black good chances: a) 14 tDf3 ? ! b4 IS tDdS tDxdS 16 exdS .i.f6 1 7 d6 'ii'c 6 1 8 c3 as 1 9 cxb4 axb4 + de la Garza-Kaidanov, Las Vegas 1 992. b) 1 4 tDd3 b4 IS .i.f4 (after IS tDdS Black wins back the pawn with l S . . . tDxdS 1 6 exdS .i.d6 1 7 g3 .i.xdS) l S . . . d6 1 6 tDa4 (not 1 6 tDdS ? tDxdS 1 7 exdS c4 -+) 1 6 . . . tDxe4 1 7 f3 (with the point that 1 7 . . . tDf6 will be met by 1 8 tDaxcS) 17 ... .i.c6 ! looks at least OK for Black. 14 tDxg4 14 . . . ':ae8 is a serious alternative for Black. After I S tDxf6+ �xf6 1 6 tDdS 'it'eS (D) White has tried:

1 77

1 8 :tad 1 d6 1 9 .i.h4 'it'f7 ! was Rantanen-Pinter, Helsinki 1 983. With pressure on f2 and a2, Black has excellent chances. b) 16 'ii'h S ':ae8 ( 1 6 . . . .i.eS ! ?) 17 .i.d2 .i.eS gives Black compensation, and after 1 8 f3 ? ! b4 1 9 tDd 1 dS 20 exdS .i.xdS Black was taking over the whole board in Morozov-Weissleder, corr. 1 995 .

•••

16 ':ae8 After 16 . . . .i.eS 1 7 tDdS .i.xdS 1 8 exdS 'it'd6 White gave back the pawn with 1 9 c3 'ii'xdS 20 .i.e3 in Stripunsky-Svirin, Smolensk 1 99 1 . This looks a little better for White because he will get pressure on the d-file. 17 .i.gS This looks a bit odd, because the bishop can­ not retreat to h4 and g3. White should probably look for improvements in other lines, but Black seems able to hold the balance: a) 1 7 'it'd 1 .i.eS 18 tDdS 'it'd6 1 9 c3 'ii'g6, with the initiative, is given by Panczyk and Ilczuk. b) 17 �h6 �eS 1 8 :'ad l d6 1 9 .i.f4 was Van Riemsdijk-Avila Jimenez, Barbera 1 999. Now Black should choose the thematic 1 9 . . . 'ii'f7 ! , with good play. c) 17 :'d l :'e6 1 8 tDdS (maybe White should just develop with 1 8 .i.e3 ! ?) 1 8 . . . 'ii'd 8 ! 1 9 .i.d2 'it'e8 20 .i.c3 was played in Zecha-Polster, corr. 1 989. Now 20 . . . 'ii'f7 is again possible, but even better is 20 . . . .i.eS ! 2 1 .i.xeS (2 1 f3 runs into 2 1 . . .hS ! , and 2 1 tDc7 backfires after 2 1 . . .'it'f7 ! ) 2 1 . . . :'xeS, winning back the pawn because 22 f3 ? is again met by 22 . . . hS ! winning the f3pawn. 17 .i.eS! 18 ':adl d6 19 tDdS 'it'f7 20 .i.f4 .i.xb2 21 l:te2 •••

a) 1 7 c4? ! bxc4 1 8 tDb6? ( 1 8 tDxf6+ 'it'xf6 1 9 f3 dS 20 exdS 'it'd4+ 2 1 Wh 1 :'xe 1 + 22 'iVxe 1 �xdS =+=) 1 8 . . . �h4 1 9 g3 'it'f6 20 f3 'it'xb6 2 1 gxh4 'iVg6+ 22 Wh 1 :'xf3 ! 23 'ii'xf3 :'xe4 -+ Carton-Crawley, London 1 986. b) 17 �f4?! 'it'xb2 1 8 tDxf6+ 'ii'xf6 19 .i.g3 :'xe4 ( 1 9 . . ..i.xe4 also looks good) 20 'ii'xd7 .i.c6 =+= Kopp-Stock, corr. 1 994. c) 17 tDxf6+ 'it'xf6 1 8 f3 dS 19 exdS ':xe 1 + (maybe Black should try 1 9 . . . 'iVd4+ ! ? 2 0 'iVxd4 l:lxe 1 + 2 1 ..ti>f2 ':xc l ! 22 'ii'e S ':xa1 23 'ii'e6+ :f7 24 d6 .l:.d 1 2S d7 h6 26 'ii'e 8+ ':f8 27 ..ti>e2, with unclear play) 20 'it'xe 1 .i.xdS 2 1 c3 'it'g6 22 'ii'g 3 'ii'x g3 23 hxg3 :e8 24 ..ti>f2 ;!; Navarro Segura-Chudnovsky, Las Vegas 1 994. The end­ ing is unpleasant but Black should draw. IS 'ii'xg4 .i.d6 16 g3 (D) This weakens the light squares but the alter­ natives are not convincing: a) 1 6 'it'h3 :'ae8 (or 1 6 . . . b4 1 7 tDdS .i.xdS 1 8 exdS .i.eS with compensation) 1 7 .i.gS .i.eS

•••

UNDERSTANDING THE MARSHAll A TTACK

1 78

Sivanandan-Negi, Indian Ch, Chennai 2008. Here 2 1 . . .:e6 is given by Panczyk and Ilczuk, while 2 l . . .hS and 2 l . . .�d4 ! also look good for Black.

B) 9 lDxd4 This move is less risky for White. 9 exd4 10 e5 Not, of course, 10 'ii'xd4? cS and 1 1 . . .c4, trapping the white bishop. IO lDe8 (D) •••

•••

w

Now White can sacrifice a pawn to acceler­ ate his development or simply recapture on d4. 178 B l : 1 1 'it'xd4 1 80 B 2 : 1 1 c3

B1) 1 1 'it'xd4 White refrains from sacrificing a pawn and brings his queen into play. Black can gain some time chasing the queen, but he needs to be careful because White still holds some ini­ tiative. 1 1 �b7 Black develops a piece and threatens . . . cS­ c4. There is a sharp alternative in the immediate 1 1 . . .cS . This allows White to attack the rook in the comer with 12 'it'e4, but Black hopes to re­ gain the time by pestering White's queen. Black has : a) 1 2 . . . lDc7 is held to be inferior but looks playable to me. 1 3 c4 bxc4 14 �c2 ( 1 4 �xc4 dS IS exd6 �xd6 is about equal; Black's active •••

pieces compensate for his slightly worse pawn­ structure) 14 . . . g6 and now: a l ) I S �h6 :e8 1 6 lDc3 :b8 1 7 'it'xc4 :b4 1 8 'it'd3 :h4 gave Black active play in Brooks­ Kaidanov, New York 1 990. a2) IS lDc3 l:tb8 ( l S . . . dS 16 exd6 �xd6 is also reasonable) 16 �a4 :b6 17 'it'xc4 ( 1 7 l:td 1 may be a better try, but 1 7 . . . .i.b7 1 8 'it'xc4 d5 1 9 exd6 .i.xd6 i s not clear) 1 7 . . .d S 1 8 exd6 .i.xd6 1 9 a3 �e6 20 'it'e2 fS ! gave Black counterplay in Van der Wiel-Ki .Georgiev, Wijk aan Zee 1 988. b) 12 . . . :b8 13 c4 ( 1 3 c3 �b7 intending . . . d6 is fine for Black, while 1 3 �dS lDc7 1 4 �f4 lDxdS I S 'it'xdS :b6 1 6 lDc3 �b7 1 7 'it'd2 :g6 1 8 lDdS was about even in Lepelletier­ Foukakis, Khania 1 994) 1 3 . . . �b7 14 'it'e2 and here: b l ) 14 . . . dS I S cxdS c4 1 6 d6 cxb3 1 7 dxe7 'it'xe7 1 8 axb3 lDc7 1 9 lDc3 ;!; Liu Dede-Djuric, Jakarta 200 1 . b2) 1 4 . . . b4 denies White the c3-square for the knight and keeps the a2-g8 diagonal closed. I S lDd2 lDc7 16 lDe4 lDe6 17 'ili'd3 fS ! 1 8 exf6 �xf6 1 9 lDxf6+ 'it'xf6 gave Black good coun­ terplay in Kholmov-Smyslov, Sochi 1 974. 12 c4 White needs to play actively in the centre to fight for the initiative. 1 2 c3 ? ! d6 1 3 �f4 dxeS 14 'it'xeS �d6 I S 'it'gS h6 16 'it'g3 'it'f6 1 7 �xd6 lDxd6 was already a little preferable for Black in Navarro Segura-Peng Xiaomin, Mos­ cow Olympiad 1 994. Therefore the main alternative for White is 12 'it'g4. Now Panczyk and Ilczuk's suggestion of 1 2 . . .f6 is interesting, but illegal, so we shall look at 1 2 . . . cS instead. 1 3 c3 d6 ( 1 3 . . . dS just gives White extra options such as 1 4 .i.c2 and 1 4 lDd2) 14 exd6 ( 1 4 e6 dS ! and 14 �f4 dxeS I S �xeS lDf6 are not promising for White) 14 . . . �xd6 (or 14 . . . lDxd6 I S �h6 �f6 1 6 �f4, which was given as ';!;' by Panczyk and Ilczuk but Black looks fine here) I S �gS and then: a) Is . . . lDf6 1 6 'it'h4 h6 1 7 �xf6 ( 1 7 �xh6 gxh6 1 8 'it'xh6 leads nowhere - at the very least Black has 1 8 . . . lDe4, when White must take a draw with 1 9 'it'g6+ 'iii>h 8 20 'it'h6+) 1 7 . . . 'it'xf6 1 8 'it'xf6 gxf6 1 9 lDd2 :fe8 20 �c2 :ad8 2 1 lDe4 1/2- 1/2 Svidler-Naiditsch, Dortmund 200S . b) I S . . . 'it'c7 ! ? 1 6 h3 'it'c6 1 7 lDd2 c4 1 8 �c2 fS 1 9 'it'f3 'it'd7 20 'it'd 1 lDc7 gave Black an

OTHER ANTI-MARSHALL LINES

active position in Kovanova-Shadrina, Russian Women' s Ch, Oriol 2006. We now return to 1 2 c4 (D):

Although Black's position remains very solid, White has a better pawn-structure and a slight initiative. IS li:)bS Black can also play I S . . . �f6. After 1 6 li:)c3 Black has : a) 1 6 . . . :tbS 1 7 �f4 �c6 I S l:tad 1 l:tb4 1 9 'ii'g 3 .l:.d4 20 �eS l:txd l 2 1 l:txd l l:teS 22 f4 'ii'e7 23 li:)dS �xdS 24 .l:.xdS ;l;; Nijboer-I.Soko­ lov, Dutch Ch, Amsterdam 1 996. b) 16 . . . :teS 17 �f4 J:.bS I S h3 �c6 19 'ii'hS 'it'd7 20 'it'aS �bS 2 1 li:)xbS axbS 22 'it'd2 ;l;; Arakhamia-Hebden, British Ch, Scarborough 2004. c) 1 6 . . . li:)bS is Black's most common idea in this position. White then has: c 1 ) 1 7 li:)xbS axbS I S �f4 cS (safer than grabbing the pawn with I S . . . �xb2 1 9 l:tad l , when White has the initiative) 1 9 l:tad l 'ii'c s 20 'it'xcs :taxcS 21 �d6 :tfdS 22 �xcS llxd 1 23 �xd l �xb2 24 �b3 g6 = J.Polgar-Nunn, Hastings 1 992/3 . c2) 1 7 li:)e4 is a better try if White wants to squeeze something out of the opening. 1 7 ... �xe4 ( 1 7 . . . li:)d4 I S li:)xf6+ 'it'xf6 1 9 �gS 'it'g6 20 'it'xd4 'it'xgS 2 1 g3 perhaps gives White a tiny edge because of his better pawn-structure) I S 'ii'xe4 �d4 (after I S . . .l:teS 1 9 'it'xeS+ 'it'xeS 20 l:txeS+ l:txeS 2 1 ..ti>n li:)d4 22 �d l White will untangle and the bishops will ensure long-term pressure) 1 9 'ii'f4 cS 20 �e3 (20 l:tbl may be even better) 20 . . . �xe3 (Black should at least try 20 . . . �xb2) 2 1 fxe3 ! ;l;; Balogh-Jenni, Euro­ pean Ch, Dresden 2007 . White has the better minor piece because he controls d4 and there is pressure against f7. 16 li:)c3 li:)xc3 1 6 . . . �f6 is common, and transposes to line 'c' in the previous note. 17 bxc3 �d6 18 �f4 'iVf6 19 �xd6 cxd6 (D) Despite the simplifications White has a nag­ ging edge because he can activate his major pieces quickly. Black should hold, but the posi­ tion is unpleasant. 20 l:.adl 20 'it'd7 l:.abS 2 1 :te3 �cS 22 'ii'a4 was also a little better for White in Dorfman-Tkachev, French Ch, Aix-Ies-Bains 2007 . 20 :tad8 21 'ii'b4 White probes the queenside. After 21 l:te3 l:tfeS 22 l:tde 1 l:txe3 23 l:txe3 g6 24 h4 hS 2S •••

B

12 bxc4 1 2 . . . d6? ! 1 3 li:)c3 dxeS 14 'ii'x dS �xdS I S cxbS axbS 1 6 li:)xbS is uncomfortable for Black because he is cramped and the eS-pawn is weak, but 1 2 . . . cS ! ? is a reasonable alternative. Black will follow up with . . . bxc4 and . . . dS . White has : a) 1 3 'ii'd l bxc4 14 �xc4 dS I S exd6 li:)xd6 gives Black a big lead in development. b) 13 'ii'd3 bxc4 14 'it'xc4 d5 I S exd6 li:)xd6 ! ? (better than I S . . :ilkxd6 1 6 li:)c3 ;l;; ) 1 6 'it'xcs li:)e4 gives Black the initiative in return for a pawn. c) 13 'it'g4 is the most active try, but Black looks OK here too: 1 3 . . . d6 (Black can also play 1 3 . . . bxc4 14 �xc4 dS , but the text-move looks even simpler) 1 4 li:)c3 b4 I S li:)dS dxeS 1 6 l:txeS �d6 intending . . .li:)f6 has been played a couple of times. Black looks fine here. 13 'ii'xc4 dS 14 exd6 li:)xd6 1S 'it'g4 (D) •••

B

1 79

•••

UNDERSTANDING THE MARSHALL A TTACK

180

For the pawn White has very easy develop­ ment and the e5-pawn gives him some extra space which he may be able to use to generate an attack on the kingside. Black must also be careful because his a8-rook can be vulnerable and White has the d5-square at his disposal. 12 d6 This is the most solid response. Black wants to eliminate the annoying e5-pawn and bring his knight back into play. There are other moves too. 1 2 . . . c6 ! ? and 1 2 . . . c5 ! ? have been rarely played but are worth considering. The main alternative is the natural 1 2 . . . .i.b7 (D), developing a piece and safeguarding the a8-rook. White has: •••

'ifb4 l:[d7 Black managed to defend in Carl­ sen-Leko, Dortmund 2007. 21.. .i.a8 22 :e3 g6? ! After this Black will lose a pawn for nothing. 22 . . . 'ii'g 5 23 .l:.g3 'ii'f6 24 'ii'b6 .l:.fe8 at least al­ lows Black to take over the e-file. 23 'ifb6 'ii'g5 24 .l:.g3 'ii'b5 25 l:txd6 l:xd6 26 'ii'xd6 ± Ivanchuk-Aronian, MorelialLinares 2008 . •

82) 1 1 c3 (D)

White offers a pawn to develop quickly and bring maximum control to the centre. 1 1 ...dxc3 Black should certainly accept the challenge. l l . . .d6 1 2 cxd4 (or even 1 2 'iff3) 1 2 . . . dxe5 1 3 'ii'f3 i.e6 1 4 dxe5 ;!; i s pleasant for White, while l l . . .d5 12 cxd4 c6 1 3 i.c2 ;!; was Short-Beli­ avsky, Warsaw rapid 2004. 12 tt'lxc3

a) 13 'ii'd 3 c5 (this is a common idea: Black creates a backward - though extra - d-pawn but gets some activity of his own) 14 i.d5 i.xd5 1 5 tt'lxd5 tt'lc7 1 6 tt'lb6 ':'a7 1 7 i.e3 d 5 1 8 exd6 tt'le6 1 9 ':'ad l 'ii'xb6 20 dxe7 ':'xe7 =+= Roman­ ishin-Tseshkovsky, Ordzhonikidze 1 978. b) 13 i.c2 d6 1 4 'ii'd 3 g6 1 5 i.h6 tt'lg7 was solid enough for Black in Woldmo-Olofson, COTTo 1 996. c) 1 3 i.f4 c5 14 'ii'g4 c4 15 .i.c2 g6 ( 1 5 . . . d5 is also possible; for example, 16 exd6 .i.xd6 1 7 :tad l 'ii'b 6) 1 6 .l:.ad l f5 1 7 'ii'h 3 was DujkovicBrenjo, Herceg Novi 2006. Now 1 7 . . . 'ii'c 8 ! ? avoids White's e6 advance and prepares . . .tt'lc7 . d) 1 3 'ii'h 5 d5 1 4 exd6 .i.xd6 ( 1 4 . . . 'it'xd6 ! ?) 1 5 i.g5 tt'lf6 1 6 'it'h4 h6 1 7 .i.xh6 ! gxh6 1 8 'ii'xh6 tt'lh7 1 9 i.c2 f5 20 .i.xf5 ':'xf5 2 1 'it'g6+ �h8 22 'ii'xf5 was unclear in Rigo-Pinter, Bu­ dapest 1 978. e) 13 tt'ld5 .i.xd5 1 4 'ii'xd5 c6 1 5 'ii'f3 ( 1 5 'ii'e4 looks more flexible) 1 5 . . .i.c5 1 6 i.f4 tt'lc7 1 7 'ii'g4? ! (this looks too slow, so White should

OTHER ANTI-MARSHALL LINES

consider 1 7 .:tad 1 or even 1 7 .:tac 1 ) 1 7 . . . ltJe6 1 8 i.h6 d5 1 9 exd6 'ili'f6 ! 20 i.e3 i.xd6 2 1 .:tac 1 c 5 was Oechslein-Schreiber, COIT. 1 988. B lack has solved his development problems and still has an extra pawn. We return to 1 2 . . . d6 (D) :

Now White has a choice: 181 B21 : 13 'ili'f3 1 82 B22: 1 3 i.d5 Other moves are rare, but not necessarily harmless: a) 13 i.f4 dxe5 14 'iWf3 (after either 14 'iWxd8 i.xd8 15 i.xe5 or 1 4 i.xe5 'ii'xd 1 1 5 ':'axd 1 i.e6 1 6 i.xe6 fxe6 White has some compensation for the pawn but no more than that) 14 . . . i.g4 (similar and perhaps even stronger is 14 . . . i.b7 ! 1 5 'ili'xb7 ltJd6 ! { not 1 5 . . . exf4? 1 6 .:txe7 ! } 1 6 'i1i'f3 exf4) 1 5 'i1i'xg4 exf4 1 6 'i1i'xf4 i.d6 1 7 'ili'f3 ltJf6 was level in Klundt-Howell, Nuremberg 1 987. b) 13 ltJd5 dxe5 ! ? ( 1 3 ... i.e6 1 4 'iWf3 trans­ poses to Line B2 1 ) and here: b 1 ) 1 4 'iWf3 i.d6 ! 1 5 ltJe7+ 'ili'xe7 1 6 'ili'xa8 ltJf6 gives Black excellent compensation for the exchange. b2) 14 ':'xe5 i.f6 15 ltJxf6+ 'ili'xf6 ! (not 15 . . .ltJxf6? 16 i.xf7+ ! ) looks safe enough for Black.

821) 13 'ili'f3 (D) White gains more time by hitting the a8rook. 13 i.e6 14 ltJd5 ..•

181

B

White can also try to keep up the pressure with 14 i.f4. After 14 . . . i.xb3 ( 1 4 . . . dxe5 1 5 i.xe5 i.xb3 transposes) 1 5 axb3 dxe5 1 6 i.xe5, the move 1 6 . . . i.d6 has been played a couple of times, but Black is uncomfortable: 1 7 ltJxb5 ! i.xe5 1 8 ':'xe5 ltJd6 1 9 ltJd4 ltJb5 20 ltJc6 (20 ltJxb5 axb5 2 1 ':'d 1 is even better according to Marin) 20 . . . ltJd4 2 1 ltJxd4 (better than 2 1 'ili'e4 ltJxc6 22 'ili'xc6 'iWd6 23 'iWc3 .:tfe8 = Balogh­ Harikrishna, George Marx Memorial, Paks 2007) 2 l . . .'iIi'xd4 22 'ili'c3 gave White pressure in Yakovenko-Kariakin, Foros 2007. 16 . . . ltJf6 may be better. After 1 7 .:tad 1 i.d6 1 8 i.xd6 cxd6 19 'ili'c6 White will win back the pawn but Black should be able to neutralize White 's slight initiative in this simplified position. 14 :tc8 A sharper idea is 14 . . . i.xd5 ! ? 1 5 i.xd5 .:tb8 1 6 i.e3 c5 1 7 .:tad 1 'fic7 1 8 e6 fxe6 1 9 i.xe6+ 'ifi>h8 20 'ii'h 3 g6 with unclear play, Timofeev­ Sutovsky, Montreal 2006, although this is not without risks. 15 i.f4 (D) •••

182

UNDERSTANDING THE MARSHALL A TTACK

15 dxe5 White is very active so Black must be careful. The text-move looks to dissipate the tension in the centre as quickly as possible. IS . . . 1i.g5 ? ! 1 6 l:.ad l 'ith8 1 7 1i.c2 ! gave White a strong initia­ tive in I.Gurevich-Nunn, Hastings 1 992/3 . 16 ttJxe7+ White can also play 1 6 1i.xeS 1i.d6 1 7 l:.ad 1 1i.xeS 1 8 l:txeS 1i.xdS 1 9 l:texdS ( 1 9 1i.xdS was played in Morales-Perez, Havana 1 999 and Black's safest move is 1 9 . . . �f6) 1 9 . . . ttJd6, as in Henrichs-Gustafsson, German Ch, Bad Konigs­ hofen 2007. White has some compensation for the pawn because his pieces are so active but with such simplifications Black's position is pretty safe. 16 :iVxe7 17 1i.xe5 ttJd6 18 1i.xe6 White also got nowhere after 1 8 �c3 'iVgS 1 9 1i.xe6 fxe6 20 b3 'iVg6 2 1 'iVc6 ttJf7 22 1i.g3 eS in Leko-Aronian, FIDE Grand Prix, Jermuk 2009. 18 fxe6 19 �c6 'iWe8 20 l:tac1 ttJc4 21 'iVe4 "Viif7 1/2_ 1/2 Adams-Aronian, Wijk aan Zee 2008. White has enough pressure for the pawn but nothing more. One possible continuation is 22 l:te2 'iVfS 23 1i.c3 'iVxe4 24 l:txe4 l:.fe8 2S l:tce l 'itf7 26 l:tf4+ 'itg8 27 l:tfe4 with a repetition of moves. .••

1 3 . . . 1i.e6?! 1 4 1i.xa8 'iVxa8 1ooks insufficient after I S 1i.f4 dxeS 1 6 1i.xeS ttJd6 1 7 ttJe2 ttJfS 1 8 ttJd4. 14 1i.e3 1i.e6 1 4 . . . cS ? ! is insufficient: I S exd6 1i.xd6 (after I S . . . 'iVxd6 1 6 ttJe4 White also has the initiative) 1 6 ttJe4 (or even 1 6 1i.xcS ! ? 1i.xh2+ 1 7 'itn 1i.d6 1 8 ttJe4 1i.e7 1 9 1i.a7) 16 . . .�c7 17 ttJxd6 ( 1 7 llc 1 also looks good) 17 . . . 'iVxd6 ( 1 7 . . . ttJxd6 1 8 1i.f4 l:tb6 is awkward, but may be a better try), S .Michailow-Muuss, Osterronfeld 1 996. Here Panczyk and Ilczuk give 1 8 l:tc 1 ± . 15 1i.a7 1i.xd5 16 1i.xb8 1i.b7 17 1i.a7 �a8 (D)

••

w

•••

822) 13 1i.d5 (D)

This move attacks a7 and g2 and looks like it gives Black good chances, but matters are not so simple. 18 exd6 1i.xd6 19 1i.d4 b4 20 ttJa4 1i.xg2 (D)

w

This is the latest wrinkle in the 8 d4 variation. 13 l:tb8 .•.

White's a4-knight is offside and his kingside is looking a bit porous. However, White has a neat resource.

OTHER ANTI-MARSHALL LINES

21 l:Ixe8! 'it'xe8 Not 2 1 . . .lhe8? 22 'it'g4. 22 �xg2 c5 22 . . . 'ii'e4+ 23 f3 'ii'f4 can be answered by 24 .i.g l . 23 .i.e3 'it'c6+ 24 f3 :e8 25 .i.f2 (D)

183

B

B

Black has two main lines here, and although Line B is considered the main continuation, Line A looks satisfactory to me as well. A : 9 dxe4 1 83 1 84 B: 9 exd4 •••

•.•

Now: a) 2S . . . c4? ! 26 li:)b6 c3 27 bxc3 bxc3 28 'ifdS 'ifc7 29 li:)c4 .i.b4 30 .l:.c 1 .l:.e2 3 1 .l:.xc3 h6 32 .l:.e3 .l:.xa2 33 .l:.e8+ �h7 34 li:)e3 .l:.xf2+ 3S �xf2 .i.c3 36 �g2 1 -0 L.Dominguez-Aronian, Wijk aan Zee 2009. b) In New in Chess, Boel gives 2S . . . .l:.eS 26 b3 .l:.gS+ 27 .i.g3 (27 �h l .l:.fS) 27 . . . fS ! 28 �f1 f4 29 .i.f2 'it'd7 ! with excellent play. After 30 li:)xcs 'it'bS+ 3 1 li:)d3 l:.dS 32 �e2 .i.eS 33 llc 1 .i.c3 White is so tied up that Black can hardly be worse.

Section 1 0 .4

9 . . . li:)xe4 is also playable. 1 0 dxeS ( 1 0 li:)xeS li:)xeS 1 1 dxeS c6 =) 1O . . . .i.e6 is an Open Ruy Lopez that would come about from the move­ order 1 e4 eS 2 li:)f3 li:)c6 3 .i.bS a6 4 .i.a4 li:)f6 S 0-0 li:)xe4 6 d4 bS 7 .i.b3 dS 8 dxeS .i.e6 9 c3 .i.e7 1 0 l:.e l 0-0. This is outside the scope of this work, but this line is rare enough that Black could consider it without taking on any great additional theoretical burden. After 1 1 li:)d4, 1 1 . . . li:)xd4 1 2 cxd4 is a little better for White, while the speculative 1 1 . . . li:)xeS 1 2 f3 .i.d6 (or 1 2 . . . .i.h4) is probably dubious, but 1 1 . . .'it'e8, as once played by Svidler, is interesting.

A) 9 dxe4 This line is considered inferior, but I have not seen a convincing reason for it to be dis­ carded. 10 li:)xe5 (D) 10 .i.b7 This is the main line, but 1O . . . li:)xeS looks quite playable to me. At first it seems that Black is just losing a pawn but matters are not so sim­ ple because Black will develop very quickly. 1 1 dxeS 'it'xd l 1 2 .i.xd l li:)d7 and now: a) 1 3 .l:.xe4 li:)cs 1 4 .l:.d4 ( 1 4 .l:.e2 .l:.d8 I S .i.c2 .i.fS ! 1 6 li:)a3 li:)d3 with compensation) 14 . . . l:Ie8 ! ? (Atalik also gives 14 . . . li:)e6 I S %:td3 li:)cs 16 .l:.d4 =) I S i.f3 i.b7 16 .i.xb7 li:)xb7 1 7 •••

1 e4 e 5 2 li:)f3 li:)c6 3 .i.b5 a 6 4 .i.a4 li:)f6 5 0-0 .i.e7 6 l:Ie1 b5 7 .i.b3 0-0 8 c3 d5 9 d4 (D) Instead of grabbing a pawn, White heightens the tension in the centre. Although this move is not very dangerous for Black, it should not be ignored. Often play resembles an Open Ruy Lopez or Yates Variation, although here Black has a pretty good version of the structures that arise from those lines. 9 d4 is played mostly as a surprise weapon. In a sense White has called Black's bluff (7 . . . 0-0) with 8 c3, but then when Black indeed goes for the Marshall with 8 . . . dS White shows that he was really the one that was bluffing . . .

•••

UNDERSTANDING THE MARSHALL A TTA CK

184

This was Khalifman-Adams, Wijk aan Zee 2002. Here Adams gives 1 7 l::td7 ! lLlc5 1 8 l::txc7 lLlxb3 1 9 axb3 �d5 20 b4 l::t ae8 2 1 �f4 ±. True enough, but Black can improve on the 1 4th move, and 1 O . . . lLlxe5 looks OK too.

B) 9 exd4 (D) •••

.i.f4 g5 ! ? 1 8 �g3 f5 1 9 exf6 �xf6 20 l::td l l::te 2 2 1 a4 l::tae8 is equal, Michielsen-Atalik, Ottawa 2007. b) 13 .i.c2 and here: b l ) 1 3 . . . .i.b7 14 .i.xe4 ( 1 4 lLld2 lLlxe5 1 5 lLlxe4 :fe8 1 6 b 3 ':ad8 1 7 �f4 f6 was harmless for Black in Gallagher-Vijayalakshmi, British Ch, Edinburgh 2003) 14 . . . �xe4 15 ':xe4 :fd8 ( l 5 . . . l:.fe8 ! ?) 1 6 lLld2 f6 1 7 exf6 Jtxf6 1 8 �f1 lLlc5 19 l::te2 a5 20 lLlf3 l::td l + 2 1 :e l ':xe l + 22 lLlxe l b4 with compensation, Kriakvin-Sargis­ ian, Saratov 2006. b2) 1 3 . . . lLlxe5 1 4 �xe4 l::tb 8 1 5 �f4 ( 1 5 .i.xh7+ �xh7 1 6 l::txe5 l::td 8 with compensa­ tion) 1 5 . . . Jtd6 1 6 Jtxe5 (or 16 Jtg3 lLlc4 17 b3 �xg3 1 8 hxg3 lLld6 19 lLld2 Jte6 = Moroze­ vich-Leko, Erevan rapid 2008) 1 6 . . . .i.xe5 1 7 Jtxh7+ �xh7 1 8 ':xe5 l::td 8 1 9 �f1 f6 20 l::te 3 l::td l + 21 ':e l ':'xe l + 22 �xe l Jtf5 23 lLla3 :e8+ 24 �d2 l:.d8+ 25 �e3 ':e8+ 26 �d2 ':d8+ '/2- '/2 Morozevich-Shirov, Tal Memorial, Moscow 2006. 1 1 lLld2 Jtd6 l l . . .lLlxe5 1 2 dxe5 lLld7 1 3 lLlxe4 lLlxe5 1 4 Jtf4 ! gave White the initiative i n Palac-Beliav­ sky, Slovenian Team Ch, Celje 2003. Black feels uncomfortable on the e-file and the c7pawn is under pressure too. 12 f4 exf3 13 lLldxf3 lLlxeS 14 lLlxeS Leading nowhere is 14 dxe5 .i.c5+ 1 5 Jte3 .i.xe3+ 1 6 l:.xe3 lLle4 = . 14 .i.xeS?! This just looks wrong. 1 4 ... lLld7 is solid, and 14 . . . c5 also looks satisfactory. 15 dxeS 1 5 l::txe5 ;!; is good too. IS Ji'xdI 16 l::tx dl lLle4 •••

••

This is the main line and Black has scored pretty well here. 10 eS 1 0 exd5 lLlxd5 is considered in Section 1 0.5 under the move-order 9 exd5 lLlxd5 1 0 d4 exd4. lo lLle4 11 cxd4 After 1 1 lLlxd4 Black should be bold and play l l . . .lLlxe5 ! 1 2 f3 ( 1 2 Jtf4 .i.f6 1 3 .i.xe5 ? ! { better i s 1 3 f3 c5, which i s line 'a' below } 1 3 . . . Jtxe5 1 4 lLlf3 Jtg4 ! 1 5 'ii'xd5 Jtxf3 1 6 'ilixe5 'ilih4 1 7 ':xe4 'ii'xe4 1 8 'ii'xe4 Jtxe4 1 9 lLld2 Jtb7 i s winning for Black, Belov-Akopian, Aeroflot Open, Moscow 2007) 1 2 . . . c5 ! (D). •••

w

OTHER ANTI-MARSHALL LINES

185

Now 1 3 0.e2? and 1 3 0.c2? both lose to 1 3 . . . c4, so White has : a) 1 3 �f4 �f6 ! 1 4 �xe5 �xe5 1 5 0.c6 �xh2+ 16 �xh2 'ii'h4+ 1 7 �gl 'ii'f2+ 1 8 �h2 'ii'g 3+ 1 9 �gl 0.g5 ! 20 �h l (20 �fl �h3 ! ) 2 0. . . 'ii'h4+ 2 1 'iii>g l 0.h3+ 2 2 gxh3 �xh3 was Kruppa-Vladimirov, Frunze 1 988. Now even White's most tenacious defence, 23 'ii'd2, leaves Black with the initiative after 23 . . . 'ii'g 3+ 24 'iii>h l �g2+ 25 'ii'x g2 'ii'xe l + 26 'ii'g l :ae8 27 0.a3 'ifd2 ! . b ) 1 3 fxe4 cxd4 1 4 exd5 (worse are 1 4 cxd4 ? ! �g4 1 5 'ifd2 0.c4 + and 1 4 �xd5 ? ! dxc3 ! with the point 1 5 �xa8 �c5+ 1 6 'iii>h l 0.d3 -+) 1 4 . . . �d6 1 5 �f4 l:te8 1 6 cxd4 0.d3 (Black could try 1 6 . . . 0.f3+ ! ? 17 gxf3 �xf4, with the initiative) 1 7 l:txe8+ 'ii'x e8 1 8 'ii'x d3 'ii'e 1 + 19 'ii'f l 'ii'x fl + 20 �xfl �xf4 2 1 0.c3 �xh2 1/2- 1/2 Ravi-Lukacs, Kolhapur 1 987. We now return to 1 1 cxd4 (D) :

( 1 4 . . . �b4? 1 5 'ii'c 2 +-) 1 5 'ii'xf3 White has ex­ cellent compensation because of the pressure on the light squares: d5, c6, f7 and h7 are all weak. b) 12 . . .�f5 1 3 a3 �e7 ? ! (Black is just losing tempi, so 1 3 . . . �xd2 has to be tried) 14 �c2 f6 1 5 0.xe4 dxe4 1 6 0.h4 0.xd4 1 7 0.xf5 0.xf5 1 8 'ii'g4 is much better for White, E.Berg-Hebden, Gibraltar 2008. c) 1 2 . . . �xd2 1 3 �xd2 �g4 1 4 �e3 0.a5 1 5 �c2 f5 1 6 �d3 l:tb8 1 7 h3 �h5 1 8 �e2 is a little better for White, because Black lacks counterplay. Matters were made worse after 1 8 . . . g5? 1 9 �xg5 ! 0.xg5 20 0.xg5 �xe2 2 1 'ii'd 2 ! 0.c4 2 2 'ii'f4 ± i n E.Berg-Olszewski, Najdorf Memorial, Warsaw 2008.

B

w

The position looks a lot like an Open Ruy Lopez, but here White has managed to play cxd4. This gives White a half-open c-file, but there is less pressure against the d5-pawn so Black can develop his queen's bishop actively. Black has two popular moves: 1 85 B l : 1 1 �g4 1 86 B2: 1 1 �f5

This is very natural. Black pins the f3-knight and creates pressure on d4. However, it turns out that the threat to double White's pawns is not much of a threat after all. 12 0.c3 White correctly forces the pace. After 1 2 �e3 Black can play 1 2 . . . 0.aS or consider 1 2 . . . �b4 again, because �e3xd2 will cost White a tempo. 12 �xf3 Black can also play 1 2 . . . 0.xc3 1 3 bxc3 'ii'd7 but if he wants this structure it is better to play 1 1 . . .�f5 because the bishop has no reason to be on g4 - White can play �c2 and h3 may be useful too. As before, Black should avoid 12 . . . �b4?! 1 3 0.xe4 ! with a strong initiative. 13 gxf3 0.xc3

•••

...

B1) 1 1 �g4 (D) •••

•..

Black can also play the odd 1 1 . ..�b4 1 2 0.bd2 ( 1 2 :e3 0.a5 1 3 �c2 f5 1 4 :r.e2 �e6 1 5 0.bd2 �e7 1 6 0.fl 0.c4 gave Black good play in Nijboer-Onishchuk, Wijk aan Zee 2005), and here: a) 1 2 . . . �g4?! 1 3 0.xe4 ! is an idea worth re­ membering. After 1 3 . . . �xe l 14 0.eg5 �xf3

UNDERSTANDING THE MARSHALL A TTA CK

186

1 3 . . . lDxf2 ? ! 1 4 �xf2 .i.h4+ looks fun, but Black should not trade two good minor pieces for a rook in this position. 14 bxc3 (D)

15 hl lDas 16 :gl White is also a little better after 1 6 .i.c2 'iith 8 1 7 J:.g 1 'ild7 1 8 .i.gS .i.xgS 1 9 ':'xgS lDc6 20 fi'c 1 fi'e6 2 1 f4 lDe7 22 a4 c6 23 .i.d3 :ac8 24 fi'e3 ;I; Ivanchuk-Yakovenko, Tal Memo­ rial, Moscow 2007 . 16 'ild7 17 'ile2 'ii'e6 18 .Jt gS c6 If 1 8 . . . lDxb3, then 1 9 axb3 intending b4 gives White an edge, so Black protects his bS-pawn. 19 .i.c2! Black was hoping for 1 9 f4 lDxb3 20 axb3, when 20 ... aS prevents b4 and gives Black the possibility of creating counterplay with . . . a4 or . . . b4. 19 ':'a7 Black wants to guard the second rank, but his pieces are not so well coordinated. 20 f4 �h8 21 fi'hS .i.a3? This is a shot in the dark. 22 :g3 White had a strong attack in Shirov-Bacrot, Bundesliga 2003/4. •••

B

•••

White has strengthened his centre on both fronts. Black's position remains very solid, but White is able to maintain some initiative. 14 fS (D) 1 4 . . . fi'd7 is also possible, but Black should be careful about delaying . . . fS for too long, else White may find it appealing to take en passant. A cautionary tale is I S �h l ! ( I S f4 lDaS 1 6 .i.c2 fS) l s . . . lDd8 ? ! ( 1 S . . . fS i s better) 1 6 f4 fS ? ! (suddenly i t i s too late, s o Black should prefer 1 6 . . . c6 1 7 .i.c2 fS 1 8 exf6 ':'xf6 1 9 fi'd3 ;1;) 1 7 exf6 .i.xf6 ( 1 7 . . .':'xf6 loses to 1 8 ':'xe7 'ilxe7 1 9 .i.xdS+) 1 8 fi'hS c6 1 9 .Jtc2 g6 20 ':'gl ! (in­ tending fS) 20 . . . .Jtg7 2 1 .l:[xg6 ! (this gives White a winning attack) 2 1 . . .hxg6 22 .Jtxg6 :f6 23 .i.h7+ f8 24 .Jta3+ :d6 2S fi'fS + ! �e7 26 .:le I + lDe6 27 fi'g6 +- Arizmendi-Nemet, Biel 2006. •••

w

82) 1 l .i.fS (D) •••

w

Instead of pinning the f3-knight, Black just develops to an active square and strengthens the e4-knight. 12 lDc3 White attacks the dS-pawn to force Black's response. After other moves another point of Black's play is revealed, namely his annoying pressure on the bl -h7 diagonal: a) 1 2 lDbd2 lDb4 1 3 lDn cS ! . This is a typi­ cal idea to fight for central control. Black had

OTHER ANTI-MARSHALL LINES

good play in Stellwagen-Vescovi, Wijk aan Zee 2006. b) 1 2 h3 'it'd7 1 3 lL'lbd2 (or 1 3 a3 a5 14 iLd2 lL'lxd2 1 5 lL'lbxd2 a4 1 6 iLc2 lL'la5 with equality, Sulskis-Malaniuk, Grodzisk Mazowiecki 2(07) 1 3 . . . lL'lb4 1 4 lL'lfl c5 illustrated the same idea in Sulskis-Brunello, European Union Ch, Arvier 2007 . 12 lL'lxc3 13 bxc3 'it'd7 (D) •••

187

15 'it'e2 1 5 iLc l offers a repetition, while Black is certainly fine after 1 5 lL'lh4 iLb2 1 6 lL'lxf5 'ilixf5 1 7 iLxd5 'it'xg5 1 8 iLxc6 :ad8 (for example, 1 9 'ilic2 iLxal 20 :xal with compensation) or Anand's 1 5 iLc2 iLb2 1 6 'it'b l iLxc2 1 7 'it'xb2 iLd3 . 15 h6 16 iLf4 After 1 6 iLc 1 Black can avoid the exchange of bishops with 1 6 . . . iLe7. 16 lL'la5 17 iLc2 lL'lc4 18 .:tad 1 iLe7 19 h3 a5 20 lL'lh2 (D) •••

•••

B

Black has a healthy position. Typical ideas are . . . lL'la5-c4 and . . . a5, which prepares both . . . b4 and the lateral . . . :a6, and . . . iLa3 to threaten . . . iLb2. 14 iLg5 White looks to exchange dark-squared bish­ ops. He can also fight for the light squares with 14 iLc2 h6 ( 1 4 . . . lL'la5 15 iLg5 iLa3 16 lL'lh4 iLxc2 17 'it'xc2 'it'g4 has been played a couple of times by Negi, and although he lost both games the idea is worth considering) 1 5 iLe3 lL'la5 1 6 lL'ld2 .:tab8 (this is an odd little move but Black intends to play . . . lL'lc4, so he is antici­ pating the opening of the b-file) 1 7 iLxf5 'ilixf5 1 8 lL'lfl lL'lc4 1 9 iLc 1 c5 20 lL'le3 lL'lxe3 2 1 iLxe3 c4 =+= Nijboer-Yakovenko, Spanish Team Ch, Sabade1l 2008. Black's queenside play is more important than anything that White is doing. Another illustration of both sides' typical ideas is 14 h3 lL'la5 1 5 �c2 lL'lc4 16 iLg5 iLa3 1 7 lL'lh4 iLxc2 1 8 'ilixc2 a5 19 :te2 .:ta6 with un­ clear play, Stellwagen-Gustafsson, Dutch Team Ch, Enschede 2005 . 14 iLa3 ! ? Black can also play 1 4 . . . iLxg5 1 5 lL'lxg5 h6 1 6 lL'lf3 lL'ld8 (Anand), when the knight will come to the blockading square e6. ••.

20 :a6! 21 lL'lg4 b4 Black had the initiative in Bacrot-Anand, Bastia rapid 2004. •••

Section 1 0 . 5 1 e4 e5 2 lL'lf3 lL'lc6 3 iLb5 a6 4 iLa4 lL'lf6 5 0-0 iLe7 6 :el b5 7 iLb3 0-0 8 c3 d5 9 exd5 lL'lxd5 (D)

w

UNDERSTANDING THE MARSHALL A TTA CK

188

If White has come this far, he almost univer­ sally plays 1 0 ll)xe5, but there are a few other lines to be aware of. 1 88 A: 10 a4 1 88 B: 10 d4 Instead: a) 10 h3 iLb7 (or 1 O . . . iLf6) I I ll)xe5 ll)xe5 1 2 :xe5 ll)f4 gives Black good play. b) 10 d3 iLb7 ( 1 O . . .'ilfd6 looks reasonable, while 1O . . . iLf6 1 l ll)bd2 ll)f4 1 2 ll)e4 ll)xd3 1 3 ll)xf6+ gxf6 1 4 iLh6 ll)xe l 1 5 'ilfxe l ll)e7 1 6 .l:.d l 'ilfe8 1 7 ll)h4 iLe6 left White scrambling for compensation in Ponomariov-Anand, Leon rapid 2007) 1 1 ll)xe5 ll)xe5 1 2 l:txe5 'ilfd7 1 3 'itf3 l:tad8 1 4 ll)d2 c5 has been defended a cou­ ple of times by Zsu.Polgar: b 1 ) 1 5 'itf5 'ilfxf5 1 6 l:txf5 ll)f6 1 7 iLc2 ll)d7 1 8 ll)f1 g6 1 9 :f4 f5 with compensation, Lon­ don-Zsu.Polgar, New York Open 1 987. The f4-rook is misplaced. b2) 1 5 ll)e4 c4 ! ? 16 dxc4 ll)f6 17 iLe3 ll)xe4 1 8 l:txe4 'itd3 1 9 l:td4 iLxf3 20 l:txd3 l:txd3 2 1 gxf3 iLf6 2 2 'itg2 l:td7 2 3 cxb5 axb5 24 f4 was agreed drawn in this unclear ending in Spassky­ Zsu.Polgar, Cannes 1 987.

A) 10 a4 (D)

a) 1 3 ll)xe5 ll)xe5 1 4 l:xe5 ll)f4 gives Black good play as usual. b) 1 3 ll)a3 iLxa3 14 bxa3 'ilfd6 is equal, al­ though unbalanced. c) 1 3 d4 exd4 1 4 ll)xd4 (or 1 4 cxd4 iLf6 = ) 1 4 . . . ll)xd4 1 5 'ilfxd4 c5 is certainly OK for Black. 13 b4! ? (D) This looks like a clean way to equalize. The solid 1 3 . . . iLf6 = is also popular. •••

14 ll)xeS ll)xeS 15 AxeS iLf6 16 l:te4 bxc3 17 bxc3 ll)xc3 18 ll)xc3 iLxc3 = Kr.Georgiev-Kaidanov, Torcy 1 99 1 .

B) 10 d4 (D)

B

This cannot be very dangerous. Neither the opening of the a-file nor the attack on the b5pawn will amount to anything. 10 iLb7 11 axbS axbS 12 ':xa8 iLxa8 13 d3 White plays very solidly. Other options: •••

This move is logical enough. White simply plays in the centre and tries to develop quickly. Nevertheless, Black also has free piece-play and should not experience any difficulties.

OTHER ANTI-MARSHAll LINES

10 exd4 1 1 cxd4 (D) I I tLlxd4 has been played more frequently, but after l l . . .tLlxd4 1 2 'ilfxd4 �b7 Black al­ ready has a lead in development and can gain even more time with . . . c5 or . . . �f6. With the text-move White prepares to develop his knight to c3 to put pressure on the d5-knight. •••

189

tLld3 14 ':e3 tLlxc l 1 5 ':xc l tLla5 1 6 �a2 �d6 1 7 tLle5 � gave White some initiative in Par­ ligras-Gyimesi, Bundesliga 2007/8 . 12 .i.b4 Now this move looks quite good. The pin is annoying and the bishop can come around to b6 to exert pressure on the d4-pawn. 13 a3 �a5 14 �a2 (D) •..

B

1 1 �f5 Black develops the bishop actively and sizes up the d3-square. There are several alternatives: a) 1 1 . . . tLlb6 was recommended by Keres. Although it has not been tried in grandmaster practice, it looks solid enough. b) 1 1 . . .�b4 attempts to disrupt White's co­ ordination, but moving the bishop again gives White some chances for an edge. 1 2 �d2 �b7 ( 1 2 . . . �g4 1 3 'ilfc 1 'ilfd7 14 �xd5 'ilfxd5 1 5 �xb4 was Nisipeanu-Handke, Sants 2005, and after 1 5 . . . tLlxb4 1 6 tLlc3 'ilfh5 1 7 tLle5 White may be a little bit better) 1 3 tLlc3 �xc3 14 bxc3 tLla5 15 �c2 tLlc4 1 6 �g5 f6 17 'ilfd3 g6 1 8 �h6 .:n was Nisipeanu-Beliavsky, Pune 2004. Nisipeanu gives 1 9 tLld2 �. c) 1 1 . . .�g4 looks natural . One example: 12 tLlc3 tLlf6 1 3 �e3 'ilfd6 14 ':c l ':ad8 15 a4 b4 1 6 tLlb l tLla5 1 7 �c2 tLld5 1 8 'it'd3 g6 1 9 tLlbd2 tLlxe3 20 'ilfxe3 (20 fxe3 c5) 20 . . . �xf3 2 1 'ilfxe7 �a8 22 'ilfxd6 ':xd6 = M.Perez­ Sargisian, Zafra 2007 . 12 tLlbd2 This looks passive. It is more natural for White to follow through with 1 2 tLlc3, because 1 2 . . . tLldb4 ( 1 2 . . . tLlb6 ! ? may be better) 1 3 a3 •••

14 � g6 Or 1 4 . . . tLlf6 ! ?; the d4-pawn is in trouble. 15 h3 �h8 16 b4 tLlc3 17 'ilfb3 tLlxa2 18 'ilfxa2 �b6 + Nisipeanu-Kasimdzhanov, Vlissingen 2003. Black has the bishop-pair and a better pawn­ structure. •••

Concl usions The slow 8 d 3 and 8 a 3 tend t o lead to quiet play that is more typical of other lines of the Ruy Lopez. However, by playing in this manner White does not have the space advantage that he almost always enjoys in the Closed lines and Black should equalize without much difficulty. The most important and topical line is 8 d4, which can lead to different kinds of positions with both sides fighting for the initiative. The counter-strike 9 d4 leads to a different kind of game altogether, but Black has done quite well here. White's deviations on move 10 are not very dangerous and Black should reach a healthy position with natural, logical play.

I ndex of Va riations

Cha pter G u ide 1 e4 e5 2 liJf3 liJc6 3 1i.b5 a6 4 1i.a4 liJf6 5 0-0 1i.e7 6 lIe1 b5 7 1i.b3 0-0 S c3 8 a4 - Chapter 8 8 h3 - Chapter 9 Other moves - Chapter 1 0 S d5 9 exd5 9 d4 - Chapter 1 0.4 9 liJxd5 10 liJxe5 Other moves - Chapter 1 O.S 10 liJxe5 1 1 l:txe5 c6 Other moves - Introduction 12 d4 1 2 l:te l - Chapter 7. 1 1 2 1i.xdS - Chapter 7.3 12 d3 1i.d6 1 3 l:te l : 1 3 . . . 'ilfh4 - Chapter S 1 3 . . . 1i.fS - Chapter 6 12 1i.d6 13 .:tel 1 3 l:te2 - Chapter 7.2 13 'ilfh4 14 g3 'ifh3 15 1i.e3 I S 'ili'e2 - Chapter 7.4 I S 1i.xdS - Chapter 7.4 IS l:.e4 - Chapter 4 Other moves - Chapter 7.4 15 1i.g4 16 'ilfd3 l:taeS 1 6 . . . fS - Chapter 3 . 1 17 liJd2 lIe6 17 . . . fS - Chapter 3.2 1 7 . . . 'ifhS - Chapter 3 . 3 lS a4 Other moves - Chapter 3 .4 Now: 1 8 . . . fS ? ! - Chapter 2 1 8 . . . bxa4 - Chapter 2 1 8 . . . 'ilfhS - Chapter 1

Cha pter- byChapter I ndex 1 e4 e5 2 liJf3 liJc6 3 1i.b5 a6 4 1i.a4 liJf6 5 0-0 1i.e7 6 l:te1 b5 7 1i.b3 0-0

Part 1 : Main Lines

•..

••.

.••

•••

•••

.••

S c3 d5 9 exd5 liJxd5 10 liJxe5 liJxe5 11 l:txe5 c6 12 d4 1i.d6 13 l:te1 'ifh4 14 g3 'ii'h3 1 : Spassky Va riation I S 1i.e3 1i.g4 1 6 'ifd3 lIae8 1 7 liJd2 .:te6 18 a4 'ilfhS 20 1 9 axbS axbS 20 A: 20 1i.xdS 2 1 B : 20 liJe4 2 1 C: 20 liJn 22 C 1 : 20 ... :fe8 22 C2: 20 . . . 1i.fS 24 C2 1 : 2 1 'ilfd l 24 C22: 21 'ilfd2 25 D: 20 'ilfn 26 D l : 20 ... .:.fe8 26 D2: 20 . . . 1i.h3 28 2: Old Main Line: 1 8 f5/ 1 8 bxa4 I S 1i.e3 1i.g4 1 6 'ifd3 l:tae8 1 7 liJd2 l:te6 1 8 a4 2.1 : 1 8 . . . fS ? ! 31 A: 1 9 'ilfn 32 B : 1 9 axbS ! ? 37 2.2: 1 8 . . . bxa4 39 1 9 1ha4 fS 20 'iii'n 'ilfhS A: 2 1 c4 40 B: 2 1 ':'xa6 41 2.3: 1 8 . . . bxa4 1 9 l:txa4 fS 20 'ifn 'ilfhS 2 1 f4 42 • . •

...

A: 2 l . . .gS ? ! 42 22 1ha6 gxf4 A I : 23 ':'xc6 43 A2: 23 1i.xf4 ! 43 B : 2 l . . .l:[b8 ? ! 44 22 1i.xdS cxdS B 1 : 23 'ilfg2 45 B2: 23 l:txa6 46 C: 2 l . . .l:.fe8 46 C l : 22 ':xa6 47 C2: 22 'ilff2 48 22 . . . gS ! C2 1 : 23 l:txa6 48 C22: 23 fxgS 49 3 : 1 5 1i.e3 : Other Li nes I S 1i.e3 51 I S . . . 1i.g4 1 6 'ifd3 3.1 : 1 6 . . . fS 51 1 7 f4 gS 1 8 'ili'n 'ilfhS 1 9 liJd2 'iti>h8 20 1i.xdS cxdS A: 2 1 a4 52 B : 2 1 fxgS ! ? 53 3.2: 1 6 . . . ':ae8 1 7 liJd2 fS 54 1 8 'ifn 'ifhS 1 9 f4 'iti>h8 20 1i.xdS cxdS A: 2 1 'ifg2 56 A I : 2 l . . .gS 55 A2: 2 l . . .l:le4 56 B : 2 1 a4 57 B l : 2 l . ..gS ? ! 57 B2: 2 l . . .bxa4 58 3.3: 1 6 ... l:tae8 1 7 liJd2 � 60 A: 1 8 1i.c2 61 B : 1 8 'ifn 63 C: 1 8 a4 64 3.4: 1 6 . . . l:tae8 1 7 liJd2 l:te6 A: 1 8 1i.d l ? ! 65 B : 1 8 �gS ? ! 66 C : 18 c4? ! 66 D: 1 8 1i.xdS 67 1 8 . . . cxdS 1 9 'ifn 'ili'hS 20 a4 D l : 20 . . . fS ? ! 68 D2: 20 . . . bxa4 69 E: 1 8 'ifn 69 1 8 .. .'ifhS 1 9 f3 ! ?

INDEX OF VARIATIONS

E l : 1 9 . . . .tf5 ? ! 70 E2: 19 .. .1:1f6? ! 71 E3 : 19 . . .ltJxe3 71 4: Modern Va riation: 1 5 l:.e4 15 lIe4 74 15 ... g5 ! A: 1 6 'ii'e 2 75 B : 1 6 'ii'f3 77 1 6 . . . .tf5 B l : 1 7 .txd5 77 B2: 1 7 .te2 78 C: 1 6 'ii'e l 80 C l : 1 6 . . . f5 80 C2: 1 6 . . . .ltf5 81 D: 1 6 'ii'f l 81 D 1 : 1 6 . . . 'ii'x fl + 81 D2: 1 6 . . . 'ii'h6 ! ? 83 D3 : 1 6 . . . 'ikh5 85 1 7 ltJd2 .tf5 1 8 f3 ltJf6 ! D3 1 : 1 9 .l:.e l 87 D32: 19 'iVg2 88 D33: 1 9 a4 89

Part 2 : Other Lines 8 c3 d5 9 exd5 ltJxd5 10 ltJxe5 ltJxe5 1 1 ':xe5 c6

5: Refined Rook- Lift: 1 2 d3 1 2 d3 92 1 2 . . . .ltd6 1 3 l:.e 1 'ii'h4 1 4 g3 'ii'h 3 1 5 .l:.e4 A: 1 5 . . . .td7 93 B : 1 5 . . . ltJf6 94 C: 1 5 . . . 'ii'd7 96 1 6 ltJd2 .tb7 17 .l:.e l e5 1 8 ltJe4 .te7 C l : 1 9 .tg5 98 C2: 1 9 a4 99 D: I5 ... 'ii'f5 99 I 6 ltJd2 'ii'g 6 D 1 : 17 a4 100 D2: 1 7 ltJfl 101 D3: 1 7 :e l 102 1 7 . . . f5 D3 1 : 1 8 f4 103 D32: 1 8 a4 104 D33 : 1 8 e4 1 06 6: El ite Eq ualizer: 1 2 d3 .td6 13 .l:.e l .i.f5 1 2 d3 .td6 13 .l:.e 1 .ltf5 108 A: 1 4 ltJd2 108

B : 1 4 'ii'f3 109 B l : 14 . . .1:le8 1 1 0 B2: 14 . . . 'ii'h4 1 12 15 g3 'iVh3 B2 1 : 1 6 ltJd2 1 13 B22: 1 6 .txd5 1 14 16 ... exd5 B22 1 : 1 7 .i.f4 1 14 B222: 1 7 .i.e3 1 14 B223: 1 7 'ii'xd5 1 15 1 7 . . . lIad8 1 8 'ii'g 2 B223 1 : 1 8 . . . 'ii'h5 1 16 B2232: 1 8 . . . 'ikxg2+ 1 18 7: Early Deviations 7.1 : 1 2 .l:.e l 120 1 2 . . . .td6 1 3 g3 A: 13 ... .l:.e8 1 2 1 B : 1 3 . . . .tf5 1 22 7.2: 1 2 d4 .td6 1 3 .l:.e2 125 A: 1 3 . . . .tg4 125 B: 13 . . .'iVh4 126 14 g3 B l : 14 . . . 'ii'h5 126 B2: 14 . . . 'ii'h 3 1 2 7 7.3: 1 2 .ltxd5 exd5 1 3 d4 .td6 1 4 l:.e3 128 7.4: 1 2 d4 .td6 1 3 .l:.e l 'iVh4 1 4 g3 'ii'h 3 15 'ii'e2 131 A: 1 5 . . . .tg4 131 B : 1 5 . . . .td7 132 7.5: 12 d4 .td6 1 3 .l:.e l 'ii'h4 1 4 g3 'ii'h 3 134 1 5 .txd5 134

Part 3 : Anti- Marshall 8: Anti-Marsha l l : 8 a4 8 a4 137 8.1 : 8 . . . b4 137 A: 9 d4 138 B: 9 d3 139 9 . . . d6 B 1 : 1 0 ltJbd2 140 B 1 1 : 1 O ... .i.e6 140 B I 2: 1 O . . . ltJa5 141 B2: 1 0 a5 142 1 O . . . .lte6 1 1 ltJbd2 'ii'e 8 B2 I : 1 2 .lte4 144 B22: 1 2 ltJe4 145 8.2: 8 . . . .tb7 147 9 d3 d6 A: 10 e3 148 1O . . . ltJa5 1 1 .ta2 e5 1 2 ltJa3 'ii'd7 1 3 axb5 axb5 1 4 .td2

191

A I : 1 4 . . . h6 149 A2: 1 4 . . . e4 149 B: 10 ltJe3 1 50 C: 1 0 ltJbd2 152 9: Anti-Marshal l : 8 h3 8 h3 155 8 . . . .tb7 9 d3 A: 9 . . . d5 ! ? 155 B: 9 . . . d6 159 10 a3 ltJa5 1 1 .ta2 e5 B l : 1 2 ltJe3 161 1 2 . . . ltJe6 B 1 1 : 1 3 .tg5 1 61 B 1 2 : 1 3 ltJh2 1 62 B 1 3 : 1 3 ltJd5 1 63 B2: 1 2 ltJbd2 164 1 2 . . . ltJe6 1 3 ltJfl .lte8 ! B 2 1 : 1 4 e3 165 B22: 1 4 .i.g5 166 B23: 1 4 ltJe3 167 1 4 . . . .te6 B23 1 : 15 .td5 167 B232: 1 5 ltJd5 1 68 1 0 : Other Anti-Marsha l l Li nes 10. 1 : 8 d3 1 70 10.2: 8 a3 1 72 10.3: 8 d4 1 73 8 ... ltJxd4 A: 9 .i.xf7+ 1 74 9 . . . l:.xf7 1 0 ltJxe5 :f8 1 1 'ii'xd4 e5 1 2 'iVd l A I : 1 2 . . . 'ii'e7 1 75 A 2 : 1 2 . . . .tb7 1 76 B : 9 ltJxd4 1 78 9 . . . exd4 1 0 e5 ltJe8 B 1: 1 1 'ii'xd4 1 78 B2: 1 1 e3 1 80 l l . . .dxe3 1 2 ltJxe3 d6 B2 1 : 1 3 'ii'f3 181 B22: 13 .td5 182 10.4: 8 e3 d5 9 d4 1 83 A: 9 ... dxe4 1 83 B : 9 . . . exd4 184 1 0 e5 ltJe4 1 1 exd4 B l : 1 l . . ..tg4 185 B2: 1 l . . ..tf5 186 10.5: 8 e3 d5 9 exd5 ltJxd5 187 A: 1 0 a4 1 88 B: 1 O d4 1 88