Kill K.I.D.: v. 1

Chess Stars www.chess-stars.com

Current Theory and Practice Series

Kill KID 1 First edition - April 2009

Cover design by Kalojan Nachev Copyright © 2009 by Semko Semkov

Printed in Bulgaria ISBN: 978-954878270-8

Kill KID 1 A White Repertoire with the Four Pawns Attack Semko Semkov

Chess Stars www.chess-stars.com

Bibliography This book offers a new approach to the Four Pawns Attack and it is based on my own analysis. It practically does not overlap with any previous work on this subject. I did check up a number of sources, but I could not "buy" anything from them so I decided to not include most of them in the list.

Books Opening for White According to Kramnik, vol. 1a and 1b, by Alexander Khalifman, Chess Stars 2007 Understanding the King's Indian, by Mikhail Golubev, Gambit 2006 Periodicals Chess Informant 1-101 Internet resources Chess Today Internet Chess Club (chessclub.com) Playchess.com ChessPublishing.com forum Databases: The Week In Chess (chesscenter.com) 10 Days (Chessmix.com)

Contents

Introduction .......................................................................................... 6 About the Structure of This Book......................................................... 8

Modern Benoni Pawn Structures Part 1 1 d4 lLlf6 2 c4 g6 3 lLlc3 ~g7 4 e4 d6 5 f4 c5 6 d5 0-0 7 lLlf3 e6 8 ~e2 exd5 9 cxd5, Lines without 9 ... E:e8 ................ 9 Part 2 7 lLlf3 e6 8 ~e2 exd5 9 cxd5 E:e8 ............................................ 47

King's Indian Defence and Volga Pawn Structures Part 3 5 f4 c5 6 d5 0-0 7 lLlf3 Rare Lines without 7... e6 ................. 72 Part 4 5 f4 0-0 6 lLlf3 Lines without 6 ... c5 ........................................81

Modern Defence and Other Rare Move Orders Part 5 1 d4 g6 2 e4 ~g7 3 c4 ............................................................. 97 Part 6 1 d4 g6 2 e4 ~g7 3 c4 d6 4 lLlc3 lLld7 5 lLlf3 .......................... 113 Part 7 1 d4 lLlf6 2 c4 d6 3 lLlc3 Lines without ...g6 .......................... 124 Index of Variations ............................................................................ 137

5

Introduction Warning! Do not kill any kid around, I only refer to the King's Indian Defence! That one, I have been trying to kill for more than 20 years. This book is based mostly on my own original analysis of the Four Pawns Attack (FPA). It does not offer just minor improvements here and there in some irrelevant variations. I devised new plans in all the major systems and I propose a detailed analysis of all the arising branches I could think of. I quit active chess long ago so I have no reason to withhold any novelty I came upon over all those years. On the contrary, I have always wanted to share my numerous discoveries. I feel that current theory of the FPA is totally messed up if not entirely wrong. However, I had a gaping hole in the main Modern Benoni line and it took me many years to strike upon a decent idea. I hope it will cause considerable headache to Black players. I have always been unhappy with the Classical variation against the KID. It accumulated such an enormous amount of theory that it is impractical to keep in pace with it even 6

for a professional player. Even bigger problem is the character of positions that arise. You probably remember Radjabov's show in the Wijk aan Zee 2007 super tournament, where he won all his three black KID games against heavily prepared top rated players. There is something basically wrong - to give the second players such attacking chances as in the Classical variation. I understand that in the KID White sacrifices time in exchange for space. But that space can also be used in reverse - to repel the enemy pieces and slice through Black in the centre. Such a strategy has obvious pluses. White is the aggressor so he limits the opponent's choice to a very narrow survival path. Instead of following a well tested attacking scheme, as it happens in most of the Classical variation games, Black must defend with only moves. The "only" drawback is that Black had a rock solid equaliser in the main line: 1 d4 l2Jf6 2 c4 g6 3 l2Jc3 ~g7 4 e4 d6 5 f4 c5 6 d5 0-0 7 l2Jf3 e6 8 ~e2 exd5 9 cxd5 Ele8 10 e5 dxe5 11 fxe5l2Jg4

Introduction

White used to play here 12 ig5 when 12 ...~b6! should lead at least to a draw. My proposition is: 12 e6! fxe613 d6! This variation has been played before, but I link it with ideas that aim for sound positional compensation instead of depending on long, forcing variations. You will find about 30 pages of original analysis as well as a critical survey of the current theory. Beside this variation, most theoretical books recommend to Black the line: 1 d4lLlf6 2 c4 g6 3lLlc3 ig7 4 e4 d6 5 f4 c5 6 d5 0-0 7 lLlf3 e6 8 ie2 exd5 9 cxd5 ig4 10 0-0 lLlbd7

The current state of theory is favourable to Black, but I believe that my approach will change that. Another 30 pages of new analysis show that White very often should checkmate the opponent with a direct attack. The lead in is 11 h3! ixf3 12 ixf3 followed by g4. A major fine point of my plan is that White does not play gel nor a4 at all. In Part 4 I reach the authentic KID waters where Black refrains from ... cS in favour of 6... lLla6 or other continuations. I am sure that White is clearly better after 6 ... lLla6 7id3! The best part of having the FPA in one's repertoire is that Black will often try to outsmart you with tricky move orders and second grade systems which commonly delay ... lLlf6 or ... g6. I wantto supply a complete repertoire based on the FPA so in Parts 5-7 I examine all the deviations of Black from the common KID move order. I do not deal, however, with 1 d4 d6. In my opinion, the best answer is 2 e4! which could lead to the Pirc. Chess Stars has published a whole 400-pages-thick book on that subject - An Opening for White According to Anand, volume 4. I am not contemplating a Kill KID 2. I hope that the readers themselves will write it with their games if they like my ideas. But Chess Stars will keep open this option. Perhaps someone is working on, say, the Saemish?!

Semko Semkov April 2009 7

About the structure of this book

I follow the Chess Stars trademark structure, introduced by "The Safest Sicilian". Each system is examined in a separate part which contains three chapters: "Quick Repertoire"; "Step by Step"; "Complete Games". You start with the "Quick Repertoire". You'll find there all the vital information that you need to start playing the variation. These chapters contain more explanation and try to pinpoint the essence of the numerous variations, analysed branch by branch in the "Step by Step" chapters. Knowledge of the key lines presented in "Quick Repertoire" should be sufficient to make you well prepared for any live tournament.

Finally, the "Complete Games" sections give practical examples and sometimes cover backup lines of the main repertoire. The focus of the commentaries is on the critical moments of the game and on typical plans of both sides. I suggest that you read all three chapters even if you are a titled player. Or you may initially skip the details and use the "Step by Step" in postmortem analysis or during online games. "Quick Repertoire" is not just a summary of "Step by Step" chapters. I often use it to explain verbally the main ideas of the variation or the reason I prefer it to other lines. This information is omitted later as I prefer to avoid repetition. I suppose that players above Elo 1800 will benefit most of this book.

Part 1

1 d4 CLlf6 2 c4 g6 3 llJc3 i.g7 4 e4 d6 S f4 cS 6 dS 0-0 7 llJf3 e6 8 i.e2 exdS 9 cxdS QUICK REPERTOIRE

The first thing a Four Pawns attacker must learn is how to cope with the position you see on the above diagram. You may have read in many books that KID style play with 5 ... 0-0 6 CLlf3 CLla6 is "very interesting" and modern, but the truth is quite different. White players feel happy in the pure King's Indian structure as they enjoy a typical for this opening space advantage, without running the risk of being crushed under a kingside assault. Even conventional wisdom, which is programmed in most engines, acknowledges this fact. Computers give albeit little, but still positive for White scores in the KID structures, while the Modern Benoni branch with 9 .. .l'\e8! is negatively rated. Most protagonists of the FPA gave it up because of that forced variation which seems to lead to a draw as White's limit. I propose a completely new approach as you shall see in the next part. I should add here that White cannot avoid transition to the Mo-

dern Benoni. It would have been nice to have a backup variation as 5 f4 0-0 6 CLlf3 c5 7 dxc5 ~a5 8 i.d3 ~xc5 9 ~e2 which hides a lot of venom. However, the neat move order with 5 ... c5! 6 dxc5 ~a5 7 i.d3 offers Black the excellent resource 7... CLlfd7! I'm not sure White can even equalise, e.g. 8 i.d2 CLlxc5 9 i.c2 CLlc6 10 CLld5 ~d8. Now I'm going to deal with less testing in my opinion systems, which are however highly praised in chess literature. Thus they are very often met in practice. A. 9 ••. bS?! 10 eS dxeS 11 fxeS llJg4 12 i.gS ~b6 13 0-0 c4+ 14 ~h1

9

Part 1 White is clearly better here. Golubev advocates 14 ...b4, but after 15 ltJe4ltJd7 he misses 16 e6 fxe6 17 hc4±. Alternatives are: a) 14 ... ltJd7 15 e6 fxe6 16 dxe6 ltJdf6 17 e7 l=l:e8 18 a4! ib7 19 axbS vtJc7 20 vtJd4±; b) 14... ltJxeS1SltJxeS heS16 d6! This move ensures a clear edge. 16... ltJd717 i.e7 i.b718if3M319 vtJxf3 hd6 20 M8 l=l:xf8 21 vtJdS±. White went on to win in DittmarKaposztas, Kecskemet 1990. B.9 ... tilbd7 This is a very solid move which eliminates all the sharpest lines after 9 ... l=l:e8 or9 ...ig4. On the other hand, Black accepts to playa cramped position. His only active plan is linked with the temporary pawn sac c5-c4 which would empty the cS-square for the d7-knight. The whole variation has been developed mostly by Hungarian players and Endre Vegh even suggests it as a main repertoire against the FPA in his "Starting Out: Modern Benoni" book (2004). White's main task should be to anticipate ... c4 and set up e4-eS. It is essential to remember the correct move order: 100-0 ge8

11 'lMfc2! a6 12 a4 'lMfc7 (Or 12 .. l=l:b8 13 i>h1!) 13 tild2! Vegh's only note about this move is: "13 ltJd2 was possibly better." Well, it certainly is! 13... gb8 14 @h1

Black's light-squared bishop is stranded on c8. Passive continuations would be grim for him because we will put a knight on c4. Therefore in practice he chooses: 14... c4 15 e5 dxe5 16 tilxc4 b5 17 axb5 axb5 18 d6 'lMfc6 White had a clear advantage in the game Gershon-Manor, Israel 2002.

c. 9 ....ig4 10 0-0 tilbd7

Some books decorate 9 ... ig4 with an exclamation mark. That reflects the current state of theory which is favourable to Black. I like it 10

5 f4 c5 6 d5 0-0 7 ttJf3 e6 8 .te2 exd5 9 cxd5 very much for White since it offers me a chance to launch a "blitzkrieg" on the kingside. Black will need an utmost precision in order to survive during the next 15 moves. 11 h3! A crucial decision. It practically commits White with further pawn advances on that wing. 11 h3 has gradually disappeared after the key game Kouatly-Kindermann, Trnava, 1987 which introduced the novelty 11 h3 .txf3 12 .txf3 E1e8 13 g4 h6! 14 h4? h5!! 15 g5 ttJg4! with a satisfactory game. Then White players began to dodge this line in favour of 11 E1el (not too successfully!). I think that White's Icing's rook belongs to f1! 11 ....bt3 12.bt3

Now let me explain my approach. 1. My general plan is to attack on the kingside with the g-and hpawns. After h2-h4-h5xg6 White will need his rook either on hI, (if Black recaptures by hxg6) or on f1. I want to put on e1 the queen's rook. 2. The reserve plan of White is to keep the g-pawn on g4 in order to enable f4-f5 (after the preliminary sacrifice e4-e5 d6xe5). In that

case the fl-rook supports the attack down the f-file. I practically never play E1fel even later in the game. In my treatment of this system the main action happens along the f- or h-file. Now let's get acquainted with several key lines which should help us find our way amongst various move orders: 1. 12 ... ttJe8 13 g4!? (13 .te3!) 13 ... ltJc7 14 g5 b5 15 h4 b4 16 ttJe2 ttJb5 17 h5 c4 18 hxg6 hxg6?

This is a typical position with an open h-file. Black is probably lost here. He should not be able to survive the attack along the h-file, combined with the coup of grace f4-f5: 19 IDg2 ~b6 20 f5!± .td4 21 ~el gxf5 22 ~h4±. 2. 12 ... a6 13 g4 h6 14 h4 ttJh715 g5 hxg516 hxg5 b5

11

Part 1 The same structure as in the previous example, but the knight is on h7. It impedes f4-fS owing to the hit on g5, but renders Black's counterplay on the opposite wing sterile. We should be concentrating forces on the kingside because we have nothing to fear from the left: 17 ~g2 ~e8 18 ~hl ~c8 (18 ... c4 19 ~e3) 19 ~el ttJhf8

White has the better prospects. He attacks with prevailing forces. Black cannot survive with a passive stand, so he will sooner or later have to push .. .f6. In that case he will at least get some pressure on e4 so White can anticipate it with 20 ~f2;!;; (although 20 ~h4 should be good, too) 20 .. .f6 21 gxf6 ttJxf6 22 ~c2, bolstering the centre and eyeing g6. White's bishop pair has great potential. See the "Complete Games" chapter, Game 1 Kouatly-Al Modiahki, Doha 1993,2 W. Arencibia-B. Gonzalez, Cuba 1993 and 3 Vaisser-Kindermann, Biel1991 to learn more about this structure. 3. 12 ... ttJe8 13 g4!? (13 ~e3!) 13 ... ttJc7 14 g5 b5 15 h4 b4 16 ttJe2 ttJb5 17 h5 c418 hxg6 fxg6! 19 ~bl!? 12

The correct capturing by the fpawn gives White the better pawn formation. It should be smashing in an endgame, but with queens on, Black can hope to take over the initiative by some piece sac. Here is a typical tactical trick: 19 ~g2?! ttJc5 20 ~c2 ttJd3 21 ~xc4 ttJxf4+! 22 ttJxf4 ~xg5+ 23 ~hl ~xf4~. White must complete his development which explains the text move. The arising position is very sharp and I advise you to study carefully the "Step by Step" chapter. I'll give here only my main line: 19 ... ttJc5 20 ~e3!? ~c7 21 ~c2 c3 22 ~bdl ~ae8 23 bxc3 ttJxc3 24 ttJxc3 hc3 25 e5! dxe5 26 d6 ~c8 27 ~d5+ ~h8 28 ~f3

White's bishop pair is terrorising the opponent's king. For instance, 28 ... exf4? loses to 29 ~xc3+!!'

S f4 cS 6 dS 0-0 7 ct'lf3 e6 8 .ie2 exdS 9 cxdS 4. 12 ... ct'le8 13 .ie3! ct'lc7 14 '!Wd2 :§:b8 IS .ie2!

start with e4-eS, another time g4gS first is more precise. The problem is that we should seriously consider piece counter-sacrifices from Black, or sneaky retreats of the f6knight, when we risk to find ourselves over-extended. Here are two negative examples:

6. 12 ... :§:e813 g4 h6! 14 gS?! hxgS ISeS This move clears the f-file to the fl-rook and relocates the lightsquared bishop to a more active position. Its idea is best seen after IS ... bS 16 eS! dxeS 17 fS a6 18 d6 ct'le8 19 a4 with tangible pressure. Another nice line is IS ...:§:e8 16 eS dxeS 17 fS e4 18 d6 ct'la6 19 fxg6 hxg6 20 :§:xf7!+-. 5. 12 ... :§:e8 13 g4 h6! 14 '!Wc2! '!WaS?! IS eS! dxeS 16.gS! hxgS 17 fxgS ct'lh7 18 ct'le4;!;

The only thing Black must NOT do is to comply with our plan by taking IS ... dxeS? Instead, he can answer IS ... ct'lxeS! or IS ... ct'lh7 16 e6 gxf4! with an excellent game. The moral of this example is that firstly White should finish development by .ie3, :§:ael. Then both gS and eS become dangerous options. 7. 12 ...:§:e813 g4 h6! 14 ~hl?! (14 '!Wc2!) 14 ... :§:c8 (14 ...bS!) IS eS dxeS 16.gS ct'lh7

Look carefully at this position! Black is doomed to a stale position. This pawn structure is crucial for my treatment of the .ig4-variation. I discovered it IS years ago, but only recently found the correct paths to achieving it. Sometimes White can 13

Part 1 Black's queen is watching the g5-square and that restricts our attack: 17 f5! ttJxg5 18 fxg6 fxg6 19 ~g4 ttJf6! (19 ... ttJh7 20 Wfc2 ttJhf8 21 ttJe4 h5 22 ~g5 Wfb6 23 ~e2 c4 24 ~adl~) 20 hg5 hxg5 21 d6 ~b8 22 ttJd5=. White does have the initiative, and the opposite coloured bishop attack is unpleasant, but Black should be able to hold on. Notice that had Black's queen left d8 before, this breakthrough would have been very strong. Now we reach the best move order for White: 8. 12 ... ~e8 13 g4 h6! 14 Wfc2! a6 15 ~e3 b5 16 ~ael ~c8 17 h4!

always meet ... c4 with .ie3. 13 ... Wfa514 Wfd2 ~fe8 15 ~d4

Black's early ... c4 gave us an ideal square for our commonly unemployed dark-squared bishop. This is the only case when we do not need to advance our kingside pawns, because Black cannot impede the central breakthrough e4-e5: 15 ... ~ad8 16 @hl (16 Wff2!?) 16 ... ttJc5 17 e5 ttJfd7 18 e6 fxe6 19 hg7 @xg7 20 dxe6 ttJxe6 21 hb7:t.

Now I can summarize: In the following diagram after 12 M3 Black is at a crucial juncture. Now it is ripe time for a clash on the kingside. 17... h5 does not work so well as before since 18 g5 ttJg4 19 hg4 hxg4 20 e5! dxe5 21 f5 e4 22 fxg6 fxg6 23 ttJxe4:t is good for White. The counterattack 17... b418 ttJdl c4 is not effective either. After 19 g5 Black has no decent retreat with the knight. 9. 12 ... c4 13 ~e3 As a rule in this system, we

14

a) He can try to restrict our pawn storm by putting pressure on e4 with 12 ... ~e8 (12 ... c4 13 ~e3 gives White an easy game - see key line 9.) 13 g4 h6! In that case we should complete

5 f4 c5 6 d5 0-0 7 tLlf3 e6 S .te2 exd5 9 cxd5 development by 14 ~c2!, .te3, fi:ael, preparing h4!, see key line S. b) He can anticipate our attack by 12 ... tLleS. Then it is out turn to make a choice. We can immediately proceed with the ultra-sharp 13 g4!? tLlc7 14 g5, key lines 1-3. It might earn easy points, but also gives Black more counterchances. I prefer the more restrained 13 .te3! tLlc7 14 ~d2 fi:bS 15 .te2!, intending to meet nearly everything by 16.e5 dxe5 17 f5! with a great game - see key line 4. c) Black also has a rare, but tricky move order: 12 ... a6. Perhaps best now is 13 .te3 b5 14 g4 or 13 ... fi:eS 14 ~c2 b5 15 fi:ael fi:cS 16 g4 h6 as in line Cl of the "Step by Step" chapter.

In line C3 I also examine 13 g4, and only then 13 ... tLleS 14 g5 tLlc7!. My analysis suggests that Black would get probably sufficient counterplay after 15 h4 tLlb5, so we must make him pay dearly for the right to install a knight on b5: 15 a4!

15 ... b5 16 h4 bxa4 17 fi:xa4 tLlb5 IS ~d3. Black has three weak pawns on the queenside to worry about.

15

Part 1

1 d4 lL)f6 2 c4 g6 3 lL)c3 i.g7 4 e4 d6 5 f4 cS 6 dS 0-0 7 lL)f3 e6 8 i.e2 exdS 9 cxdS STEP BY STEP

In this part we shall consider: A. 9 ... b5?! B.9 ... tLlbd7 C.9 ....tg4 The most critical move 9 ... !i:e8 is the subject of the next part of the book. Minor alternatives are: a) 9 ... a6 10 0-0 b5 11 e5 dxe5 (1l ... tLle812 .te3 tLld713 a4;!;) 12 fxe5 tLlg4 13 d6! There is no reason to shift the black queen to a better square with 13 .tg5 ~b6.

b) 9 ... tLla6. This rare line is underestimated by theory, but it is not bad at all. White cannot punish this development by an early e4-e5 so he should probably switch to central plans: 10 0-0 (10 e5? tLld7 11 e6 fxe6 12 dxe6 tLlb6) 1O ... tLlc7

11 !i:el.

11 e5 tLld7! is roughly equal (11 ...

13 ... tLlxe5 14 ~d5 tLlxf3+ (or 14 ... tLlec6 15 ~xc5 .tb7 16 a4;!;) 15 M3 .td4+ 16 i>h1 !i:a7 17 .tg5 ~a5 18 .te7 .te619 ~g5 !i:e8 20 !i:ael±. 16

dxe5? 12 d6 e4 13 dxc7 ~xd114 !i:xd1 exf3 15 M3 .tg4 16 f5! is clearly in White's favour) 12 exd6 tLle8 13 f5 tLlxd6 14 fxg6 hxg6 15 .tg5 ~b6! (practice has only seen 15 ....tf6 16 ~d2 tLle5 17 M6 ~xf6 18 !i:ael;!;) 16 ~d2 tLlf6=. 11 h3 does not work due to 11...tLlh5 12 .tc4 b5! 13 tLlxb5 tLlxb5

5 f4 c5 6 d5 0-0 7 ctJf3 e6 8 i.e2 exd5 9 cxd5 14 hb5l'!b815 i.c4l'!b4! (15 ... hb2 16 hb2 l'!xb2 17llNc1llNb6 18 ctJd2±) with compensation. 11 ...i.g4 (11 ...b5 12 e5) 12 h3 hf3 13 hf3 ctJd7! (13 ... l'!b8 14 e5±) 14 i.e3 l'!b8 15 a4. White's game is somewhat easier due to his better centre. A. 9 ••. bS?! 10 eS dxeS Bulgarian grandmaster Bobotsov introduced in 1959 the move 1O ... ctJfd7, but later White discovered the simple 11 hb5 dxe5 12 0-0 i.a6 (or 12 ...i.b7 13 l'!el exf4 14 hf4 ctJf6 15 d6 ctJc6 16 l'!cl±) 13 a4 with a considerable advantage. 11 fxeS lLJg4 11 ... ctJfd7 is bad since Black cannot develop normally after 12 i.f4, e.g. 12 ...b4 13 ctJe4±. Even better is 12 e6 ctJe5 13 i.f4 ctJxf3+ 14 hf3±.

12 i.gS \19b6 12 .. .f6 13 exf6 hf6 offers White a juicy choice. Virtually every active continuation gives him a huge advantage. My computer likes 14 d6±. Humans prefer 14 llNd2 hg5 15 ~xg5 llNxg5 16 ctJxg5 with a difficult ending for Black owing to the tremendous strength of the d5-pawn: 16 ... ctJe3 17 @d2 ctJxg2 18 l'!afl a619 3xf8+ @xf8 20 l'!fl + i.f5 21 i.d3 ctJh4 22 i.e4 h6 (22 ... ctJd7loses a piece to 23 l'!f4) 23 d6 l'!a7 24 ctJe6+ @f7 25 .'Dxc5 ctJd7 26 b4±. Finally, 14 hf6 llNxf6 15 llNd2 b4 16 ctJe4 llNe7 17 ttJfg5 fits best in the true spirit of the FPA. Black should be crushed after something like 17... h618 d6llNe519 i.c4+ @g7 200-0-0 hxg5 21 i.d5±.

130-0 c4+ 14 h1

Al. 14 ... ctJd7 A2.14 ... ttJxe5 Alternatives: Golubev advocates 14 ...b4, but after 15 ctJe4 ttJd7, he does not consider 16 e6 fxe6 17 hc4±, for instance, 17...i.a6 18 ha6 llNxa6 19 llNb3 l'!ae8 20 d6±. 14 ... ctJf2+ 15 l'!xf2 llNxf2 16 ttJxb5 is reputed to be in White's favour. Let us analyse it a little further: 16 ...i.g4 and now 17 e6!?± is probably best, while 17 ctJc7 is more forceful: 17... hf3 18 hf3 llNxb2 19 i.f4 ctJd7 20 l'!bl llNxa2 21 e6±. A1.14 ... ttJd715 e6fxe616 dxe6 ttJdf6 17 e7 E:e8

17

Part 1 18 a4! This is more precise than 181Wd4 h6 191Wxb6 axb6 20 lLlxbS ~xe7 21 hc4+~h722.ih4;l;gS23.ig3which

is also better for White though. 18 ....ib7 19 axb5 Vf!c7 20 Vf!d4± White is more active and he even has a material advantage. A2. 14 ... lLlxe5 15 ~xe5 .ixe5

16 d6! I think that only this move ensures a clear advantage. The fine point is to compel the enemy knight to go to d7. Practice has only seen 16 .ie7 ~e8 17 d6, when Black has 17... lLlc6! 18 lLldS1WcSl9 b4 (19 lLlf6+ ixf6 20 hf6 .ifS+) 19 ...1Wd4 and White can win the exchange by 20 1Wxd4 lLlxd4 21 ~ael lLlxe2 22 ~xe2 .ig7 23 lLlc7 .id7 24 h3, but Black's bishop pair is a fearsome force. 16 ... ~d7 Black can give up material by 16 ....ie6 17 .if3 lLld7 18 has ~xa8 191Wf3 ~e8 20 ~adl± or 16 ....ib717 d7. In all the cases he should not obtain adequate compensation. 17 .ie7 It would be interesting to insert 18

17 a4!? 17 ....ib7 18 .if3 .ixf3 19 Vf!xf3 .ixd6 20 .ixf8 ~xf8 21 Vf!d5± White went on to win in DittmarKaposztas, Kecskemet 1990.

B. 9 ••• ~bd7 This is a very solid move which eliminates all the sharpest lines after 9 ... ~e8 or 9 ....ig4. On the other hand, Black accepts to play a cramped position. His only active plan is linked with the temporary pawn sacrifice cS-c4. I first faced this move in Semkov-Perenyi, Saint John 1988 and failed to respond in the best way. 10 0-0 ~e8

11 Vf!c2! So far so good. White should keep open the option of pushing eS. According to Megabase, my move was a novelty back then. I still think it is the only way to aspire to the advantage. The frequently played 11 lLld2 c4 (1l ... a6 12 a4 c4 13 ~hl lLlcS 14 eS dxeSlS fxeS ~xeSl6 lLlxc4 ~e817 .igS .ifS 18 d6;l;) is roughly equal: a) 12 ~hl lLlcS 13 eS dxeS 14 fxeS ~xeS IS lLlxc4

S f4 cS 6 dS 0-0 7 ttJf3 e6 8 i.e2 exdS 9 cxdS elude 1l ... a6 12 a4 since the bSsquare could be useful to White in some variations: 1l ... c412 hc4 ttJcS 13 eS i.fS (13 ... dxeS 14 fxeS ttJg4 1S e6±) 14 ~d1 dxeS 1S fxeS ttJg4

1S .. .2'l:e8! The rook will be doomed to perish on fS, although White must be precise. Tregubov-Vuilleumier, Clichy 2008 saw further 1S ... E1fS16 i.f4 !LJfe4 (16 ... gS 17 i.eS) 17 ttJxe4 ttJxe4 18 ttJe3 hb2, when instead of grabbing the rook by 19 ttJxfS hfS 20 2:b1 ~f6!=, White had 19 E1b1! ttJc3 (19 ... ~f6 20 ttJg4) 20 ~c2 ttJxb1 21 !LJxfS hfS 22 ~xb2±. 16 i.gS h6 17 M6 I do not believe that White could be better without his dark-squared bishop, but 17 i.h4 ttJce4 18 ttJxe4 2:xe4 is not inspiring either. 17 ... M6 18 ttJbS (In TregubovGruenfeld, Bastia 2004, Black equalised after 18 d6 by 18 ...hc3!? 19 bxc3 i.fS) 18 ... a6 19 ttJbd6 E1f8 20 ~c2 Black cannot avoid the sacrifice on fl, but it is not that terrible. 20 ...i.g7! 21 ttJxfl E1xfl 22 E1xfl ~xf7 23 E1f1 + i.fS 24 ttJe3 ~b6! 2S g4 ~xb2=. b) 12 eS dxeS 13 ttJxc4 ttJb6 14 fxeS ttJfxdS 1S ttJd6 ttJxc3 16 bxc3 gf8!=, Minescu-Marin, Baile Tusnad 200S. 11 ... a6 Commonly Black prefers to in-

16 e6!? 16 i.f4 ~b6 (16 ... ttJxeS 17 ttJxeS heS 18 heS E1xeS 19 ~d4) 17 ~d4 ttJxeS 18 ttJxeS heS 19 heS E1xeS 20 E1adU ttJd7 21 E1f4 is also pleasant for White. 16 ... fxe6 17 h3 ttJeS 18 i.bS! ttJxf3+ 19 E1xf3 E1f8 20 i.e3 ~b6 21 E1cl±. Here g4 is a nasty threat. 12a4~c7

Preparing ... c4 which is vital for Black's development. 12 .. E1b8 seems mundane as ... bS is impossible. However White must be cautious. The correct response is 13 Wh1! Note that 13 eS?! dxeS 14 fxeS walks into 14 ... ttJxeS! 1S ttJxeS E1xeS 16 i.f4 i.fS 17 ~b3 ttJe4+. 13 E1e1 also allows tactical tricks - 13 ... bS!? 14 axbS axbS 1S hbS ttJxdS! 16 exdS E1xe1 + 17 ttJxe1 hc3 18 bxc3 E1xbS 19 c4 E1aSoo. Finally, 13 as justifies Black's last move: 13 ...bS 14 axb6 ~xb6 1S ttJd2 ~c7 16 ttJc4 E1b4+t. 13 ttJd2! It is essential to impede ... c4. In 19

Part 1 the source game I made one prophylactic move too many and Perenyi took over the initiative after 13 Whl? c4 14lLld2 lLlcS. 13••• Etb8 14 c;t>h1

White has a clear advantage. The game Gershon-Manor, Israel 2002 went on 19lLlxeSlLlxeS 20 fxeS 1:!xeS 21 i.f4 i.fS 22 ~dl1:!cS 23 i.f3t. Perhaps 19 lLlaS!? ~a6 (19 ... ~xd6 20 lLlxbS ~b6 21 lLlc6±) 20 b4t is even better.

c. 9 ...J.g4 10 0-0 tObd7

Now this move is already good. Black is at a juncture. He should find some employment of his lightsquared bishop which is stranded on cB. The problem, however, is that passive continuations risk to aggravate Black's situation: 14...b6 15 lLlc4 (calm developmentas 15 1:!el is also good) IS ...bSI6 axbS axbS17lLlaS b41BlLlbS ~b619 lLlc4 ~xbS 20 lLlxd6 b3 21 ~dl ~b4 22lLlxeBlLlxeB 23 eSt or 14 ... lLlfB 15 as. Therefore in practice Black commonly speeds up play by: 14•.. c415 e5 dxe5 16 tOxc4 b5 17 axb5 axb5 18 d6 ~c6

20

a) 1O ... 1:!eB transposes to the main line after 11 h3. Instead, 11 eS i.xf3 12 i.xf3 dxeS 13 fxeS 1:!xeS 14 d6 lLlc6 15 i.f4 1:!eB 16 lLlbS 1:!fB is rather unclear although White has considerable compensation for the pawn. b) 1O ...i.xf3?! is imprecise, to say the least. This move is based on the assumption that White cherishes his knight so much, that he would play lLlgS to avoid its exchange. My treatment of White's "hand" however envisages a kingside pawn storm with gS and h2-h4-hS. In that case the early exchange on f3 presents us with an extra tempo, saved on h3. c) 1O ... a6 is a rare move which has many drawbacks and no pluses. White can continue with his general plan or force play by 11 eS dxeS

S f4 cS 6 dS 0-0 7 tLlf3 e6 8 ~e2 exdS 9 cxdS (11...tLle8 12 e6 fxe6 13 tLlgS±) 12 fxeS tLlfd7 13 e6 fxe6 (the computer suggestion 14 ~e3!? may be even stronger) 14 dxe6 he61S tLlgS ~e8 16 tLlxe6 ~xe6 17 ~f3 ~d4+ 18 cj;Jh1 tLlf6 19 hbn.

11 h3!? Here is the first important moment. This move has gradually disappeared after the model game Kouatly-Kindermann, Trnava, 1987 which introduced the cunning idea 11 h3 hf3 12 hf3 Ele8 13 g4 h6 14 h4 hS! lS gS tLlg4! with a satisfactory game. White players began to dodge this line in favour of 11 Elel. Perhaps you should be following how the favourite setup of Jobava develops. He retreats the e2-bishop to f1 in order to take on f3 by queen, then completes development by ie3-f2, and watches his time for slicing through the centre with e4eS, for example: 11 Ele1 Ele8 12 ~f1 ~c813 h3 hf314 ~xf3 a61S a4 c4 16 ~e3 tLlcS

My general plan is to attack on the kingside with the g-and h-pawns or by the pawn sacrifice e4-eS, followed up by f4-fS. I practically never play Elfe1 even later in the game. In my treatment of this system the main action happens along the f- or h-file.

11 ...ixf3 12 ixf3

Cl. 12 ... Ele8 C2. 12 ... tLle8 C3.12 ... a6 C4.12 ... Elb8 CS.12 ... c4 C1.12 .. J'!e8

17 ~f2 tLlb3 18 Elad1 tLld7 19 cj;Jh1 20 eS dxeS 21 d6t, Jobava~ebolsina, Benidorm 2007. Now let me explain the essence of my approach. ~aS

This move considerably restricts White's kingside pawn storm. It directly threatens with ...bS. For instance, if White tried some "finesse" as 13 cj;Jh1, then 13 ...bS! would take over the initiative. Note that 13 ... a6?! 14 g4 h6 lS ~c2 c4 16 eS! dxeS 17 gS or 13 ... c4?! 14 ~e3 bS lS a3 as 16 Ele1 Elb8 17 ~d4 b4 18 tLla4;!;; would be rather pleasant for White. We shall discuss such positions later in the main line. 1394 h6!

21

Part 1 as a safe equaliser and it convincingly discouraged White players from the pawn storm idea. They switched attention to the centre and the currently most topical line is 1ll"1e1, refraining from h3. I have always felt however, that White should not renounce this plan, but rather improve his implementation. My first thought was to save 14 h4 in favour of 14 g5!? hxg515 e5 Black has no decent retreat for the f6-knight, so the prophylaxis is indispensable. Engines also propose the absurd redeployment 13 ... tLlb6?! 14 g5tLlfd7, when we can go on 15 h4 c4 16 .ie3 l"1c8 17.id4! (17 h5 is persistent, but it misses the goal: 17... tLlc5 18 hxg6 hxg6 19 f5 hc3 20 fxg6 fxg6 21 bxc3l"1c7 22 .ig4l"1xe4 23 ~f3oo) 17... tLlc518 hg7 It>xg7 19 ~d4+ It>g8 (19 .. .f6 20 .ig4 tLlbd7 21l"1ael±) 20 h5tLlbd7 21 h6 f6 22 .ig4 l"1c7 23 l"1ael±. The text also sets up a curious positional trap: 14 h4 h5!!.

Now 15 gxh5? tLlxh5 16 hh5 is disastrous, so the source game Kouatly-Kindermann, Trnava 1987, went on 15 g5tLlg416 hg4 hxg417l"1el c418 .ie3 hc319 bxc3 l"1xe4=. This game has been accepted ~xh4+

22

In blitz this idea is extremely effective as Black automatically answers 15 ... dxe5? and gets into a murky position after 16 fxg5 tLlh7 17 tLle4, for instance, 17.. .f5 18 gxf6 tLldxf6 19 d6 or 17... c4 18 .ie3. We shall meet often this pawn structure. Remember that it is very pleasant for White! However, Black can intercept the initiative by counter-sacrificing a piece: 15 ... tLlxe5 16 fxe5 l"1xe5 or even better, 15 ... tLlh7! 16 e6 gxf4 17 exd7 ~xd7. Such a scenario is totally unacceptable to me, so I decided to postpone g5 for a while. The first candidate for an improvement was 14 It>hl. The reason behind this move is clear: It is always good to leave the dangerous diagonal gl-a7 and free the gl-

S f4 cS 6 dS 0-0 7 lLlf3 e6 8 ie2 exdS 9 cxdS square for a rook. At the same time, we can ignore nearly any Black activity on the queenside as 14 ... a6, 14 .. J~c8 or c4. To spare you irrelevant details, I shall say that White obtains a promising initiative in those cases, but 14 ... bS! proved to be a too hard test. Then lS gS hxgS 16 eSlLlh717 e6 gxf418 exf7+! 'it>xf7 19 hf4 is already in White's favour as Black has not the vital check from d4 to free g7 for his king. However, the other piece counter-sac remains good: 16 eS lLlxeS! 17 fxeS ~xeS. Perhaps White should have started with lS eS, I asked myself. But it does not work either, due to lS ... dxeS 16 gS lLlh7+. I'll not bore you any more with my futile investigation of earlier deviations which aimed to revive my idea. The summary is: to make it work, White needs to enable a third possible breakthrough, namely f4fS! This goal is best served by: 14 Wc2!

14 ... a6

Alternatives are: a) 14 ...bS1SlLlxbS c416 'it>h1! lLlcS 17 eS! dxeS 18 fxeS ~xeS 19 if4

White's bishop pair dominates in the centre and should ensure some edge, for example, 19 ... lLlxdS (19 ... ~xdS 20 lLlc7) 20 heS heS 21 ~ad1lLld3 22 ~xd3 cxd3 23 '&xd3 lLlc7 24 '&e2 lLlxbS 2S '&xeS '&d3 26 'it>gl ~f8 27 idS '&d4+ 28 '&xd4±. b) 14 ... '&aS?! It is really over-optimistic to expose the castling position to danger at this moment. We should not miss the chance: lS eS! dxeS 16 gS lLlh7 and now we see the main reason behind White's 14th move: 17 fS! with a dreadful attack. Instead, 16 ... hxgS 17 fxgS lLlh7 18 lLle4;!; would lead to my favourite pawn structure. c) 14 ... c41S ie3 bS

16 ~ae1! This move completes the mobilisation of White's forces and enables all kinds of breakthroughs on the 23

Part 1 kingside. Of course, the less committing redeployment 16 .tf2 a6 17 4:Je2 is also possible: 17... 4:Jb6 (Or 17... g5 18 fxg5 hxg5 19 4:Jg3 4:Je5 20 .te3:t; 17... 'Wc7 18 a4 b4 19 e5 dxe5 20 d6) 18 4:Jd4 4:Jfd7 with a tangled position. 16 ... a6 16 ...b4 17 4:Jb5 4:Jb6 18 .tf2:t will leave Black over-extended on the queenside. 17h4! It turns out that we did not completely discard this idea. It only needed a better framing - good development and ... a queen on c2! 17... h518 g5 4:Jg419 hg4 hxg4

24 E1xf7! 4:Jxh4 25 E1efl! 4:Jf3+ 26 E11xf3 gxf3 27 E1xf3! 'Wxd5 28 4:Jf6+ hf6 29 gxf6+-. I must confess that these lines conclusively won me for the 14 'Wc2! manoeuvre.

15.ie3 It is still early for 15 h4 h5 16 g5 4:Jg4 17 e5 dxe5 18 f5 e4!. 15 ... b5 16 l'!ae1 l'!c8 16 ... b4 17 4:Jdl E1c8 transposes to the main line. 17 h4!

Now we realise another thematic thrust: 20 e5!± dxe5 21 f5 Black's position is aboutto crumble down. 2l...4:Jf8 22 fxg6 4:Jxg6 meets 23 Ei:xf7! 4:Jxh4 24 Ei:eflEi:e7 25 E1xe7 'Wxe7 26 'We4 E1d8 27 'Wxg4 4:Jg6 28 4:Je4 iWd7 29 'Wh5±, while 21...e4 22 4:Jxe4 4:Je5 23 fxg6 4:Jxg6 (23 ... fxg6 24 4:Jf6+ hf6 25 gxf6 ~f7 26 h5 'Wd6 27 .td4 4:Jf3+ 28 ~hl ~f8 29 .tc5 iWxc5 30 'Wxg6 'Wa7 31 E1e7 E1xe7 32 fxe7+ 'Wxe7 33 iWxg4+-) is not too different: 24

Now it is ripe time for a clash on the kingside. 17 ... b4 White is on top after 17... h518 g5 4:Jg419 .txg4 hxg4 20 e5! dxe5 21 f5 e4 22 fxg6 fxg6 23 4:Jxe4:t. Black can also stay passively

5 f4 c5 6 d5 0-0 7 tLlf3 e6 8 with something like 17... tLlh7 18 g5 ~d4, but White's space advantage on the kingside provides good conditions for building up an attack after 19 tLldl. 18 tLld1 c4 For 17... iWa5 18 b3 see game 5 Semkov-The Cookie Lover, 3 min. Playchess.com 21.12.2008. The knight is closer to the kings ide from d1, so 18 ... h5 is hardly advisable: 19 g5 tLlg4 20 hg4 hxg4 21 tLlf2 iWe7. I have also analysed 21...g3 22 tLlh1 iWa5 23 tLlxg3 iWxa2

~e2

exd5 9 cxd5

~f4+-)

30 @g2+-. 25 l"1d1 b3 26 iWe2 iWb7 27 h5 d5 28 iWf3 tLlb6 29 hxg6 fxg6 30 f5 and White's pawn avalanche finally got to the enemy king. Let us return to the more reasonable 21...iWe7. We should continuewith: 22 ~d2 ~d4 23 @g2 hf2 24 l"1xf2 as 25 l"1fe2t. Black is pretty helpless against the powerful pawn centre. White can even walk the king to g4 before proceeding further with the attack. 1995tLlhS To 19 ... tLlh7 we can open the black castling position with 20 gxh6 hh621h5~.

20 ~xhS gxhS

24 e5! Rybka insists on 24 l"1a1 iWxb2 25 iWxb2 hb2 26 l"1xa6 claiming a White's advantage after 26 ... c4 27 l"1xd6 l"1c7 28 h5 gxh5 29 tLlf5 b3 (29 ... l"1xe4 30 l"1c6 l"1b7 31 @f2±) 30 ~g2 l"1xe4 31 l"1c6) 31 l"1b6 l"1xe4 32 3b8±. "She" (in most Slavic languages rybka means fish, but it is also a slang for chick) may be right, but I am an adept of the "human" approach. So we go for a mate: 24 ... iWxd5 Or 24 ... dxe5 25 f5 e4 26 f6 ~f8 27h5 tLle5 (27 ... gxh5 28 g6) 28 hxg6 tLlxg6 29 iWh2 iWxd5 (29 ... ~d6 30

The attack is running by itself, for example: 21 f5 hxg5 22 hxg5 b3 23 axb3 cxb3 24 iWg2 l"1c4 25 ~f2 ~d4 26 hd4 l"1xd4 27 f6 tLlf8 28 iWf3 l"1e5 29 iWxh5 tLlh7 30 l"1f5;!;.

C2.12 ... tLle8 A modern and challenging plan. Black is going to manoeuvre his knight to d4 via e8-c7-b5.

25

Part 1 through e4-e5, to be followed up by f4-f5. It is fairly safe as White does not expose his king too much and he enjoys a firm control of the centre.

This can be achieved by repelling the c3-knight from c3 with the help of b7-b5-b4, or by the modest a7-a6, tLlc7-b5. Black's counterplay is clear, fast and straightforward. I think that White is helpless to hold his queenside against the enemy's pawn majority and prevailing forces. So he should discard mundane development and focus on an active plan of his own rather than play on restriction. I explored two plausible courses of action: C2a. 13 g4!? and C2b. 13 .te3! The first one is extremely sharp. It features a lot offorceful variations. You can use it to "examine" your opponent. There is a fair chance that you win without making a move of your own. However, there is too little strategy in this approach and too much is on stake on every turn. If you like to play "All-in", this line offers a lot of fun. It looks especially eligible against a higher rated opponent since his better understanding will be of lesser use. The second line is completely different. White prepares the break26

C2a.13 g4!? White's next moves will be g4g5, h3-h4-h5xg6. 13 ...h6 would not stop g5 in view of 14 mg2 :1'1b8 15 h4! and the pawn is immune. (15 ... ~xh4 16 g5) 13 ••• tlJc7 14 g5 b5 Or 14 ... c415 .te3 b516 ~d2 tLlb6 17a3±. 13 ... a6 is considered in line C3a. 15h4

15 ... b4 Black can also try 15 .. .f5 16 exf5 gxf5. (16 ... :1'1xf5 17 .tg4 :1'1f8 18 h5 .td4+ 19 mg2 tLlb6 20 hxg6 hxg6 21 f5 b4 22 f6 bxc3 23 ~d3 ~e8 24 bxc3± .txf6 25 gxf6 or 17... ~f8 18 ~c2 tLlb6 19 i.xf5 gxf5 20 a4 b4 21 tLlb5±) White obtains an initiative by attacking the weak f5-pawn and threatening to open the g-file: 17 h5 .td4+ 18 mhl tLlb6 19 tLle2 .th8 20 :1'1g1 ~d7 (20 ... ~e7 21 :1'1bl :1'1ae8 22 b3t) 21 :1'1bl :1'1ae8 22 b3 :1'1f7 23 g6 :1'1fe7 24 ~d3 c4 25 bxc4 bxc4 26 ~c2

5 f4 c5 6 d5 0-0 7 ltJf3 e6 8 ~e2 exd5 9 cxd5 .'Llcxd5 27ltJg3 Ei:f8 28 hd5+ ltJxd5 29 ~xc4 ~e6 30 ~d2ltJb6 31 gxh7+ :t>xh7 32 ~d3 ~f6 33 Ei:belt. 16 tlJe2 tlJb5 17 h5 c4 It is important to push quickly c4-c3 to get real counterplay. 17... ltJd4 18 ltJxd4 hd4+ 19 \t>g2 only helps White in his plan. 17... ~b6 also loses an important tempo: 18 hxg6 fxg6 (18 ... hxg6 19 \t>g2) 19 ~g4 c4+ 20 \t>g2 ltJc5 21 1e3 Ei:ae8 (21...hb2? 22 ~e6+ \t>h8 23 Ei:bl +-) 22 ~e6+ \t>h8 (22 .. J"1xe6 23 dxe6 ~c6 24 ~d5 ~xd5 25 exd5 :z:Jd3 (25 ... hb2 26 hc5 dxc5 27 ~ae1±) ) 23 Ei:h1±. 18 hxg6 fxg6! This is the only move, but in blitz the vast majority of players opt for 18 ... hxg6?!. It allows White to get decisive attack along the h-file: 19 ~g2 ~b6 20 f5!±

to a forced draw. Only most precise play retains the initiative:

19 Ei:b1!? Alternatives: a) 19 \t>g2?! ltJc5 20 ~c2 ltJd3 21 ~xc4ltJxf4+ 22ltJxf4 ~xg5+ 23 \t>hl Ei:xf4--+; b) 19 ~c2 c3 20 bxc3 ltJxc3 21 ltJxc3oo; c) 19 ~e3 hb2 20 Ei:bl c3 21 ~g4 ttJc5, see 19 ~g4; d) 19 ~g4!? ltJc5 20 ~e3!? Unfortunately, 20 f5? stumbles into 20 ...hb2! 21 hb2 ~xg5. Much more entertaining is 20 ~c2 c3 21 bxc3ltJxc3 22ltJxc3 hc3 23 Ei:bl ~e7 24 ~d2oo Ei:ae8 25 Ei:bel h6!? (obtaining counterplay) 26 hc3bxc3

20 ... gxf5 21 exf5 f6 22 g6 Ei:fe8 23 ~h1±;

20 ...~d4 21 ~el gxf5 22 ~h4±; 20 ... ~e5 21 ~el c3 22 Ei:hl+-. After the correct capturing, Black has more chances to maintain the balance. Although White's pawn formation hides great potential, his exposed king allows various destructive sacrifices that often lead

27 e5! hxg5 28 ~xg6+ ~g7 29 gxf4 30 Ei:f2 c2 31 ~xc5 (or

~xd6

27

Part 1 31 gg2 gxe5 32 .ie6+ ~h8 33 gel gg5=) 31...gc8 32 \!!!fd6 c1\!!!f 33 \!!!fe6+ gO 34 gxc1 gxel + 35 gn gxf1 + 36 ~xf1 \!!!fh7 37 .if3=. 20 ...hb2! Black would be lost if he let the white bishop to reach e6: 20 ... ltJd3 21.ie6+ ~h8 22 b3 \!!!fc7 (22 ...ha1 23 bxc4) 23 gb1 gae8 (23 ... ltJb2 24 gxb2 hb2 25 ~c2 1J.g7 26 \!!!fxc4±) 24 f5 .ie5 25 ~g2 gxf5 26 exf5 ltJa3 27 gc1 ltJxc1 28 hel cxb3 29 axb3 \!!!fc2 30 ~xc2 ltJxc2 31 f6 ltJd4 32 ltJxd4 hd4 33 1J.f4+- or 20 ... ltJxe4 21 ie6+ ~h8 22 \!!!fc2 ltJc5 23 \!!!fxc4 gb8 24ltJd4 ltJxd4 25 hd4 ~c7 26 gacl. 21 gb1 c3 (21...ltJc3?! 22 ltJxc3 hc3 23 ie6+ ltJxe6 24 dxe6 h6 25 f5 hxg5 26 ~g4!--t) 22 hc5 dxc5 23 \!!!fb3

This extremely unbalanced position is rather dangerous for Black, but he should escape after: 23 ... ltJd6! The attack on e4 forces play. 23 ... \!!!fb6 24 ~g2 a5 is very risky: 25 ie6+ ~h8 26 ~b2 a4 27 \!!!fc2 b3 28 gxb3 axb3 29 axb3 ~a6 30 gf2oo. I suspect that White is better here. At least, his game is easier because he is the more active side, for ex28

ample: 30 ... gfb8 31 f5 ~d6 32 fxg6 \!!!fe5 33 gf3 ltJd4 34 ltJxd4 cxd4 35 gh3 gb7 36 iO gxf7 37 gxf7 ~xg5+ 38 gg3 \!!!fe5 39 \!!!fe2± ~f4 40 \!!!ff2 ~xe4+ 41 ~f3 \!!!fxf3+ 42 ~xf3+­ c2 43 ggl gf8 44 ~e4 gxf7 45 ~xd4 gc7 46 gel ~g7 47 d6. 24 ie6+ ~h8 25 gxb2 ltJxe4 26 ltJxc3 bxc3 27 ge2 gxf4 28 gxf4 \!!!fxg5+ 29 gg4 \!!!fel+ 30 ~h2 ltJd2= with perpetual check. 19 ... ltJc5

19 ... \!!!fa5 deprives Black's king of adequate defence: 20 ig4 ltJc5 21 ie3ltJd3 22 .ie6+ ~h8 23 ~g2! (23 f5 gxf5 24 exf5 ltJc7 25 ltJf4 ltJxf4 26 gxf4 ltJxe6 27 fxe6 gxf4 28 hf4=) 23 ... ltJc7 24 ltJd4 ltJxe6 (24 ...~xa2 25 f5) 25 ltJxe6 ltJxb2 26 \!!!fe2 ltJd3 27f5 gxf5 28 ghl--t. The h-file turns to be fateful again. 19 ... c3 20 bxc3 ltJxc3 21 ltJxc3 bxc3 22 ie3 ~a5 23 \!!!fe2 leads to a position where the mobility of White's central pawns causes Black a lot of trouble. The main threat is ig4 and f5, but e5, followed up by f5 is also an option.

20 1J.e3!?

20 ~c2 is not in the spirit of the

S f4 cS 6 dS 0-0 7 ctJf3 e6 8 ie2 exdS 9 cxdS variation. It does win the c4-pawn, but hands the initiative to the opponent. It could also lead to a forced draw: 20 ... ctJd4 (20 ... c3? 21 bxc3 ctJxc3 22 ~xb4±) 21 ctJxd4 hd4+ 22 li>g2 ~c8 23 ig4 (23 Wxc4 ctJe6) 23 ... ~e8 (23 ... ~c7 24 Wxc4) 24 ~el ctJd3 (24 ... c3 2S bxc3 hc3 26 ~e3 2:c7 27 ib2 hb2 28 Wxb2 as 29 ~d4;!;) 2S ~dl ~xe4 26 ~xd3 cxd3 27 ~xd3 We8 28 hc8 Wxc8 29 Wxe4 ~g4+=. Alternatives are risky for White: 21 Wxc4 ctJxf3+ 22 ~xf3 ~c8 23 ie3 Wd7 24 ~bf1 (or 24 hcS 2:xcS 2S Wd3 ~fc8 26 fS ieS 27 ~f2 gxfS 28 exfS Wf7co) 24 ... Wg4+ 2S ctJg3co. 20 •. .'II~·c7 21 Wc2 c3 22 ~bd1 gae8 22 ... cxb2 23 eS ctJc3 24 ctJxc3 bxc3 2S Wxc3 ctJe6 26 Wd2! is unpleasant for Black. He must give up a piece without sufficient compensation: 26 ... ctJxgS 27 fxgS heS 28 id4 hd4+ 29 Wxd4 Wb6 30 Wxb6 axb6 31 ~bl ~xa2 32 ie2;!;. 23 bxc3 tLlxc3 24 tLlxc3 hc3 Or 24 ...bxc3 2S eS WaS 26 ig4 dxeS 27 fS Wb4 28 ih3 gxfS 29 ixfS e430Wh2;!;. 25 e5! dxe5 26 d6 'Wc8 27 i.d5+ li>h8 28 ~f3

White's bishop pair is terrorising the opponent's king. The nicest variation here is 28 ... exf4? 29 Wxc3+!!' More reasonable alternatives are: a) 28 ... Wg4+ 29 Wg2 Wxg2+ 30 li>xg2 ttJd7 31 ha7 ~xf4 32 a4±; b) 28 ... ~fS 29 ~h3 hS 30 d7 ttJxd7 (30 ... Wxd7 loses to 31 ie4 ~xgS+ 32 li>h2+-) 31 ~xhS+ gxhS 32 WxfS exf4 33 'Wxf4t; c) 28 ... ttJe6 29 fxeS ctJd4 30 ~xf8+ ~xf8 31 Wg2 with an initiative, for example, 31...ttJfS 32 ib7 We6 33 ha7 heS 34 dn. C2b. 13 .ie3! tLlc714 'Wd2 ~b8 14 ...bS is obviously premature due to IS eS, but Black has two other reasonable options: a) 14 ... a6 IS i,e2!

We are all set now for e4-eS. IS .. .fS The only way to impede White's plan. IS ...bS is rather difficult. (1S ... We7 16 eS ctJe8 17 e6± ttJb6 18 fS gxfS 19 ~xfS fxe6 20 ~xf8+ li>xf8 21 i,gS) 16 eS dxeS 17 fS. Next we push d6 with total domination. 16 exfS gxfS The pawn structure has changed 29

Part 1 significantly and we should adjust our plans accordingly. Now the centre is fixed and we must define other objectives to pursue. I would list them in the following order: 1. To restrict Black's play on the queenside; 2. To activate the e3-bishop via f2 and eventually h4; 3. To put the f1-rook on the efile. (Well, this is an exception to the rule!) 4. To bind the enemy queen with the defence of the f5-pawn. Exchanges of rooks do not help Black as that pawn will become more vulnerable. We can begin with 17 a4, when the computer suggests 17...b5 18 axb5 i.xc3 19 bxc3 axb5. Then we follow up with number 2 of our todo list: 20 i.f2 'tJb6 21 ~ab1 ~d7 22 .ih4 h6 23 ~fe1 ~fe8 24 i.h5 ~e4 25 ~xe4 fxe4 26 f5±. b) 14 ... ~e8!? 15 g4! Black has discouraged our planned manoeuvre i.e2 and we must urgently revive our positional aim - to cramp the opponent by e4-e5, f4-f5. It may seem that White can insert first 15 ~ad1 ~b8 and only then play 16 g4 b5 17 e5. That would be true if Black had to answer 17... 'tJ b6 18 ~f2 'tJc419 'tJe4 'tJxe3 20 ~xe3±, but he can capture on e5: 17... dxe5! 18 f5 b4 (18 ... e4 19 'tJxe4 'tJe5 20 ~f2±) 19 'tJe4 'tJb5 20 fxg6 hxg6 21 d6 'tJd4 22 'tJg5 'tJf6 23 'tJxi7 ~d7 and Black is holding, e.g. 24 i.xd4 cxd4 25 g5 ~xf7 26 gxf6 i.xf6 27 30

i.d5 ~xd5 28 ~xf6 ~e6 29 ~g5 ~xf6 30 ~xf6 @h7=. 15 ...b5 16 e5 'tJb6 17 ~ad1. Black has pushed b5, but it has stricken at thin air. 17... 'tJc418 ~f2 'tJxe3 19 ~xe3

White's threat is 20 'tJe4±, so 19 ... dxe5 seems the only move, but then 20 f5! e4 21 'tJxe4 puts White on top. 15ie2! b5 Alternatives are: 15... a6 16 a4 b5 17 axb5 axb5 18 e5 dxe5 19 f5± or 15... ~e8 16 e5 dxe5 17 f5 e4 18 d6 'tJa619 fxg6 hxg6 20 ~xf7!+-.

16 e5! dxe5 17 f5 a6 18 d6 18 'tJe4 c4 19 b3 is also interesting. 18 ... 'tJe8 19 a4 White has tangible pressure.

5 f4 c5 6 d5 0-0 7 tLlf3 e6 8 ie2 exd5 9 cxd5 C3.12 ... a6 Perhaps White's best answer now is 13 ie3!? and play will most probably transpose to another main line. For instance: 13 ... E1e8 14 Vf1c2 b5 15 E1ael E1c8 16 g4 h6 - see line Cl, or 13 ... b5 14 g4, when the move ... a6 might prove to be useless. Next I will analyse positions of independent significance where Black executes the manoeuvre ... tLlf6-e8-c7-b5. 13 9 4

ging 15 ... c4 in view of 16 tLld4. You can see the game in the "Complete Games" chapter. The game Bagaturov-Grigorian, Belgorod 1989 went on with 14... f5 15 exf5 gxf5, when simplest is 16 a4±. 15 a4! It looks imperative to slow down the enemy counterplay with ... tLlb5. 15 ig4ltJ b5 16 e5 tLlxc3 17 bxc3 dxe5 18 f5 e4 19 f6 tLlxf6! allowed Black in game 4 Nogueiras-Velimirovic, Reggio Emilia 1986, to balance the game. I have fiddled for many days with 15 h4 tLlb516 tLle2 (16 id2, as in Adianto-Hulak, Jakarta 1986, cannot be serious, nor could be 16 Vf1d3 c4!) 16 ... tLld417 tLlxd4 cxd4 18 h5 E1e8 19 hxg6 hxg6 (19 ... fxg6 20 f5) 20 ~g2

C3a. 13 ... tLle8; C3b. 13 ... h6

C3a.13 ... lbe8 Intending tLlc7-b5-d4. 14 g5lbc7 Game 7 Semkov-Z.Ilic, Saint John 1988, saw 14 ...b5. This is rather inconsistent as Black could have prepared this advance by tLle8c7 or ... E1b8, saving a tempo on a6. I could have answered with the thematic 15 h4, following the model of line C2a, (15 h4 c4 16 ie3 Vf1c7 17 h5 ltJc5 18 Wc2 ltJd3 19 hxg6 hxg6 20 ie2 Wd7 21 f5-+) but I decided to take the maximum of the concrete move order by 15 ltJe2!?, discoura-

The open c-file gives Black just enough counterplay to maintain the balance: 20 ...Vf1c7 (or 20 .. .f6 21 ig4 tLlc5 22 f5! We7 23 E1el±) 21 f5 ie5 (21...E1ac8 22 f6 Wc2+ 23 E1f2±) 22 id2 Vf1c4 23 fxg6 fxg6 24 ig4 ltJc5 25 Vf1f3 E1f8 26 Vf1h3 Vf1c2 27 ie6+ ltJxe6 28 Wxe6+ ~g7=. I could not find an improvement on this variation. 31

Part 1 15 ...b5

16h4 Aiming for h4-hSxg6. Perhaps 16 @g2!? is slightly more precise, because it cuts off some side variations. Anyway, play is similar to the main line: 16 .. bxa4 17 l:'lxa4 ctJbS 18 ~d3 ctJb6 19 l:'la2 c4 20 ~c2 ctJd4 21 ~f2 ctJxf3 22 ~xf3 l:'le8 23 ~e3;l;. 16 ... bxa4 A typical Modern Benoni approach. Black revives the idea with ctJc7-bS-d4. The alternatives do not create immediate threats so they offer White a free hand on the kingside: a) 16 ... c417 ~e3 b418 ttJa2 as 19 l:'lc1 ~xb2 20 l:'lxc4 ctJa6 21l:'lc6 ctJdb8 22 l:'lc2 ~g7 23 ctJc1 ~d7 24l:'lh2; b) 16 ...b417 ttJe2 ~e718 @g2 (18 hSl:'lae8 19 @g2 f6 20 hxg6 hxg6 21 gxf6 M6 22 ctJg3;l;) 18 .. .f6 19 gxf6 hf6 20 eS ctJxeS (20 ... hh4 21l:'lh1) 21 fxeS ~xeS 22 ~gS;l;. 17 gxa4 It is tempting to try 17 hS ctJbS 18 hxg6. Then 18 ... fxg6?! 19 ctJxa4! (Black was threatening with 19 ... a3.) 19 ... ttJd4 20 ~g4 ~e7 21 l:'la3 l:'la7 22 l:'lh3± would be advantageous to 32

White. Note that 19 ctJxbS axbS 20 ~g4 c4 21 ~e6+ @h8 22 ~e3 hb2 23l:'lb1 (23 @g2 hS) 23 ... ~g7 24 @g2 ctJcS 25 hcS dxcS 26 l:'lxbS l:'lxf4 27 l:'lxf4 a3 28 ~g4 a2 29 l:'lfl a1~ 30 l:'lxa1 hal 31 ~f4 is probably level. Black should recapture by 18 ... hxg6! 19 ctJxa4 l:'le8 20 @g2 ctJd4 with initiative. 17.. )ilb5 18 ~d3 This way of meeting ... ctJ bS would have been impossible without the inclusion of 15 a4 and 16 17 l:'lxa4 now Black has not 18 ... c4. By the way, I'm not too sure the text is the best move. 18 ctJe2!? might be stronger, to avoid the simplification I mention after the next diagram. Anyway, Black has to prove first that he is OK in my main line, before starting to worry about this one. 18 ••• tLlb6 19 ga2 19 l:'la1!? is also worth consideration. 19 ... c4 Or 19 ... ~d7 20 @g2 ctJd4 21 ~d1 fS 22 gxf6 hf6 23 hS;l;. 20 ~c2

White's king looks a bit exposed, but the superb pawn centre

S f4 cS 6 dS 0-0 7 tLJf3 e6 8 i.e2 exdS 9 cxdS keeps the enemy pieces at bay. Notice that his game is rather easy as he only should doubt about which pawn to push first - hS or fS, while Black's choice is more difficult. For instance, if he trades his knight for the f3-bishop, he will reduce the number of our attacking pieces, but will leave without counterplay: 20 ... tLJd4 21 Wf2 tLJxf3+ 22 Wxf3l=l:e8 23 i.e3. Besides the kingside, Black has to think about three pawn weaknesses as well - 23 ... tLJd7 24 l=l:a4 gb8 2S l=l:xa6 l=l:xb2 26 tLJd1 l=l:b4 27 gxd6 if8 28 l=l:a6 tLJcS 29 fS l=l:b7 30 fxg6 hxg6 31l=l:f6;!;. In case of20 .. .fS 21ie3, 21...tLJd4 (21...fxe4 22 ig4±) is already late due to 22 ixd4 ixd4+ 23 ~g2;!;. 20 ••. l=l:e8 21 tLJe2 YlYc8 22 ~g2 Everything is ready for further expansion by hS. The computer offers as best 22 ... aS 23 hS YlYcS 24 hxg6 hxg6 2S l=l:h1 a4 to organise counterplay, but then 26 fS ieS 27 gh6! ig7 28 l=l:h3 (28 fxg6 ixh6 29 gxf7+ ~xf7 30 ihS+ ~e7 31 ixe8 hgS 32 ixgS+ ~xe8 33 Wd2;!;) 28 ... tLJd7 29 f6± is close to winning. C3b. 13".h6 14 h4 tLJh7

The idea ofline Cl14 ...hS?! here only serves to lose a pawn without compensation after lS gS (lS gxhS? tLJxhS 16 ixhS YlYxh4!) lS ... tLJg4 16 ixg4 hxg4 17 YlYxg4 id4+ 18 ~g2 We7, when we can go forward without lingering: 19 fSl=l:fe8 20 f6 tLJeS 21 We2 YlYd7 22 hS ixc3 23 bxc3 WbS 24 c4 YlYxc4 2S Wxc4 tLJxc4 26 hxg6 l=l:xe4 27 g7 tLJeS 28 if4±. 15 g5 bxg5 16 bxg5 b517 ~g2 l=l:e818l=l:h1l=l:c8 (18 ... c419 ie3) 19 YlYe1 tLJhfS

White has the better prospects. He is attacking with superior forces on the kingside, while the enemy counterplay on the opposite wing is rather unimpressive without the second knight. Black cannot survive with a passive stand, so he will sooner or later have to push .. .f6. In that case he will at least get some pressure on e4 so White can prepare to meet it with 20 ~f2;!; (although 20 Wh4 should be good, too) 20 ... f6 21 gxf6 tLJxf6 22 Wc2, bolstering the centre and eyeing g6. White's bishop pair has great potential. See game 1 KouatIy-Al Modiahki, Doha 1993, 2 W. Arenci33

Part 1 bia-B. Gonzalez, Cuba 1993 and game 3 Vaisser-Kindermann, Biel1991 about this structure.

C5. 12 ... c4 13 .te3

C4.12 .. J~~b8 13 g4 h6

Only this move is of independent significance. 13 ...bS 14 gS ct:Je8 transposes to line C2. 14 h4 tLle8 15 g5 h5

We know from line C2 that an open h-file gives White a crushing attack. Now the situation has changed and White must alter his plan accordingly. The new target is g6! So White's aim should be to push eS, followed up by fS. That becomes good when Black cannot immediately blockade the dS-pawn by... ct:Je8-d6. Thus we should continue developing, waiting for the opponent to shift his knight from e8: 16 Wl'c2 b5 17 .te3 tLlc7

Or 17...b4 18 ct:Je2 ct:Jc719 eS :1'l:e8 20 i,f2 dxeS 21 fS e4 22 he4 hb2 23 fxg6±. 18 e5 dxe5 19 f5

White achieved his objective and he is clearly on top. 34

With his last move Black discouraged 13 g4, because both his knights would find perfect stands after 13 ... ct:JcS 14 gS ct:Jfd7. This advance has its obvious drawbacks though. Black abandons its important outpost on d4. Furthermore, the d4-square becomes the fulcrum of White's play. For instance, i,e3d4 would support the breakthrough e4-eS, while ct:Jc3-e2-d4 could also become an option. As a rule in this system, we always meet ... c4 with i,e3. 13 ... WI'a5

13 ... :1'l:e8 can be answered by 14 ~d2 as in the main line, but the greedy comp prefers to snatch a pawn: 14 ~a4!?, when neither 14 ... ct:Jxe41S ct:Jxe4 fS16 ~xc4 fxe417 i,g4, nor 14 ... :1'l:c8 lS ~xa7 bS 16 a3 ct:JcS 17 hcS± should satisfy Black. Only 14 ... ~c8 saves the pawn, but then lS :1'l:ac1 a6 16 ~b4 ~c7 17 a4 :1'l:ac8 18 :1'l:fe1 sees White dominating on both wings. The plan with g4 is also possible, as you can see from the commentaries of game 8 Blokh-Morozevich, Moscow 1992.

S f4 cS 6 dS 0-0 7 ttJf3 e6 S .ie2 exdS 9 cxdS 14~d2

This move looks the most harmonious to me, although 14 '!!ie2, hitting c4, is probably sufficient for obtaining some advantage: 14 .. J'lacS (14 ...bS stumbles into IS a4 b416 ttJbS) IS Elael (IS g4 offers Black the initiative after IS ... ttJcS! 16 gS ttJfd717'!!ixc4 ElfeS.)

White's plan is to wait for ttJcS and push eS. Then he can go further on with e6 or take on d6 in order to make a passer on the d-file. Meanwhile he has plenty of useful moves as ~hl, '!!if2, .id4 or .ig4. Black should devise a plan of redeployment of his pieces, since the pawn advance IS ... bS would only make a weakness and isolate Black's queen from the centre. White will answer 16 a3, bolstering the c3-knight. a) IS ... ttJb6. I would not even consider this passive move, but the engines suggest it for a second best line. White has many attractive continuations, for instance, 16 ~hl ttJfd7 17 .ig4 ElcdS IS hd7 Elxd7 19 fSt with serious initiative. b) IS ... ElfeSI6 ~hl (16'!!if2 eliminates to a roughly equal endgame: 16 ... ttJcSI7 hcS'!!ixcSIS '!!ixcS ElxcS 19 eS dxeS 20 fxeS ttJd7 21 ttJe4 Elc7

22 e6 fxe6 23 ttJd6 EldS 24 dxe6 ttJeS 2S e7 Elxe7 26 .idS+ ~hS 27 ttJf7+ Elxf7 2S i1.xf7 bS with full compensation for the exchange, e.g. 29 Ele2 ttJd3 30 .ieS b4 31 Ele7 .ih6 32 Elff7 ~gS 33 .ibS ttJxb2 34 Elxh7 .ifS 3S Eld7) 16 ... ttJcS 17 eS dxeS IS fxeS ElxeS 19 '!!ixc4 EleeS.

White has a slight edge here thanks to his bishop pair and a strong passed pawn in the centre. Perhaps he should not force play, but I was curious about the endgamefollowing20b4!?'!!ia621'!!ixa6 lIJxa6 22 lIJbS lIJxb4 23 ttJd6 ttJc2 24 lIJxcS ttJxel 2S Elxel ElxcS 26 ha7t. It turned out that Black's defence is not easy: 26 ... ElaS 27 .ie3 Elxa2 2S d6± ttJd7 29 hb7 Elc2 30 .if4. c) IS ... ttJcS The first choice of the engines. However, I suspect that it is the worst one. 16 eS! ttJfd7 (16 ... dxeS 17 fxeS lIJfd7 IS e6±) 17 e6 ttJb6 IS fS! (I make such moves in the FPA automatically.) IS ... ElceS. (1S ... fxe6 19 fxg6 hxg6 20 dxe6±) My practice has convinced me that when White's dark-squared bishop has an access to the kingside in such positions, there is always some destructive tactical blow at hand: 3S

Part 1

19 i.hS!! fxe6 20 fxg6+-. Now let's return to 14 'l!;Yd2:

1S ... :aad8 Black has not an active plan of his own so he puts his rook in an opposition to the d2-queen. 1S ... llJcS 16 eS is clearly in White's favour: 16 ... llJfd717 exd6llJd318 ,txg7l!?xg7 19 ie2±. 16 <;t>h1 (16 'l!;Yf2!?) 16 ... c!l)cS 17 eS c!l)fd718 e6 fxe619 .ixg7 <;t>xg7 20 dxe6 c!l)xe6 21 .ixb7t

14 .. JUe8 1S .id4

Black's king is definitely underprotected.

36

Part 1

1 d4 c!iJf6 2 c4 g6 3 c!iJc3 J.g7 4 e4 d6 5 f4 cS 6 dS 0-0 7 c!iJf3 e6 8 J.e2 exdS 9 cxdS COMPLETE GAMES

1. Kouatly - AI Modiahki Doha 1993 1 d4 c!iJf6 2 c4 g6 3 c!iJc3 J.g7 4 e4 d6 5 f4 c5 6 d5 0-0 7 c!iJf3 e6 8 J.e2 exd5 9 cxd5 J.g4 10 0-0 c!iJbd711 ge1 ge812 h3 i.xf313 i.xf3 a6 14 g4 h6 15 h4 b5 16 g5 hxg5 17 hxg5 c!iJh7

18 g2 Kouatly chooses the plan with !'!hl. It is very efficient when White had not lost a tempo on !'!el. In the current situation I would think about the other thematic plan with the pawn sac e4-e5 and attack on g6. This idea works when White is fully mobilised, so he can start

with 18 ~c2 c4 19 .te3 !'!c8 20 .tg4 !'!c7 (or 20 .. c!iJhf8 21 e5) 21 e5! dxe5 22f5

This possibility is yet another argument against playing !'!el - the rook has nothing to do on el, but it can return to fl: 22 ... c!iJxg5 23 fxg6 c!iJc5 24 ~d2 c!iJge4 25 c!iJxe4ltJxe4 26 ~g2ltJd6 27 gxf7+!'!xf7 28 .te6 !'!xe6 29 dxe6 !'!f6 30 !'!fl±. 18 ... c!iJb6 This setup does not offer serious counterplay. I think that only 18 ... c4 19 .te3 !'!c8 leaves some chances. Then White should persist with his plan by 20 !'!hl ltJc5 (or 20 ...b4 21 ltJe2 ltJc5 22 ltJd4 ltJxe4 23 ltJe6t) 37

Part 1 21 a3 ttJd3 22 Wd2, because 20 Wd2 ttJcS 21 ixcS?! ElxcS is hardly good. Black will open the kingside with .. .f6 at an opportunity. 19 gh1 ga7 20 ttJe2llJc4 Perhaps Black discovered that 20 ... Elae7 21 Wc2 ttJc4 22 Elbl would leave him without any threats. 21 gb1 Wi'aS 22 a3 b4?! It is understandable that Black is trying to organise some counterplay, but his knight on h7 is too far from the queens ide and White easily overtakes the initiative on that wing. Black obviously did not expect the abrupt turn of events and got into a lost endgame: 23 Wi'd3! Wi'bS 24 gd1! llJaS 2S Wi'xbS axbS 26 axb4 cxb4 27 .te3 gaa8 28 b3 gac8 29 .tb6 llJb7 30 gbc1 The game is practically over. 29 ... llJf8 31 .tg4 gxc1 32 gxc1 gxe4 33 gc7llJcS 34 .txcS dxcS 3S <;t>f3 ge8 36 gxcS gb8 37 @e4 fS+ 38 gxf6.txf6 39 llJd4 ge8+ 40 llJe6 .tc341 gxbS llJxe6 42 .txe6+ @f8 43 @f3 @e7 44 gb7+ <;t>d6 4S <;t>e4 .td2 46 gd7+ <;t>cS 47 gc7+ <;t>d6 48 gc2 .tc3 49 gg2 1-0

2. W. Arencibia - B. Gonzalez Cuba 1993 1 d4 llJf6 2 c4 cS 3 dS e6 4 llJc3 exdS S cxdS d6 6 e4 g6 7 f4 .tg7 8 llJf3 0-0 9 .te2 .tg4 10 0-0 llJbd7 11 h3 .txf3 12 .txf3 ge8 13 g4 h6 14 h4?! llJh7?! 1S gS hxgS 16 hxgS White achieved the same pawn structure as in the previous game, 38

but having saved Elel. That puts his opponent into a difficult situation. He correctly assumes that staying passively on the kingside will be too risky and attempts to loosen the grip by: 16 ... f6

Notice that sooner or later Black should play this move anyway, for instance: 16 ... a6 17 ~g2 bS 18 Elhl c419 WeI ttJhf8 20 Wh4 f6, but it is too late due to 21 fS fxgS 22 ixgS .tf6 23 fxg6. 17 gxf6± Wi'xf6 17... ttJhxf6 is well met by 18 We2 intending Wg2. 18 .te3 ge7 19 @g2 bS 19 ... Elae8 would not stop 20 eS dxeS 21 d6 Ele6 22 i.dS+-. Now White wins a piece. 20 eSllJxeS 21llJe4 Wi'h4 22 gh1 llJc4 23 .tc1 1-0

3. Vaisser - Kindermann Biel1991 1 d4 llJf6 2 c4 g6 3 llJc3 ig7 4 e4 d6 S f4 0-0 6 llJf3 cS 7 dS e6 8 .te2 exdS 9 cxdS .tg4 10 0-0 llJbd711 ge1 ge8 12 h3 .txf3 13 ixf3 Wi'aS14.te3 gac81S g4 h616

5 f4 c5 6 d5 0-0 7 ct:Jf3 e6 8 ie2 exd5 9 cxd5 h4 b5 17 g5 hxg5 18 hxg5 ttJh7

In comparison to the previous examples, here Black adopted a more aggressive setup. He brought his queen onto the queenside and he is about to overrun the opponent by ... b4, ... c5-c4-c3. However, it is White's turn and Vaisser employs the proven and very efficient breakthrough in the centre: 19 .ig4 ~cd8 20 e5! dxe5 21 f5 e4! 22 fxg6 fxg6 23 .ie6+ ~xe6? Black panics and commits a decisive mistake. Commonly these positions with a white bishop on e6 are lost indeed, but here the opposition of the d8-rook with the white queen perhaps can help Black save the day. The critical moment comes after 23 ... \tJh8 24 iWg4ct:Jdf8:

I did not find a win for White: a) 25 ixc5 ct:Jxe6 26 dxe6 b4 27 ct:Jxe41"ld5 281"lac1 hb2 291"lc4 b3 30 iWe2 iWxa2 31ct:Jd61"lxg5+! (31...1"lxc5 321"lxc5 id4+ 33 \tJh1 iWxe2 341"lxe2 hc5 35 ct:Jxe8 ct:Jxg5 36 ct:Jc7 +-) 32 \tJfl 1"lf8+ 33 ct:Jf7+ 1"lxf7+ 34 exf7 1"lf5+ 35 \tJg1 ie5=; b) 25 ct:Jxe4 ct:Jxe6 26 dxe6 hb2 27 1"lad1 1"lxd1 28 1"lxd1 iWxa2 29 e7 1"lxe7 30 hc5 1"le8 31 ct:Jd6 1"lf8 32 ct:Jxb51"lb8 33 iWe2 \tJg8 341"ld2 iWa1+ 351"ld1=. These variations illustrate once again that White's rook is misplaced on el. 24 dxe6 ttJe5 25 e7 ~e8 26 'lWd5+ tJih8 27 tJig2 ~xe7 28 ~h1 ~f7 29 ~af1 ~xf1 30 ~xf1 1-0

4. Nogueiras - Velimirovic Reggio Emilia, 1986 1 d4 c5 2 d5 g6 3 e4 .ig7 4 c4 d6 5 ttJc3 ttJf6 6 f4 0-0 7 lLlf3 e6 8 .ie2 exd5 9 cxd5 .ig4 10 0-0 lLlbd7 11 h3.ixf3 12 .ixf3 a6 13 g4 lLle8 14 g5 lLlc7

In the "Step by Step" chapter I recommend 15 a4, while I also consider 15 h4. Nogueiras, however, 39

Part 1 chose the thematic pawn sacrifice: 1S .ig4?! llJbS 16 eS llJxc3 17 bxc3 dxeS 18 fS

In such tactical, highly unbalanced positions, Velimirovich is in his element. He is sure to prevent the blockade on e4, even at the cost of a piece: 18 ... e4! 19 f6llJxf6 20 gxf6 .ixf6 21 .ih6

This game is a milestone in the theory of the 9 ...1Lg4 variation. It shows what a powerful weapon could be a black knight on d4. More importantly, we see that even with an extra rook White could not claim an advantage. The moral is that White should aim at the first place to restrict the enemy counterplay and not to win material. In this variation activity is very often more valuable than material.

S. Semkov - The Cookie Lover 3 min. Playchess.com 21.12.2008 1 d4 llJf6 2 c4 g6 3 llJc3 .ig7 4 e4 d6 S f4 0-0 6 llJf3 cS 7 dS e6 8 .ie2 exdS 9 cxdS .ig41 00-0 llJbd7 11 h3 .ixf3 12 .ixf3 l'!e8 13 g4 h6 14 Wfc2 a6 1S .ie3 bS 16l'!ae1 b4 21 ... ~d6!?

I remember that this move has made a strong impression on all the commentators. It reminds me a good bluff in poker - everybody took Velimirovic's word and rejected 22 .bf8 Elxf8 23 1Wd21Lg7 24 ~g2 fS 2S 1Le2 bSoo. This position is extremely tangled indeed, but it may be slightly favourable for White. The simple 21...1Lg7! 22 .bg7 ~xg7 23 d6!? fS 24 1WdS 1Wh4+ would have underlined the weakness of White's king. After the text, the tension quickly recedes to simplification and a drawish endgame: 22 Elxf6 ~xf6 23 ~d2 l'!ad8 24 l'!d1 l'!fe8 2S d6 e3 26 .ixe3 l'!xd6 27 ~xd6 ~xd6 28 l'!xd6 l'!xe3 29 l'!b6 l'!xc3 30 l'!xb7 l'!c1+ 31 'it>f2 l'!c2+ 32 Ii:?g3 l'!xa2 33 l'!c7 %-%

40

This move order, and especially Black's next attack, is not best. It allows White to skip h4 and switch to the plan with e4-eS. That becomes possible because the black queen released its control on gS. 17 llJd1 ~aS 18 b3 l'!ac8

19 h4

Black's 17th move enabled the breakthrough 19 eS dxeS 20 gS hxgS

5 f4 c5 6 d5 0-0 7 lLlf3 e6 8 i.e2 exd5 9 cxd5 (20 ... lLlh7 21 f5~) 21 fxg5 lLlh7 22 lLlf2 lLlb6 23 d6± lLlf8 24 lLle4. However the text is also good and it leads to a critical position which White could not avoid had Black started with 16 .. J:k8 19 ... lLlb6 19 ... c4!? 20 bxc4 Vfic7 opens files so it deserves attention. Still, White's attack is very strong and easy to conduct, for instance: 21 g5 hxg5 22 hxg5 lLlh5 23 Vfih2 Vfixc4 24 lLlf2 i.c3 25 i.e2!± Vfic7 26 l"i:el Wg7 27 f5 l"i:h8 28 lLlg4 lLle5 29 lLlh6. The text is however more consistent. 20 g5 hxg5 21 hxg5 lLlfd7 22 e5 dxe5 23 f5 e4

The following combination took me only 7 seconds (it was a 3" blitz game!) - Black's king is so exposed that I did not need long calculations. 24 fxg6! exf3 25 gxf7+ xf7 26 §'h7!! lLle5 It turns out that things are not that simple and I have to part with more material, but the rest is still enough for winning! 27 ~f4! f2+! 28 ~xf2 lLlf3+ 29 ~xf3 ~xe1+ 30 Ii?f2+I like very much this position.

It shows well the essence of White's play in this variation. We should aim to destroy the pawn shield of the enemy king, then tactics will work to our benefit. 30 ... ~xa2+ 31 Ii?xe1 ~e8+ 32 ~e3+ e7 33 ~xg7+ d8 34 g6 ~d7 35 ~f7 ~xb3 36 g7 lLle5 37 ~xe8+ xe8 38 g8~+ d7 39 ~e6+ c7 40 ~xe5+ b6 41 ~d6+ b5 42 ~d7+ Ii?c4 43 ~g4+ b5 44 ~d2 ~c4 45 ~xc4+ xc4 46 d6 1-0

6. Roepert - Marin Budapest, 1988 1 d4 ~f6 2 c4 c5 3 d5 d6 4 ~c3 g6 5 e4 i.g7 6 f4 0-0 7 lLlf3 e6 8 ~e2 exd5 9 cxd5 ~g4 10 0-0 ~ bd7 11 h3 ~xf312 ~xf3 ~e8 13 ~c2?! White had to insert 13 g4 h6 since now Black can save ...h6 thus avoiding weaknesses. 13 ... c414 g4 ~c5 15 g5 ~fd7 Black has reached the maximum in this variation. He accommodated both his knights on perfect positions in the centre and gained space on the queenside. On the next move he will even push ... b5 without having to prepare it with ... a6. It is all 41

Part 1 the more curious to see that even in these ideal circumstances Black's game is not that easy. Thanks to his space advantage, White retains powerful threats. 16 lLld1 b5 17 lLlf2 h6?! Marin had an excellent feeling of the KID, as I had the chance to experience myself, (on the receiving end) but this move is dubious to me since it unnecessarily opens the h-file and weakens g6. White would push h5 anyway, so why give him tempi?! Interestingly, Rybka also likes it. 18 h4 hxg5 19 hxg5 gc8 20 i,e3

20 ... f6?! Perhaps something like 20 ... ltJd3 21ltJxd3 cxd3 22 ~xd3 ltJc5 would have been more reasonable. I suppose, Marin overestimated his position. He obviously thought that it was time to get down to business and launch a crushing counterattack. His 17th move and now 20 ... f6 confirm that. At the same time, White still has superior forces on the kingside! He should remember, that retreating to defence in this position is fateful. Instead, 21 42

e5! would have faced Black with a hard choice. For instance, 21...fxe5 22 ~xg6 exf4 23 lLlg4 ltJe5 24 ltJxe5 dxe5 25 hc5 ~xc5 26 b4 cxb3 27 axb3 ~d7 28 ~xa7 ~xa7 29 ~xe8+ .tf8 30 Wg2± would be rather unpleasant to him. Perhaps 20 .. .f5 would be the lesser evil, but very few players would go for it after an attempt to open up play on the previous move. 21 lLlh3? This is awful. Never play like this! It is better to sacrifice this knight rather than putting it aside, even temporary. Now Black's approach triumphs. 21 ... lLld3 22 gxf6 lLlxf6 23 lLlg5 lLlxd5 24 exd5 gxe3 25 \!;Vh2 \!;Vf6 26 i,g4 \!;Vd4 27 gf2 gee8 28 i,e6+ g3xe6 0-1

7. Semkov - Z. llie Saint John, 1988 1 d4 lLlf6 2 e4 e5 3 d5 g6 4 lLle3 i,g7 5 e4 d6 6 f4 0-0 7 lLlf3 e6 8 i,e2 exd5 9 exd5 i,g4 10 0-0 lLlbd7 11 h3 i,xf3 12 i,xf3 a6 13 g4 lLle8 14 g5 b5

15 lLle2

5 f4 c5 6 d5 0-0 7 CDf3 e6 8 ~e2 exd5 9 cxd5 I still like this idea. My course of thinking was roughly the following: Black wants to play ... c4. Of course I will meet it with ~e3, but then he will have .. J'k8 or 'Wic7 (15 h4 c4 16 ~e3 Wic717 h5 CDc5 18 'Wic2 CDd3 19 hxg6 hxg6 20 ~e2 'Wid7 21 f5 is worth attention though). In that case he will benefit from the delay of ... CDc7. I hate to let a black knight to d3, so the only way to discourage ... c4 must be CDe2. Then 15 ... c4 16 CDd4 CDc5 17 ~e3 CDd3 18 'Wid2;!:; will be a pleasant trade off since the d4-square in combination with the pawn centre should ensure an edge. If Black renounces the plan with ... c4, White will proceed with his kingside attack by pushing the hpawn. The e2 knight will turn handy on that wing. Perhaps Ilic had the same feeling as he decided to attack my centre by 15 .. .f5?! I can add now that 15 ... CDc7 was probably the best try. White responds with 16 'Wic2, defending b2 and preparing to develop the bishop on e3 or d2. The computer suggests 16 ...b4 17 h4 CDb5 18 h5 CDa3 19 Wid1 ge8, but 20 hxg6 hxg6 21 f5! gxf5

22 CDg3! (22 exf5 hb2 23 hb2

'Wixg5+ 24 ~f2 'Wie3+ 25 ~g3 'Wig5+ 26 ~f2 is only a perpetual) 22 ..fxe4 (22 .. .f4 23 ~h5 CDc4 24 'Wie2±) 23 he4 CDf8 24 'Wig4 'Wie7 25 ~f4 hb2 26 gael± should be convincing. 15 ... f5?! 16 exf5 gxf5 Black allows my bishop to e6. My practice shows that Black rarely survives that. He should have captured by pawn: 16 ... gxf5 I like then 17 gb1!? (17 'Wic2 CDb6 or 17 h4 CDc7 18 h5 CD b6 19 h6 ~h8 are not too clear) 17... CDb6 18 b3. This is not to trade the dark-squared bishops from b2, but to restricts the black knight. White needs that bishop to cover the invasion squares on the e-file: 18 ... CDc7 19 ~e3 Wie7 20 'Wid2 'Wif7 21 gbd1 gae8 22 CDg3 CDbxd5 23 ~f2 with an initiative. 17 i.g4 gf8 18 ~h1?! 18 ~e6+ ~h8 19 h4--+ CDc7 20 h5 was very strong. 18 ... CDC7 19 .id2!?

If Black accepts the gift on b2, my bishop pair will terrorise the enemy knights from the both edges of the board. The variation 19 ... hb2 20~a5gb8 21~e6+ ~h8 22 gb1~g7 23 f5 gxf5 24 CDf4--+ looks dreadful, so Ilic declined the sac: 43

Part 1 19 ... b4! 20 ~b3 i>hS 21 ~ad1 as A critical moment. Black should have tried 21...ttJb6 22 ~c1 \We8 23 h4 as although White's initiative would still be tangible. Now the hfile turns to be the decisive factor in the game. 22 ic1 :!'!a7 23 h4 ttJb6 24 h5 gxh5 25 ixh5 ~d7 26 i>g2 a4 27 ~f3 ttJc4 2S :!'!h1 i>gS 29 ig4 ~e7 30 ~d3+- ixb2 31 ~xc4 ihS 32 ttJg3 :!'!eS 33 liJf5 ~e4+ 34 ~xe4 ~xe4 35liJxd6 :!'!d4 36 liJcS :!'!b7 37 d6 1-0

S. Blokh - Morozevich Moscow, 1992 1 d4 liJf6 2 c4 g6 3 liJc3 ig7 4 e4 d6 5 f4 0-0 6 liJf3 c5 7 d5 e6 S ie2 exd5 9 cxd5 ig4 10 0-0 liJbd7 11 h3 ixf3 12 ixf3 c4 13 ie3 :!'!eS 14 g4 I also consider 14 \Wa4 in the "Step by Step" chapter. 14 ... h6 15 h4

More solid is 15 \Wc2 b5 16 :!'!ae1 transposing to "Step by Step" line Cl. The text is double-edged as it allows 15 ...h516 g5 ttJg417 hg4 hxg4. 44

This position is a better version for White of the famous game Kouatly-Kindermann, but the same strategic ideas still work: 18 \Wxg4 hc3 19 bxc3 :!'!xe4 20 ~d4 \We7 21 \Wh3 (or 21 h5 ttJe5 22 \Wg2 :!'!xd4 23 fxe5 :!'!d3 24 hxg6 \Wxe5=) 21...:!'!e8 22 h5 ttJe5 23 hxg6 ttJxg6oo. Such course of events does not comply with my approach to the variation. Instead, I would rather sacrifice material: 18 e5!? dxe5 19 f5 gxf5 20 :!'!xf5 e4 21 \Wxg4 ttJe5 22 :!'!xe5!! :!'!xe5 (22 ... he5 23 ttJxe4~) 23 h5

You should be getting used to such positions if you have read the previous pages. The dynamic factors favour White. Main threat now is 24 h6. Black cannot buy off himself with 23 ... :!'!xd5 24 ttJxd5 \Wxd5 since 25 :!'!d1 \Wc6 26 g6± continues the attack. 23 ... \Wc8 removes the hit from d5 and also favours White: 24 \Wg2 :!'!e7 25 :!'!fl--+. Perhaps best is 23 ... :!'!e7 24 h6! (24 ttJxe4? stumbles into 24 ... \Wd7! (24 .. hb2 25 :!,!b1~) 25 \Wg2 \wf5 26 ttJf6+ ~h8 27 ~d4 \Wd3+) 24 ... ~e5 25 ttJxe4. I would always take White in such positions. Morozhevich chose the much more passive:

5 f4 c5 6 d5 0-0 7 tLlf3 e6 8 ie2 exd5 9 cxd5 1S ... tLlh7 16 gS hxgS 17 hxgS ~aS

18.id4? I cannot understand how a FPA player can even think about trading queens at this moment! He should have followed up with 18 cj;>g2! tLlc5 when 19 id4 hd4 20 vtIxd4 would already be good. Black still would be able to swap queens by 20 ... tLld3 21 ie2 vtIc5, but White would have all the files and diagonals already open to his favour: 22 vtIxc5 tLlxc5 23 hc4 tLlxe4 24 ib5 'ge7 25 'gael tLlxc3 26 2:xe7 tLlxb5 27 'gxb7 tLld4 28 'gcl±. 18 ....ixd4+ 19 ~xd4 ~eS 20 E:ad1 a6 21 .ig4 tLlhf8 22 'gfe1 bS 23 @g2 ~xd4 24 'gxd4 tLleS

The Modern Benoni endgame

is commonly easier for Black since he has a clear plan to expand on the queenside while White's breakthrough e5 is seldom efficient. On his next move Morozhevich allows White to trade his bishop for a knight, which was hardly a good idea. (25 .. J'l:b8!? instead) 2S @f3 tLlfd7 26 .ixd7 lDxd7 27 eS?! Now Black definitely takes the upper hand and confidently converts his initiative into a full point. 27 ... dxeS 28 'gde4@g7! 29 fxeS lDeS 30 'g4e3 'gh8 31 lDe4 lDd3 32 'gxd3 'gh3+ 33 lDg3 exd3 34 'gd1 'ge8 3S 'gxd3 'gxeS 36 d6 'gxgS 0-1 9. Banikas - Pigusov Komotini,1993 1 d4 lDf6 2 e4 eS 3 dS e6 4 lDe3 exdS 5 exd5 d6 6 e4 g6 7 f4 .ig7 8 lDf3 0-0 9 .ie2 .ig4 10 0-0 lDbd7 11 h3 .ixf3 12 .ixf3 a6 13 a4 'ge8 14 g4 lDf8 15 g5 lD6d7 16 .ie3 'ge8 17 @g2 ~a5 18 We2 e4 19 .ig4 'ged8

I do not like a4 in general as it spends a tempo and weakens the b3 and b4-squares. Still, White is

45

Part 1 harmoniously developed and now he must choose an active plan. The thematic 20 eSlooks strong enough. Then 20 ... 1t1cS would be an obvious mistake due to 21 ,bcS ~xcS 221t1e4 ~xdS 23 1:!adl± so Black must accept the sacrifice: 20 ... dxeS 21 fSt. I would not like to be in his shoes though. Instead, White opts for the awful move:

22 axbS ~xbS 23 1:!a4;!;. 22 Wxe4 gbS 23 ge1 tDe5 24 e5!

20 tDd1?

A really awful move! I'm sure White calculated some very deep variations to even start pondering it, but he apparently missed something. The whole idea to eat a pawn instead of staking on activity is fundamentally wrong in the FPA. Here it does not even work. It is a mistery to me why Black did not answer 20 ... 1t1cS! 211t1f2 (21 if3 ltlb3) 21...~b4 22 as tDfd7 with excellent counterplay. 20 ... b5? 21 .td2! b4 It seems that Black missed 21

id2 and panicked. Perhaps 21...~b6 would have been more stubborn:

46

24 ... tDxa4 25 exd6 tDb6 26 Wxb4 Wa2 27 WaS tDa4 2S d7 gedS 29 geS f5 30 gxdS gxdS 31 WxdS Wxd5+ 32 .if3 Wxd2+ 33 gf2 Wxd7 34 .id5+ ~hS 35 Wxd7 tDxd7 36 .ie6 tDdb6 37 ge2 .ifS 3S J.xa4 tDxa4 39 geS ~gS 40 gaS tDe5 41 ~f3 ~f7 42 tDf2 .te7 43 ga7 ~fS 44 h4 ~f7 45 b4 tDe6 46 tDd3 tDd4+ 47 ~e3 tDe2+ 4S ~d2 tDd4 49 gxa6 ~eS 50 ~e3 tDb5+ 51 ~e4 tDe7 52 ga7 1-0

Part 2

1 d4 tiJf6 2 c4 g6 3 tiJc3ilg7 4 e4 d6 5 f4 cS 6 dS 0-0 7 tiJf3 e6 8 ie2 exdS 9 cxdS ~e8

QUICK REPERTOIRE

This variation is the critical test of the FPA. Anyone who advocates the system for White must propose a decent alternative to the known drawing variations after 10 eS dxeS 11 fxeS lLlg4 12 igS IWb6! 13 0-0 lLlxeS. Unfortunately, I have not seen so far anything sensible neither in practice, nor in print. The ideas with long castling are worthless, while Jobava's 12 e6! fxe6 13 igS!? may be interesting, but I fail to grasp completely White's intention. You can find a survey of White's problems in the well known lines in the "Step by Step" chapter. This whole part is based solely on my original analysis, so do not wonder about the absence of references to other games. The reason to write this book is that I want to propose a new approach to this variation. It has at least several advantages: 1. It is unexplored; 2. It has a very sound positional foundation so White is not exces-

sively dependent on long forcing variations. In all cases he can count on certain compensation for the sacrificed pawn. 3. I have extensively checked my analysis with last versions of popular engines so Black's attempt to find the best defence by simply switching on the computer should not bring an easy success. The second players will have to navigate in a maze of false trails because the arising positions are strategically very unbalanced and the engines commonly fail to evaluate them correctly, at least at some reasonable depth of 16-20 half-moves. I should add that most of the variations are not forceful and both sides have plenty of other possibilities. I chose as main lines either the most logical continuations, or the preferred choice of the engines which served me as an opponent during my analysis. 4. I have tested my ideas in blitz and I must say that commonly Black totally fails to cope with his defensive task. 47

Part 2 So we go on with: 10 eS dxeS 11 fxeS tDg4 12 e6! fxe6 13 d6!

d6-pawn and allows White to win material. The c8-bishop cannot be easily brought into play. If it goes to c6, then e6 becomes vulnerable.

Key lines: 1. 13 .. :~b6 14 etJg5! etJh6 15 0-0

etJc6 (or 15 ... etJf5 16 Elxf5!!) 16 etJge4 etJf5 17 ig5 etJcd4 18 ~a4

I'll try to explain the most important features of this structure firstly in words, then in concrete examples. Black's kings ide is weakened. If we trade dark-squared bishops, the f6 and h6 squares would be very sensitive. In some lines we can even sacrifice the exchange in order to kill the g7-bishop. My practical advice: if you are unsure what to play, look for ideas which would help you trade the bishops. This strategy could be extended to the king's knights, too. The reason for that is simple - the d6 pawn significantly cramps Black and impedes the relocation of his queenside pieces to the other wing. White's space advantage ensures greater mobility to his pieces so he can easily change the direction of his attack. For instance, etJc3-b5-c7 is a major threat. Sometimes White also hits c5, or transfers the queen to h4 via a4. In many cases exchange of queens enhances the power of the 48

Black covered solidly his kingside but suddenly the d6-pawn is about to leap forward. It can be sacrificed to eliminate the g7-bishop: 18 ... etJc6 19 etJf6+! ixf6 20 ixf6 etJxd6 21 ~h4l"1f8 22 ih5!!± The threat of taking on g6 is impossible to deflect. 2. 13 ... ~b6 14 etJg5! etJe5 15 0-0 etJbd716 a4!

This pawn move threatens with etJb5, but more importantly, it en-

S f4 cS 6 dS 0-0 7 ltJf3 e6 8 ~e2 exdS 9 cxdS ~e8 abIes the rook lift ~al-a3-h3. So Black answers 16 ... c4+ 17 @hl ~c6 (planning ... b6) 18 ~d4 (hoping to reach h4) 18 ... h6 19 ttJge4 b6

20 ttJf6+! ttJxf6 21 ~xeS ttJh5 22 Note that White does not object to enter an endgame due to his very active pieces. ~bS±.

3. 13 ... ttJeS 14 0-0 ttJbc6 IS CLlxeS Simple and consistent. We aim to trade dark-squared bishops. IS ... heS 16 ~f4 hf4 17 ~xf4 eS 18 ~f2 ttJd4 19 ~a4

4. 13 ... ttJc6 14 a-a! ttJd4 15 ttJe4

White will be fighting for the f6square with ~gS. Now IS ...b6? loses immediately to 16 ~g5 ~d7 17 ttJxd4 hd4+ 18 ~xd4!+-, 15 ... ttJf6? is bad owing to 16 ~gS±. More resilient are: a) 15 ... l"1f8 16 ~gS ~b6 17 ttJfd2! with good control over the critical dark squares; b) 15...h616 ttJxd4! hd4+ 17@hl

The subsequent tactical interchange is in White's favour: 17... ttJxh218 l"1f4!± gS19 l"1f2 hf2 20 ttJxf2 l"1f8 21 ~d3! The d6-pawn makes the difference in this position: 19 ...~fS 20 ~bS ttJxbS 21 ttJxbS or 19 ... ~e6 20 ~c4 @h8 21 he6 l"1xe6 22 ttJe4 leave White with the initiative.

5. 13 ... ~d714 0-0 ~b615 ~el Whenever Black decides to win the d6-pawn by ... ~b6, we should aim to transfer our queen to h4. We have seen in key line 1 the manoeuvre ~a4. Here we shall consider

49

Part 2 another route for the queen. IS ... CDc6 16 'lWh4 CDceS 17 CDxeS CDxeS 18 CDe4

We see White's strategy in its purest form - the aim is to conquer the dark squares. Black is unable to prevent that: 18 ... 'lWb4 19 .tgS'lWxb2 20 .tf6 l"If8 21 l"Iael! or 18 ... l"If8 19 .th6 'lWxb2 20 hg7 ~xg7 21 'lWe7+ ~g8 22 CDgS with initiative.

6. 13 ....td714 0-0 .tc6 IS CDgS! This move is good only here and in the case of key line 1. White must play concretely or he risks to lose the initiative. IS ... CDeS 16 .te3 b6

Recommendation of the ECO, vol. A, 4th edition. Black keeps his knight on b8 in order to meet 17 CDbS by 17... CDa6. 17'IWd2 ECO only mentions 17 CDge4?!.

SO

17 CDbS!? also deserves consideration. 17... CDbd7 Black protected everything, but his position is passive. We should seek targets on the queenside 18 CDbS h6 19 CDh3 CDg4 20 .tf4 gS Or 20 ... hbS 21 hbSt. 21 CDxgS! hxgS 22 hgS CDgf6 23 'lWf4. White maintains pressure on the kingside.

Let me summarize now: Black has a vast choice in the position after White's 13 d6. It is not easy to say which is his best option. In any case White preserves sufficient positional compensation for the pawn even in an endgame. He can complete development and find a target depending on the enemy's play. It would be nice to put the queen on h4 via el or a4, but CDbS and a piece attack on the queenside are not that impossible, too. It is important to remember that our advantage consists in the greater mobility of our pieces. We do not need immediate tactical achievements since our pawn on d6 is awkward enough to the opponent. Note the two cases when we must make an exception to this rule and play CDgS! early - in key lines 1 and 6. I perfectly realise how complex this variation is. It needs many more days of analysis and fresh contribution from other players. But I am sure that Black will have a hard time when facing this variation unprepared.

Part 2

1 d4 liJf6 2 c4 g6 3 liJc3 .ig7 4 e4 d6 5 f4 cS 6 dS 0-0 7 liJf3 e6 8 ie2 exdS 9 cxdS 1'!e8 STEP BY STEP

10eS

White has no real alternatives to this breakthrough. I think it is the only way to fight for an advantage. I must confess that I sinned more than once in blitz with the incorrect sac 10 O-O? tLlxe4 11 tLlxe4 E1xe4 12 1d3 E1e8 13 f5, but it counts only on the cheapo 13 ... tLld7 (13 ...ixf5+) 14 tLlg5 tLle5 15 tLlxf7 tLlxf7 16 fxg6 tLle5 17 ~h5 hxg6 18ixg6 tLlxg6 19 ~xg6 1''lf8?? 20 1h6+-. Instead, 19 ... E1e7 20 1d2 ~e8+ or 19 ... E1e5+ repel the attack. 10 tLld2 is much more solid, but nowadays it is considered innocuous. Perhaps simplest is to develop the knight to a6 - 1Q ... tLla6 11 0-0 tLlc7 12 a4 b6 13 E1e11a6 14 E1a3 ixe2 15 E1xe2=, Spassky-Marin, France, 1991. 10... dxeS

I was surprised to discover that the seemingly stupid 1Q ... tLlfd7 is not that bad at all. The only certain thing now is that 11 e6? fxe612 dxe6 tLlf6+ leaves White over-extended.

The allegedly best 11 exd6?! achieves the same result after 11 ... a612 a4 (120-0 b5+) 12 ... tLlf613 0-0 ~xd6! (A Chess Base source mentions only 13 ...1g414 ~b3 b6 15 h3 ixf316ixf3 ~xd6171d2±) 14 tLle5 ~d8! 15 1f3 tLlbd7 16 E1e1 tLlxe5 17 fxe5 tLld718 e6 tLle5. Now White has to take on f7 and think about maintaining the balance, e.g. 19 exf7+ tLlxf7 20 E1xe8+ ~xe8 211f41d4+ 22 cJrh1 tLle5 23 d6 tLlxf3 24 ~xf3 1d7 25 ~xb7 E1b8 26 ~d5+ ~f7 27 ~xf7+ cJrxf7 28 E1e1=. My attempts to improve on this variation proved in vain. Eventually I understood that Black's tender point in the diagram 51

Part 2 position is t7 and we should attack it by: 11 O-O! dxeS 12 ttJe4! 12 fxeS ttJxeS 13 if4+ would transpose to a popular position where White has not sufficient compensation in my opinion.

The first thing I considered here was: a) 12 ... exf4 13 ttJd6 Ele7 14 ixf4 ttJb6 (14 ... ixb2 IS igS±) IS ttJgS. White has full compensation and making a couple of natural moves suffices to convince the engines: lS ... if616 ttJge4 id4+ 17 ~h1 fS 18 ttJc3 ttJa6 19 ttJdbSt. Black's king is rather vulnerable. lS ... id4+ 16 ~h1 f6 17 ttJxc8 '.Wxc8 18 ttJf3 Ele4 (Alternatively, 18 ...ie3 19 d6 Ele6 20 ixe3 Elxe3 21 ~d2 Ele6 22 ttJgS Elxe2 23 ~xe2 fxgS 24 '.We7 ttJc6 2S ~t7+ ~h8 26 Elae1 ttJd7 27 ~e6 ttJd4 28 ~dS ttJfS 29 ElxfS gxfS 30 '.WxfS±; 18 ... ixb219 Elb1 ic3 20 d6 Ele8 21 Elxb6 axb6 22 ibS± and the threat of 23 ~b3+ regains some material.) 19 ih6 (It is also tempting to eat the important dark-squared bishop by 19 ttJxd4 Elxd4 20 ~b3 ~g7 21 Elad1 ttJ8d7 22 ie3 Elxd1 23 Elxd1~) 19 ... ttJ8d7 20 id3 EleS 21 ttJxd4 cxd4 22 Elc1 ~d8 S2

23 if4 ElxdS 24 ie4t. Then I checked up the purely computer move: b) 12 ... Elf8. The trick is that 13 fxeS ttJxeS 14 igS ~b6 IS ie7 would face lS ... ttJbd7! with unclear play. Other continuations give White some initiative: lS ... ttJxf3+ 16 Elxf3 ~xb2 17 ttJf6+±; lS ... ~xb2!? 16 ttJf6+ (16 ttJfgS!?) 16 ... ixf6 17 ixf6 ttJbd7 18 Elb1 (18 ttJxeS ttJxeS19 Elb1 ~d4+ 20 ~xd4 cxd4 21 ixeS Ele8 22 ixd4 Elxe2 23 Elfe1 Elxe1 + 24 Elxe1 id7 2S Ele7 Eld8 26 ~f2t. This endgame should be a draw, of course.) 18 ... ~a3 19 Elb3 ttJxf3+ 20 Elbxf3 ~xa2 21 ie7 ttJeS 22 Ele3 ttJc4 23 ixc4 ~xc4 24 ixf8 ~xf8 2S Elef3 fS 26 Ele1oo. Anyway, my impression was that Black's defence was too often based on checks along the gl-a7 diagonal, so I focused on: 13 ~h1!? exf4 14 ixf4 ttJf6 (14 ... ixb2 IS igS~) IS ttJxf6+ ixf6 16 Elc1~ ttJd7 17 ~d2 ~b6 18 b3 Ele8 19 ic4t

It is not obvious how Black can disentangle his pieces. His immediate concern should be how to complete development. The only reasonable move in that direction is

5 f4 c5 6 d5 0-0 7 lLlf3 e6 8 .ie2 exd5 9 cxd5 !%e8 19 ... lLlf8, (since 19 ... lLleS? loses to 20 WTe3) when 20 .igS .ig7 21 WTf2 admonishes about the weakness of t7, e.g. 21.. ..id7 22 lLleS! EixeS 23 ~xt7+ mh8 24 .if6 .ixf6 2S Eixf6 ~d8 26 d6±. As a whole, my proposition of 11 O-O! dxeS 12 lLle4! leads to a strong White's initiative without excessive risk. I should also mention 10 ... lLlg4 11 h3 lLlh6. Now best seems the restrained development by 12 0-0 (or 12lLle4 first) 12 ... lLlfS (or 12 ... dxeS13 fxeSlLlfS 14lLle4lLld71S e6) 13 lLle4 dxeS 14 fxeSlLld71S e6 fxe616 dxe6 with an edge, for instance, 16 ... lLlb6 17 .ibS .ixe618 .ixe8;!; WTxe819lLlxcS idS 20 Eie1 or 16 ... lLleS 17 WTxd8 ~xd8 18 lLlxeS .ixeS 19 e7.id4+ 20 ilh2 Eie8 21.igS±.

11 fxeS lLlg4 11...lLlfd7 is already dubious because White's pawn arrives at e7: 12 e6 fxe6 13 dxe6lLlb6 14WTxd8 Eixd8 IS lLlbSlLla6 16 e7 Eie8 17 .igS h6 18 .'2ld6 .id719lLlxe8 Eixe8 20 .ie3 Eixe7 21 mf2 lLla4 22 Eiad1±.

It took me too many years to ripen for this move. The next pages are irrelevant to the proposed repertoire, but they might be of interest to Black players who can face the widespread 12 .ig5. When I started playing the FPA more than 20 years ago, I believed firmly in this move. I had a lot of original analyses and it brought me many memorable victories over strong opponents. However, two ideas completely destroyed my favourite repertoire. They remain milestones in the theory of the FPA as White has not found anything to revive the variation: Neurohr-Schlosser, St. Ingbert 1988: 12 ...WTb6 13 0-0 lLlxeS 14 d6 lLlxf3+ IS .ixf3 WTxb2 16 lLldS .id4+ 17 mh1 WTxal 18 WTxal .ixal19 Eixa1 EieS!! and Black is somewhat better; Semkov-Marin, Berga, 1990 14lLlxeS .ixeS1S .ic4WTxb2 16 d6 Eif8! 17.ixt7+ mg7!+.

White has many other options, of course, but they all lead to a forced draw as best. For many years I could not understand how White players still persisted with playing this line while it was obviously dead. I myself switched to dxcS early in the opening. Then Golubev presented me with his very interesting book

Understanding the King's Indian, Gambit 2006, where I was really stunned to read that 12 .igS WTb6 13 12 e6!

WTd2, intending long castling, was rather dangerous to Black!! (Some research on this subject pointed out S3

Part 2 .ic4 (or IS d6 .ie616 ~hl c417 ttJe4 ttJd7+) IS ...'lWb4 16 'lWb3 .ifS 17 d6 'lWxb3 18 axb3 .ixd6 19 ttJdS is not a playable option either. Only in the most optimistic frame of mind I would assess it as roughly equal. Anyway, we cannot speak about any White's advantage. I was not going to give up easily and began looking for new ways. Obviously 12 ...'lWaS takes the sting of the whole idea with d6 and ttJdS. However, White obtains other options, as 13 e6! (Formally, this is not a novelty as it had already been played in the game Hartmann-Erwes, Germany 1994, but White connected it with a wrong idea - 13 ... Now the hint about castling fxe614 O-O?!) 13 ...fxe614 d6 (14 h3 queenside does not help since the ttJf6!) enemy queen is much more dangerous than it was from b6, and it does not block the advance of the b7-pawn. On the other hand, the b2 pawn is not hanging, so 13 0-0 should be in favourable circumstances compared to the 12 ... 'lWb6 line. Detailed analysis did not confirm my hopes though. After 13 ... ttJxeS White usually chooses The more I analysed this posi14 d6, but then 14 ... ttJbc6! (I'd be glad to take Konikowski's word that tion, the more I liked it for White. "White has the better prospects" afI shall explain the fine points ter 14 d6, but my own inner voice of this pawn structure later in this tells me just the opposite.) IS ttJdS chapter. White's play is linked with c4! turned to be extremely unpleas- the vulnerable dark squares in ant, e.g. 16 ttJf6+ .ixf6 17.ixf6 ttJg4. Black's camp and the hanging state (By the way, Konikowski does not of the g4-knight. Of course, the d6consider 14... ttJbc6 at all. I suspect pawn is also lying in ambush only that it is not the only good option of two squares away from queening. Black though.) Perhaps Black should activate his The alternative 14 ttJxeS .ixeS IS queen by:

to a 200S Chess Base opening survey by Konikowski.) And off I went, myoId love coming back at full bloom. I started playing blitz games to test Golubev's recommendation, but soon discovered that most of my opponents answered 12 .igS with 12 ... 'lWaS.

S4

5 f4 c5 6 d5 0-0 7 l2lf3 e6 8 ie2 exd5 9 cxd5 l"l:e8 14 ... c4 Alternatives are 14 ... id7 15 0-0 c4 16Wd2 CLJc6 (16 ... l"l:f8 17 @h1CLJc6 18 h3 CLJge5 19 ~h6 Ei:ac8 20 Ei:ae1 Ei:f5 21CLJxe5 Ei:xf1+ 22 Ei:xf1Wxe5 23 hg7 @xg7 24 hc4 CLJa5 25 ~d3t) 17 CLJe4! Wxd2 (17 ... Wf5 18 Wc2) 18 CLJfxd2 regaining the pawn with a plus; 14 ... CLJd7 15 O-O! (15 h3 CLJge5 160-0 CLJxf3+ 17 l"l:xf3 c418Wd2CLJe5 19 l"l:f4 ~d7 20 l"l:afl l"l:f8) 15 ... c4 16 @h1 CLJb6 17 We1 ~d7 (17 ... Wb4 18 CLJd2±) 18 Wh4± CLJe5 19 CLJe4. 15Wd2 CLJc6 The seemingly dumb 15 ... CLJd7 might be better. Then 16 hc4CLJde5 17 CLJxe5 Wxe5+ 18 ~e2 l"l:f8! 19 d7 hd7 20 Wxd7 CLJe3 is rather risky for White, so he should better develop and play all over the board: 16 0-0 Wc5+ 17 @h1 b5 18 h3CLJge5 19 a4 bxa4 20 l"l:xa4 ~b7 21CLJxe5 CLJxe5 22d7t. 16 0-0 ~d7

~c6

is satisfactory for him. 17 @h1 is more consistent, but the forced play 17... CLJge518CLJxe5CLJxe519 ~h6 hh6 20 Wxh6 ~c6 21 We3 b6 22 b4 cxb3 23 axb3 Wc5 24 Wxc5 bxc5 25 Ei:a5 CLJf7 (25 ... CLJd7 26 Ei:a6 ~b7 27 Ei:aa1 Ei:f8 28 ~c4) 26 CLJb5 Ei:ed8 27 CLJc7 Ei:ab8 28 Ei:xa7 CLJxd6 29 CLJxe6 Ei:e8 30 ~c4 @h8 31 h4CLJxc4 32 Ei:ff7= leads only to a draw. The text allows White to conquer the f6-square: 17... Wd5 Alternatively: 17... Wb6+ 18 @h1 Wxb2 19 Ei:ab1 Wxd2 20 CLJfxd2±; 17... Wxd2 18 CLJfxd2±; 17... Wf5 18 Wc2 Ei:f8 (Or 18 ... Wa519 Wxc4 Wb6+ 20 @h1CLJe3 21 he3 Wxe3 22 CLJc5) 19 @h1 retaining a lot ofthreats. 18 Wf4 Wf5 (18 ... CLJd419CLJxd4±) 19CLJfd2 Wxf4 20 Ei:xf4CLJe3 21CLJf6+ hf6 22 Ei:xf6t. I was happy with my find 13 e6! to 12 ... Wa5 and was getting more confident about the whole system, when trouble came, as usual, from the main line with 12 ... Wb6. Of course 13 Wd2 looked quite fresh, but gradually I began to realise that White's play had not a stable background and only counted on some tactical tricks. Look at the position after 13 ... CLJxe5 14 0-0-0:

17CLJe4! A very important resource which

exploits the placement of Black's knight on g4. That's why moves like 17 h3 should be made only by concrete reasons. In our case Black would have welcomed it as 17 ... CLJge5 18 CLJxe5CLJxe5 19 ~e7 Wc5+ 20 @h1 55

Part 2 Black has tried here probably a dozen of moves but somehow White managed to generate threats. The whole question is who would be faster. Black will put his bishop on fS and White's king will be in a critical situation, if it had not been for the resource g2-g4 which repels the dangerous attacker. However, Black has 14 .. .lt:lxf3! and White's design is completely destroyed. Then 15 gxf3 .tfS (lS ....txc3? 16 W1xc3 Elxe2 17 Elhel±) 16 Elhe1 ttJd7=t or 15 .txf3 ttJd7 16 d6 ttJeS 17 ttJdS ttJd3+! 18 ~b1 W1xd6 19 W1xd3 .tfS+' are rather grim for him. I attempted to get around this obstacle by changing the move order: 14 ttJxeS instead of castling, but 14 ....txeS 15 0-0-0 turned out to be not much better since Black had:

lS ....txc3 (It is another question that lS ... ttJd7 may be even better for Black: 16 Elhe1 .td4 17 ttJbS ttJeS 18 ttJxd4 cxd4 19 W1xd4 .tfS=t or 16 d6 ttJf6 17 Elhe1 .txc3) 16 W1xc3 Elxe2 17 Elhe1 (17 .th6 f6) 17....tg4 18 .th6 f6 19 Elxe2 (19 d6 W1xb2+ 20 W1xb2 Elxb2 21 Ele8+=) 19 ....txe2 20 W1h3 (20 W1e3 ~f7+) 20 ... W1d6 21 W1c8+ ~f7 22 Ele1 gS 23 Elxe2 W1f4+ with perpetual. Of course, this forced 56

draw is sufficient, albeit not exclusive, argument against 13 W1d2. I was already ripe for the thought of burying once again the FPA, but at that point it dawned on me that White could try a completely different, purely strategic approach. This brings us back to my proposed repertoire: 12...fxe6 13 d6

Apparently I'm not alone to think that White's position hides some venom as two strong grandmasters have recently tested the new move order 13 .tgS!?, and, notice, in very important team tournaments - the Olympiad and the Russian championship! Jobava-Bologan, Turin 2006, saw further 13 ... ttJf6 14 d6 h6 15 .txf6 W1xf6 16 ttJbS .td717 ttJc7 ttJc618 ttJxa8t. White came also better in Tarasova-Lomineishvili, EICC Plovdiv 2008, after 13 ....tf614 .txf6 ttJxf6 15 d6 ttJdS 16 ttJxd5 exd5 17 W1xd5+ .te618 ~xc5 ttJd719 ~d4t. The critical line 13 ... ~b6 14 ~d2 appeared in 1. Popov-Inarkiev, Dagomys 2008, when 14... c4 15 0-0-0 exd5 16 Elhe1 ttJc6 17 .txc4 .tf5 18 .txd5+ ~h8 19 ttJh4 was very tan-

5 f4 c5 6 d5 0-0 7 ttJf3 e6 8 ~e2 exd5 9 cxd5 l:l:e8 gled, but probably slightly better for White. I have no idea though what White had in mind against the computer's suggestion 14... hc31S bxc3 exdS. Perhaps we will learn the answer in future. In the diagram position I analysed: A.13 ... ~b6 B. 13 ... tiJeS C.13 ... tiJc6 D.13 ... ~d7 A.13 ••• ~b6

I start with this move, because it is absolutely critical for my idea. It bans 14 0-0 due to 14 ... c4+, and discourages the thematic development of the c1-bishop to gS since the b2pawn is hanging. White has not too much of a choice. At first I pondered over 14 ~a4 tiJf6 IS 0-0

with the following options: 22 ...b6 23 tiJe4; 22 ... E1f8 23 ttJdS; 22 ... tiJf6 23 E1dfl tiJdS 24 ~f2±) 22 tiJe4 E1f8 23 tiJfgS tiJxgS 24 tiJxgS tiJf6 2S ~c4. All the white pieces come into play to ensure a stable advantage after 2S ... a6 26 E1de1 bS 27 he6+ he6 28 ~xe6+ ~xe6 29 tiJxe6 E1f7 30 tiJxcSt. Black's problems sprang from the weakness of the e6-pawn. Therefore the knight should go not to d7, but to c6: lS ... ~xd6! 16 ttJe4 (Now 16 ~h4 tiJc6 17 ~h6 hh6 18 ~xh6 l:l:f8 19 tiJgS ~e7 is much less effective for White in comparison with the previous variation.) 16 ... ~e7 17 ~gS ~d7 18 ~b3 ~c6 19 tiJxf6+ hf6 20 ~c4 and Black has various ways to level the game as 20 ... tiJa6 (but not 20 ...~xgS?! 21 E1ae1!) 21 hf6 ~xf6 22 gad1 tiJc7 23 tiJd4 ~eS 24 tiJxc6 bxc6 2S gfel~ or 20 ... bS 21 hf6 ~xf6 22 hbS a6 23 ~c4 tiJd7. Remains: 14lDg5!

It seems that White's attack is running very smoothly on the dark squares: lS ... tiJbd7 16 ~h4 ~xd6 17 ~h6 hh6 (17 ... ~e7 18 tiJgS hh6 19 ~xh6 ~g7 20 ~h3~) 18 ~xh6 tiJg4 (Or 18 ... ~e7 19 tiJgS E1f8 20 gad1± tiJeS 21 tiJce4) 19 ~gS tiJgeS 20 gad1 ~e7 21 ~e3 tiJf7 (21...tiJxf3+ 22 E1xf3

14... lDh6 The knight is heading for fS from where it covers the f-file, attacks d6

S7

Part 2 and does not bar the g7-bishop. However, 14 ... ltJeS IS 0-0 ltJbd7 also looks possible. IS ... ltJbc6 does not control the f6-square so we should be able to rapidly destroy the enemy's castling position, for instance: 16 'lWelltJd4 (16 ... 'lWb417 a3) 17 'lWh4 hS (17 ... ltJxe2+ 18 ltJxe2 hS 19 ltJe4±) 18 i.xhS! ltJfS 19 'lWh3 gxhS 20 'lWxhS ~d7 21 'lWh7+ i>f8 22 ltJce4+-. Then only 16 a4! poses significant problems to Black.

Instead, 16 'lWeI c4+ 17 i>hl 'lWxd6 18 'lWh4 h6 19ltJge4 'lWe7 or 16ltJge4 c4+ 17 i>hl 'lWc618 'lWeI bS or 16 ~e3 'lWxb2 17ltJbS :1'&f8 18 :1'&xf8+ ltJxf8 are insufficient for obtaining an advantage. The pawn move is threatening with ltJbS, but more importantly, it enables the rook lift :1'&al-a3-h3 with a devastating effect: a) 16 ... a617 :1'&a3 :1'&b8 (17 ... c4+ 18 i>hlltJcS 19 ~e3 'lWc6 20 hc5 'lWxcS 21ltJce4 'lWdS 22 'lWbl bS 23 ~f3ltJxf3 24 :1'&axf3 ~b7 2S :1'&f7 :1'&f8 26 :1'&xf8+ :1'&xf8 27 :1'&xf8+ i>xf8 28 'lWfl + 'lWfS 29 ltJxe6++-) 18 ltJd5! White's pieces are incredibly active. Whatever idea I tried, everything worked perfectly! S8

18 ... exdS 19 'lWxdS+ i>h8 20 :1'&h3+-; b) 16 ... c4+ 17 i>hl 'lWc6 18 'lWd4 h6 19ltJge4 b6

20 ltJf6+! ltJxf6 21 'lWxeS ltJhS 22 'lWbSt.

150-0

We should build our play on the better coordination and space advantage. Every attempt to grab quickly

5 f4 c5 6 d5 0-0 7 ltJf3 e6 8 ~e2 exd5 9 cxd5 E:e8 material usually ends up favourably to Black: 15 lLlb5? E:d816 ~f4 lLlc617 lLlc7 'lWxb2 18 Elbl 'lWc3+ 19 ~d2 'lWd4 20 lLlxa8 'lWxd6 21 lLle4 'lWb8t. In the diagram position, main branches are: AI. IS ... lLlc6 A2. IS ... lLlfS A1. 1s ... lLlc6 16 lLlge4! Initially I let myself being enticed by the beautiful variation: a) 16 lLla4 'lWb4 17 @hl lLlfS 18 d7 .b:d7 19 'lWxd7 'lWxa4? 20 lLlxe6 ~f6 (20 ... 'lWb4 21 ~gS lLleS 22 'lWdS @h8 23 lLlxg7 @xg7 24 g4±; 20 ... @h8 21 ~gS .b:b2 22 Elael ~c3 23 ~d3 'lWxa2 24 .b:fS gxfS 2S E:e3±) 21 ~gS .b:gS 22 ElxfS Ele7 23 'lWdS lLlb4 24 Elafl!!+~h6 2S'lWd2!, but19 ... Elad8!! quickly sobered me down: 20 'lWf7+ @h8 21 lLlxcS Ele7 22 'lWxe7 (22 lLlcxe6 Eldd7 23 a3 'lWaS 24 ElxfS 'lWxfS 2S 'lWxfS gxfS 26 g4 lLld4 27 ~c4 h6 28 lLlxg7 Ele1+ 29 @g2 hxgS+) 22 ... lLlcxe7 23 lLlf7+ @g8 24 lLlxd8 'lWh4! 2S ~f4 lLldS 26 g3 'lWxd8 27 lLlxe6 'lWb6 28 lLlxg7 lLlxf4 29 lLlxfS lLlxe2 30 lLle7+ @g7 31 lLldS 'lWe6 32 lLlf4 lLlxf4 33 Elxf4 'lWe2=. This is a dead draw. b) 16 lLlce4 does not achieve the goal either owing to 16 ... lLld417 ~c4 lLlhfS 18 l"lxfS lLlxfS 19 d7 .b:d7 20 'lWxd7 Ele7+. In this system, when you are uncertain how to continue, point your pieces at the kingside dark squares. Such a strategy works here, too: 16 ... lLlfS

16 ... lLlf7 fails to 17 d7 .b:d7 18 'lWxd7 ~d4+ 19 @hl E:e7 20 Elxf7! and White emerges with a better ending after 20 ... @xf7 21 'lWd6 lLleS 22 ~gS Eld7 23 'lWxb6 axb6 24 lLlf6 Elc7 2S lLlxh7 @g7 26 lLlf6±. 17 ~gS lLlcd4 18 \!l!fa4

Black covered solidly his kingside and I was unable to find anything decisive on that wing. Fortunately, we have another valuable resource - the d6-pawn. 18 ... lLlc6 This retreat does not look natural, but it is the first line of Rybka. "She" thinks that White has nothing better but return to dI. Alternative is 18 ... 'lWc6 19 'lWxc6 lLlxc6 20 g4 lLlfd4 21 lLlxcS lLleS 22 ~dU. 19lLlf6+! ~xf6 20 i.xf6lLlxd6 21 \!l!fh4 ~f8

S9

Part 2 Black's position was difficult. For instance, 21...ltJd4 22 .ieS Wd8 23 Wf4 ltJ6fS loses a piece to 24 g4 ltJxe2+ 2S ltJxe2 b6 26 ltJc3 .ib7 27 gxfS±. 22.ih5! An exquisite move. Black still can put up some resistance, but the result of the opening battle is already clear. The threat of taking on g6 is impossible to deflect: 22 ... Wb4 (22 ... ltJf7 23 Wg3±) 23 WgS ltJfS (23 ... ltJf7 24 Wg3±) 24 ~f4 Wxb2 2S ~b1 Wd2 26 hg6 We3+ (26 ... h6 27 .if7+ @xf7 28 WhS+ @xf6 29 ltJe4++-) 27 @h1 h6 28 Wg4ltJeS 29 heS WxeS 30 hfS+ @h8 31 Wg6 exfS 32 Wxh6+ @g8 33 ltJdS±.

A2.15 ... ltJf5

favour ever since the first second. But, amazingly, going deeper, the engine rejected this option! 16... gxf5 17 .ib5 Ensuring an access to hS and removing a defender of the crucial e6square. 17... :gf8

White's attack is also very strong following 17....id7 18 Whs Wxd6 (18 ... hbS 19 Wxh7+ @f8 20 Wg6 @g8 21 ltJxe6 c4+ 22 @h1 Wd4 23 .ih6+-) 19 Wxh7+ @f8 20 ltJf3 .ic6 (20 ... We7 21ltJeS Wf6 22 ltJg6+ @f7 23 .ie2±; 20 ...hbS 21.ih6 hh6 22 Wxh6+ @g8 23 ltJxbS Wf8 24 Wg6+ @h8 2S ltJgS ~e7 26 ltJxe6+-) 21 .ih6 hh6 22 Wxh6+ @g8 23 Wg6+ @h8 24 ~e1, for instance, 24 ... WfS 2S ltJgS Wg7 26 WhS+ @g8 27 .ic4 ltJd7 28 he6+ ~xe6 29 ~xe6 tLlf6 30 Wh4t. 18.ic4

16 :gxf5!! I have not analysed any other

continuations since this hit proved to be good enough. Rybka and Fritz 11 at depth 1S still do not see it all. Later I discovered that Deep Junior 10 comported more humanly. It picked up this sacrifice in a flash and correctly evaluated it in White's 60

Black's queenside pieces are unable to rescue its king on time: A2a.18 ...h6;A2b.18 ... @h8;A2c. 18 ....id4+ A2a. 18 ... h6 19 he6+ he6 20 ltJxe6 .id4+ 21 @hl Wxd6

5 f4 c5 6 d5 0-0 7 lLlf3 e6 8 ~e2 exd5 9 cxd5 Eie8 Again, this hit is better than grabbing back material by 20 ltJxe6 .be6 21 hh6 .bc4 22 .bg7+ xg7 23 ~g5+=. 20 •••.td7 Or 20 ... a6 21 ltJc7 ~b4 22 ~f4 ~xc4 23 ~e5; 20 ... ltJc6 21i.f4. 21.tf4 tOc6 22 tOt7+ gxt7 23 YfJxt7 tOa5 24 ti'xd7 tOxc4 25 gfl e5 26 .txh6 .txh6 27 ti'e6 .te3+ 28 hl gfS 29 gf3+-. 22 ti'b3! 22 lLlxf8 ~xf8 23 ~h5 lLlc6 24 hh6 ti'f7 25 ~h4 ~g6 26 E1f1 is unclear. White's knights are dreadful attackers and should be "traded" for rooks only with concrete benefits. We have no urgent reasons to start regaining the sacrificed material. 22 •. JU7 23 .tf4 ~b6 24 tOb5 tOc6 25 .tc7 ~a6 26 tOd6± tOe5 27 h3 tOc4 28 tOxt7 ti'xe6 29 tOxh6+ We have already levelled the material while our attack is still going on.

A2c. 18 •••.td4+ 19 hl ti'xd6 20 YfJh5 YfJe7 21 .lf4 ti'g7 22 tOb5! 22ltJxe6 .be6 23 .be6+ h8 24 i.h6 ~e5 25 M8 YfJxe6 26ltJb5ltJa6 27 ltJxd4 cxd4 28 i.a3~ should only be enough for a draw. 22 •••ti'g6 23 ti'xg6+ hxg6 24 tOc7b5

A2b. 18 ••• h819 ~h5 h6

20 tOb5!

25hb5 The alternative 25 ltJgxe6 bxc4 26 tOxf8 xf8 27 E1c1ltJa6 28 ltJxa8 seems to allow Black to equalise: 28 ... ltJb4 29 Eixc4 ltJd3 30 i.d6+ f7 31 h4ltJxb2 32 Eic2 i.b7 33 .bc5 (33 ltJc7 i.e4 34 .bc5 .bc2 35 .bd4 ltJc4 36 .ba7 i.b1=) 33 ....bc5 34 Eixc5 .ba8 35 Eia5 i.c6 36 E1a6 i.e8 61

Part 2 37 <j;>h2±. The small plus is for the opposite coloured bishops attack, but it does not mean White has real chances to win. 25 ... e5 26 ~e1!! A very nice tactical blow which underlines that White's attack is not yet exhausted. The bishop is immune in view of 26 ... exf4? 27 l'!:e7 ig7 28 ct:Je8± setting up a mating net. 26 .. .lljd7 27 ic4+ ci>h8 28 lLlxa8 exf4 29 ~e7 lLlf6 3 0 ~xa7 id7 31 lLlc7 ixb2 32 h4 ic6 33 lLlce6 ~a8 34 ~c7t I have the feeling that White can improve somehow on this variation. B.13... lDeS This move aims to prevent igS, followed up by ~d2 and possible long castling. 140-0

ct:JbS ~d7 (17 ... ~b618 ct:Jxd4 cxd419 ih6 ct:Jd7 20 ~d2±) 18 ct:Jxd4 cxd4 (18 ... ~xd4 19 ~e2 ct:Jc6 20 ie3 ~eS 21 ~f2 ct:Jd4 loses to 22ic6) 19~d3~ ct:Jc6 20 ie4 ~g7 21 igS eS 22 if6 ~f7 23 ~g3 id7 24 h4~. Probably, any reasonable plan for attack on the kingside should work, too. 1SlDxeS Simple and consistent. We aim to trade dark-squared bishops. 15 ct:Je4 ct:Jxf3+ would misplace our bishop as from f3 it blocks the f-file, 16 ixf3 l'!:f8 17 igS ~b6. 1S••• ixeS IS ... ct:JxeS does not look good. I analysed in response 16 ie3 b6 17 ct:JbS l'!:f8 18 ~b3 l'!:xf1+ 19 l'!:xfll'!:b8 20 ct:Jc7 ~xd6 21 ct:Je8±. 16 if4 .ixf4 Or 16 ... id4+ 17 <j;>hll'!:f8 (17 ... eS? 18 ih6) 18 ~d2 id7 19 ih6 l'!:fS 20 ig4 leaves Black very passive, e.g. l'!:xfl + 21 l'!:xfl ct:JeS 22 ~f4 ~e8 23 ct:Je4. By the text Black wins tempi for development and a strong outpost on d4 for the knight. However, the structural defects on the kingside make his defence difficult as the dark-squares are gaping holes. 17 gxf4 eS 18 gf2 lDd4 19 ~a4

14••• lDbc6 Black can win the d6-pawn by 14 ... ct:Jxf3+ 15 bf3 id4+ (IS ... ct:Jc6 16 bc6 bxc617 ct:Je4±) 16 <j;>hl ~xd6, but the lack of his dark-squared bishop should cost him dearly: 17

62

5 f4 cS 6 dS 0-0 7 lLlf3 e6 8 ie2 exdS 9 cxdS l%e8 Suddenly the d6-pawn becomes the focus of the battle. Look at the following variation: 19 ...ifS 20 ib5 lLlxb5 21 lLlxb5 Wld7 22 Wlc4+ i.e6 23 Wle2 l%f8 (23 ...i.g4 24 ~f1 b6 25lLlc7 ~xd6 26 lLlxa8 ~a8 27 h3±) 24 Eiafl Eixf2 25 ~f2 Eid8 26 Eif6! White triumphs on the dark squares: 26 ...ha2 27 lLlc3 i.fl 28 ~xe5 b6 29 lLld5 hd5 30 ~xd5+ ci!lg7 31 Eie6 Eie8 32 ~e5+ ci!lfl 33 Wlf6+ ci!lg8 34 Eie7 ~e7 35 dxe7± ~e8 36 h4 Wlfl 37 Wle5 ~e8 38 ~e6+ ci!lg7 39 g4 hS 40 g5. This variation perfectly illustrates White's positional goal. 19•••J.e6 20 J.c4 ~h8 21 be6 l::!xe6 22lLle4 Wlh4 23 l::!e1 l::!xd6 24 g3 Wlg4 25 Wlc4 l::!b6 26 Wlf1iii Black's defence remains difficult, for instance, 26 .. .'~c8 27 lLlg5 lLle6 28lLlfl+ ci!lg8 29 Eixe5±.

c. 13••• tDc614 0-0 The thematic 14 i.g5 is not precise due to 14 ... ~b6 and White must put his queen on d2 which might prove not too useful: 15 ~d2 lLld4 16 lLlxd4 cxd4 17 hg4 dxc3 18 bxc3 i.e5't.

C1.14... Eif8 C2. 14... lLlf6 C3. 14... lLld4 C4.14 ... i.d7 14... e5? hardly deserves attention due to 15 lLlg5 lLlh6 16 ~dS+ ci!lh8 17 Eifl±. C1.14 .. JU8 An attempt to anticipate White's play on the f-file. The problem with this move is that Black's queen becomes very passive after: 15 J.g5 Wld7 15 ... ~b6? is bad duw to 16 lLla4 ~a5 17 Eicl±. 16 tDe4 b6

17 tDfd2! tDge5 18 l::!xf8+ ~xf8 19 J.b5 .ib7 20 Wlb3iii Despite Black's effort, the f-file and the f6-square will be in White's control. The b3-queen is rather mobile along the third rank all the way to h3. C2. 14... tDf6 15 .1g5 h6 This move solves the problem with the awkward pin, but produces a fatal weakness on g6. 63

Part 2 C3a. 15 ... b6? 16 .tg5 tiJxd4hd4+

16.lxf6 16 .th4 loses a tempo: 16 ... gS 17 .tf2 b6oo. 16 ....lxf6 17 Wfd3 @g7 18 ~ad1 .ld7 19 Wfe4 ~f8 20 .ld3 .le8 21 .lc4;t Black cannot defend everything: 21.. ..td7 22 lLlbS or 21...lLld4 22 lLleS with heavy pressure.

C3. 14 ... lLld4 This move looks Black's most natural reaction. I often face it in blitz games. 15 lLle4

~d7

18 ~xd4!!+- exd4 19 hg4 20 b3 . A funny position. I have won already a couple of games like that, so keep it in mind. ~a4

C3b. 15 .•. tLlf6? 16 .tg5± tLlxe4 17 hd8 ~xd8 18 ~el tLlxd6 19 ~dl± W.Arencibia-Peredun, Toronto 2003. C3e. 15 ••. tLlf5? 16 .tg5 ~b6 17 d7 e4+ 18 @hl hd7 19 ~xd7 h6 20 he4±, Berkovich-Shahal, Beersheba 1991. C3d. 15 ... ~f8 16 .tg5 ~b6

Reviving the plan with .tgS. C3a.1S ...b6?; C3b.1S ... lLlf6?; C3c. lS ... lLlfS?; C3d. lS ... E1f8; C3e. lS ... h6 64

17

5 f4 c5 6 d5 0-0 7 lOf3 e6 8 ie2 exd5 9 cxd5 ~e8 17 tLlfd2! I could not find a tangible advantage after 17 ttJxd4 ixd4+ 18 ~h1 gxf1+ 19 ~xf1 ~c6 (19 ... ~xb2 20 gel ttJe5 21 gb1 ~xa2 22 ih6 id7 23 gxb7+-) 20 if3 (20 ixg4 ~xe4 21 ~f3 ~e5 22 ~e7 ~d7 23 gel ~f5 24 b3 gb8 25 ~e2 b5 26 ~g4=) 20 ... ~xe4 21 he4 ttJf2+ 22 ~xf2

23 ~f3 gb8 24 a4 b6 25 gfl ~a6 26 ~f4 offers good compensation, but the text is linked with the clear positional idea to kill the fS-knight. Then White's knights and the d6pawn will rule over the board. 23".a6

hf2

~f6

26

~xc4

28

24 i.xf5! gxf5 25 gel I have devoted a lot of time to this endgame. White wins the exchange with hb7, followed by d7. The question is, is that enough for an edge. My final judgment is that the position should be drawish. So we better switch to a slower positional treatment. Our goal for the next few moves will be not to win immediately, but to enforce domination in the centre. 17••• gxf1+ Alternative is 17... ttJe5 18 Elxf8+ ~xf819 ~fl + ttJi7 20 ttJc4 with strong pressure, for instance, 20 ... ~c6 21 ~f3 ~b5 22 a4± ttJxf3+ 23 ~h1! ~d7 24 ~xf3+-. 18 ~xf1 tLle5 19 .tf6 tLld7 20 h.g7 rtixg7 21 J.d3 tLlf5 21...~xb2 loses to 22 lOc4 ~b4 23 gb1 ~a4 24 ~f4+-. 22 tLlc4 ~d8 23 tLlc3!

~f2b5

Or 26 ... ~d4 27 gxe6 ~xf5+-.

27 tLla5 Black's queenside is not less vulnerable, for instance, 27...h6 28 rtih1 ~g6 29 ~f3 gb8 30 ~c6 rtih8 31 ~c7±. 27".tLle5 28 tLlb3 c4 29 tLlc5 rtif7 30 a3 tLld3 31 tLlxd3 cxd3 32 gd1 J.d7 33 gxd3±. C3e.15".h6 The previous variations showed that Black suffers when White's bishop reaches g5. Therefore, the text is a logical and testing attempt. Its big drawback though is the weakening of the g6-square. 16 tLlxd4! hd4+ 17 rtih1 It seems that Black is lost, be he still has a nice counterblow: 65

Part 2

17••• ~xh2 18 ~f4!;!; g5 19 ~f2 ~xf2 ~f8 21 VUd3!

.hf2 20

This ends the tactical interchange in White's favour. 21lLle4 b6 22 ~d3 ~d7 23 @xh2 is bad in view of 23 ...i.b7 24 i.e3 E1ad8 2S E1d1 ~h7 26 lLlc3 ~xd3 27 hd3 i.dS't. 21. •• ~xf2 22 VUg6+ @f8 23 ~xh6+ @g8 24 VUg6+ @h8 25 i.d2±.

D.13 ... .id7 At first I thought that the somewhat slow regrouping of Black should give White a chance for opposite attacks and analysed 14 i.gS ~b61S ~d2 i.c6? 16 h3lLleS17lLlxeS heS180-0-0!

Incredibly, Black might be beyond salvation here! He seems helpless against h4-hS, for instance, 66

18 ... lLld7 19 h4! hg2 20 E1hg1 i.h3 (Or 20 ... i.dS 21 hS ~xd6 22 hxg6 hxg6 23 i.h6 @h8 24 ~e3+- intending 24 ~h3) 21 hS i.fS 22 hxg6 hxg6 23 i.h6 ~b4 (23 ... i.d4 24 E1xg6+ hg6 2S ~gS @h7 26 E1h1 +-) 24 i.d3 ~d4 2S E1xg6+ hg6 26 ~g2 i.g7 27 ~xg6lLlf8 28 ~gS+-. However, ls ... lLlc6! puts to the test the whole White concept. After many fruitless attempts, I had to give up. I must say I was rather disappointed. I had to return to the positional treatment with short castling: 140-0 White achieves good results with 14 lLlgS, but Rybka's first line 14 ... lLlh6 is rather unpleasant. Then 1S 0-0 lLlfS16i.d3 h6? fails to 17 hiS gxfS 18 ~hS!! hxgS 19 hgS ~b6 20 i.h6 c4+ 21 @h1 ~d4 22 hg7 @xg7 23 ~gS+ @f8 24 E1f3~, but 16 ... ~f6! repels the attack. You can also see game 10 Vasilchenko-Kovalev, Katowice 1990.

D1.14 ... ~b6 D2.14 ...i.c6 14 ... lLlc6 hardly deserves detailed examination in view of 1S i.gS! (I fail to conceive the reason

5 f4 cS 6 dS 0-0 7 lDf3 e6 8 .ie2 exdS 9 cxdS ~e8 behind 15 i>hl ~f8 16 .ig5 ~a5 17 .id2oo, EI Gindy-Atabayev, AI Ain 31.12.2008) 15 ... ~b6 1614 ttJa4. 01. 14 ...Wb6 Computer's suggestion. Black is eager to win more material - the exchange or the d6-pawn. The flipside is that he leaves his king at the mercy of our pieces. Perhaps White has various good options, but I was satisfied with the first and most straightforward variation I tried: 15 We1 ttJc6 Our attack runs smoothly after 15 ... ~xd6 16 ~h4 ttJe5 (16 ... h5 17 h3 ttJe518 ttJe4 ttJxf3+ 19 ~xf3 ~d4+ 20 ~e3 ~xb2 211"1afl±) 17 ttJe4 ttJxf3+ 18 1"1xf3 ~e519 ~g5 ~c6 20 ttJf6+ hf6 21 M6 ~xe2 22 ~h6 1"1e7 23 he7 lDd7 24 1"1f2±. This variation clearly shows that Black's active 14th move is rather dubious. 16 Wh4 ltJce5 17 ltJxe5 ltJxe5 1SltJe4

1"1f8211"1ael! 21 hg7 i>xg7 22 ~e7+ i>g8 would transpose to the main line. 21...~d4+ 22 i>hl1"1ae8 23 1"1f4! h6 24 ~g3! 24 hg7 1"1xf4 25 ~xh6 ttJf7 26 ~xf4 i>xg7 271"1fl1"1f8 28 ttJxc5 ~xf4 29 1"1xf4 e5 30 1"1£1 ~f5 31 g4 1"1c8 32 gxf5 1"1xc5 33 d7 gxf5 34 1"1xf5 ttJd8 35 1"1g5+ i>f8 36 ~f3 also favours White, but the text is much more entertaining. 24 ...~c6 24 ... 1"1f7 25 1"1dl Wb2 261"1f2! cuts the black queen off from e5, winning material: 26 ... 1"1ef8 27 ~f3 ~xf2 28 ~xf2 hf6 29 ttJxc5±. 25 ~f3 g5

An amazing position where all the pieces on the board seem to be cross-pinned. White wins now with the cold-blooded 26 h4 ttJxf3 27 hxg5.

Our strategy is simple - we aim to remove the g7-bishop. 1S ... 1"1fS Black can discourage 18 ~h6 by 18 ... ~b4. Then we can reroute our bishop to f6: 19 ~g5 ~xb2 20 ~f6

19 .ih6 Wxb2 20 .ixg7 i>xg7 21 We7+ i>gS 22 ltJg5 The endgame after 22 ttJf6+ 1"1xf6 23 Wxf6 ~xe2 24 1"1ael ltJf3+ is unclear. As a rule in this system, an extra exchange does not guarantee White an advantage even in the endgame. It is better to have an attack. 67

Part 2 22 ...Wid4+ 23 @h1 Wih4 24 25 ~f1 ~e8 26 Wif6 h6 27 liJh7 Wixf6 28 liJxf6+ @g7 29 liJxe8+ .ixe8 30 ~c1 b6 ~xf8+ ~xf8

This version of the endgame is better for White since his rook is very active. Play can continue with 31 ~c3liJt7 32 ~d3 .ib5 33 l"i:e3liJxd6 34l"i:xe6 (34 hb5liJxb5 35l"i:xe6 mt7 36 l"i:c6liJd4 37l"i:c7+ me6 38l"i:xa7t) 34 ... he2 35 l"i:xd6 .ic4 36 a3 mt7 37 mgl±.

02. 14 ....ic6 Black plays very solidly. After ttJb8-d7 the critical f6-square will be well defended, the e4-square will also be under control, all the threats connected with advancing the d-pawn will evaporate. The only drawback of this approach is that Black is too passive and e6 becomes vulnerable. White's main resource is ttJb5, threatening with ttJc7. 15 liJg5 I rejected 15 .ig5 '\Wb6 16 '\Wd2 due to 16 ... h6! (16 ... ttJd7 was unclear in Kantorik-Likavsky, Ostrava 2005: 17l"i:ael c4+ 18 mhl '\Wc519 h3 liJge5 20 .ie3 '\Wa5 21 ttJd4l"i:ac8oo) 17 68

.ie7 c4+ 18 mhl ttJe3 18 ... liJf2+ 19 l"i:xf2 '\Wxf2 20 l"i:fl is promising for White: 20 ... '\Wc5 21 ttJh4 g5 22 ttJg6 (22 .ih5 gxh4 23 .it7+ mh7 24 he8 h3 25 '\Wc2+ mh8 26 hc6 ttJxc6 27 .if6=) 22 ....id7 (22 ... liJd7 23 .ig4 '\Wd4 24 he6+ mh7 25 '\Wc2 '\Wd3 26 '\Wf2+-) 23 .if6 (23 ttJf8 l"i:xf8 24 hf8 liJc6 25 hg7 mxg7 26 '\Wc2 '\We3 27 hc4 l"i:f8't) 23 ... ttJc6 24 h4~. After the text however, White has nothing better than the drawish variation: 19 l"i:fel ttJd7 20 ttJdl ttJf5 21 hc4 ttJc5 22 he6+ ttJxe6 23l"i:xe6.ixf3 24 gxf3 l"i:xe7 25 l"i:xe7 ttJxe7 26 dxe7=. 15 ... liJe5 Again, it is wrong to grab the exchange: 15 ....id4+ 16 mhl ttJf2+ 17 l"i:xf2 hf2 18 '\Wfl h6 19 'lWxf2 hxg5 20 '\Wg3 with attack. 16.ie3

16 ... b6 Recommendation of the ECO, vol. A, 4th edition. Black keeps his knight on b8 in order to meet 17 ttJb5 by 17... ttJa6. This is an arguable concept, since after 18 .if4 White has an obvious initiative, for instance: 18 ... c4 (Or 18 ... h6 19 '\Wbl ttJb4 20 ttJh3 '\Wd7 21 he5 he5 22 '\Wxg6+

5 f4 cS 6 dS 0-0 7 ttJf3 e6 S Jie2 exdS 9 cxdS geS V9g7 23 V9xg7+ hg7 24 ttJc7 hb2 2SEiab1 Jid4+ 26 <j;;>h1 ttJd5 27 ttJxaS gxaS 2S ttJf4 ttJxf4 29 Eixf4 EidS 30 gbflt) 19 ttJd4 V9xd6 20 hc4 ttJcS 21 Eic1 h6 22 he5 heS 23 ttJxc6 hxg5 24 V9xd6 (24 ttJxe5 V9xe5) 24 .. "bd6 25 b4 ttJb7 26 Eicdlt. Still, I will examine as a main line another variation, which also provides full compensation for the pawn. Let us examine now 16 ... ttJbd717 lLlb5 (17 Vge1 is another interesting option) 17... h6 1S lLlh3 hb5 1S ... ttJg4 19 hg4 hb5 20 i.e2 he2 21 V9xe2 leaves Black with several weaknesses, e.g. 21...gS 22 lLlf2 EifS 23 h4t. 19 hb5 V9b6 20 V9a4 gadS 21 lLlf4 gS 22 lLlh5 i.hS 23 gadl. White's game is easier as his pieces are more active.

Eixe6 20 V9b3 lLlfS 21 gxfS+ V9xfS 22 V9xe6+ V9f7 or 17 a4 lLlbd71S lLlbS h6 19 lLlxe6Eixe6 20 V9b3 lLlfS 21 EixfS+ V9xfS 22 V9xe6+ <j;;>h7 23 gflEieS, but it seems that Black is holding there. By the way, ECO only mentions 17 lLlge4?!. 17 ... ltJbd7

Black protected everything, but his position is passive. We should seek targets on the queenside 18ltJb5 hS 19 ltJh3ltJg4 20 .if4 g5

Or 20 ... hb5 21 hbS;!:;.

17~d2

I suppose that this calm development might be most unpleasant to Black. Beside 17 lLlb5!?, which I mentioned in the previous paragraph, we should also have in mind 17 Eic1 lLlbd7 1S lLlbS h6 19 lLlxe6

21 ltJxg5! hxg5 22 .ixg5 ltJgfS 23 ~f4.

White maintains pressure on the kingside.

69

Part 2

1 d4 tDf6 2 c4 g6 3 tDc3 ig7 4 e4 d6 5 f4 cS 6 dS 0-0 7 liJf3 e6 8 ie2 exdS 9 cxdS 13e8 COMPLETE GAMES

It was very difficult to find a game

to comment for this chapter. The variation with 12 e6 is too rare and at least one of the sides plays rather badly in the unexplored positions that arise. Finally I chose one game. White commits a common mistake in the opening and quickly gets into trouble, but then recovers to even get the upper hand. The game is rich of typical tactical motifs. 10. Vasilchenko - Kovalev Katowice 1990 1 d4 tDfS 2 c4 gS 3 tDc3 ig7 4 e4 dS 5f4 0-0 S tDf3 c5 7 d5 eS 8ie2 exd5 9 cxd513e8 10 e5 dxe5 11 fxe5 tDg412 eS fxeS 13 dS tDcS 14 tDg5

This manoeuvre is quite popular, but it is often a mistake. Correct is 14 0-0. 14... tDhS!

The same retreat is also extremely unpleasant to 13 ...i.d7 14 ttJg5. The knight is heading for f5 from where it covers the critical ffile while threatening (and usually capturing!) the d6-pawn. 150-0 tDf51S tDge4 hS!

White's play on the kingside is over. His last hope is now linked with the weak e6-pawn: 17 tDb513f8 18 tDc713b8 19 ic4

19 ... tDcd4

Now White is back in the game 70

5 f4 c5 6 d5 0-0 7 lLlf3 e6 8 ie2 exd5 9 cxd5 ~e8 again. Had he found the cunning prophylactic move 19 ... ~h7!, White would have been much worse, for instance, 20 he6 (or 20 if4 hb2) 20 ....id4+ 21 ~h1 lLlxd6 22 lDxd6 ~xd6-+. Note that the immediate 19 ....id4+ would be unclear after 20 ~h1lDxd6 21lDxd6 ~xd6 22 lDb5. 20 lDxe5 fJxd6 21 lD5xe6 .ixe6 Black still could obtain a small edge with 21...~b4! 22lDxg7+ ~xc4 23 lDxf5 .ixf5 24 b3 lDe2+ 25 ~h1 ~g4 26 lDd5 .ie4 27 lDe3 ~h5 28 E:xf8+ E:xf8 29 .ia3 lDg3+ 30 ~gl ~e2 31 ~xe2 lDxe2+ 32 ~h1 E:f2=t. 22 lDxe6 fJe6 22 ... lDxe6! was imperative: 23 ~xd6lDxd6 24 he6+ ~h7 25 E:d1=. 23lDxg7+! A good decision. White's game is already preferable due to his bishop pair. 23 lDxf8+ ~xc4 24 lDxg6 was risky and Black would have at least equality: 24 ... lDe2+ 25 ~h1 lDfg3+ 26 hxg3lDxg3+ 27 ~gl.id4+ 28 E:f2 lDe2+ 29 ~h1.ixf2=. 23 ... ~xg7 24 id3 lDh4 25 fJg4 gxf1 + 26 ixf1 lD hf5

Alternatively: 26 ... ~c5 27 hh6+ ~xh6 28 ~xh4+ ~g7 29 ~h1lDf5 30 ~f4±. Now the same hit 27 hh6+! was winning, but White makes a few mistakes: 27 id2? ge8 28 ge1? This exchange is positionally bad because having the bishop pair, White should aim to keep a rook, too. The combination of E:+li vs. E:+lDlD in an open position is much stronger than mere l i vs. lDlD, especially with queens on board. Moreover, the text loses a pawn in three after 28 ... ~b6 29 ~h1 E:xe130 he1 ~xb2. 28 ... gxe1? 29 .ixe1 fJe6 Now 29 ... ~b6 would be met by 30.ic3. 30 .if2 fJe5 31 fJd1 h5 32 fJe1 fJf4 33 id3 h4 34 .ixf5?! White loses nerve and bails out to a draw by parting with one of his bishops. 34 ~c3 ~h6 35 .ift!: kept things under control although the enemy knights are a real nuisance. 34 ... lDxf5 35 fJe3+ ~h7 36 .ixa7 fJa4 37 if2 fJxa2 38 fJe5 fJb1+ 39 ie1 b5 40 h3 fJd3 41 ie3 fJd1+ 42 ~h2 fJd6 43 fJxd6 lDxd6 44 .if6 lDf5 45 ~g1 ~h6 46 ~f2 g5 47 ~f3 ~g6 48 ie5 rJ/f7 49 .ie3 lDd6 50 ib4 lDf5 51 ~e4 ~e6 52 .ie1lDd6+ 53 ~d4lDf5+ 54 ~e5 lDe3 55 .id2 lDxg2 56 ixg5 ~f5 57 idS lDf4 58 rJ/xb5 lDxh3 59 ixh4 lDf4 60 rJ/e4 ~e6 61 b4 rJ/d7 62 ig3 lDe6 63 rJ/b5 %-%

71

Part 3

1 d4 ltJf6 2 c4 g6 3 ltJc3 .tg7 4 e4 d6 S f4 cS 6 dS 0-0 7 ltJf3 Rare Lines QUICK REPERTOIRE

In practice I often observe attempts from Black to deviate from the main lines. Instead of taking a pawn and struggling in a cramped position which requires good home preparation, many Black players prefer to part with a pawn, but take over the initiative with a Volga stile approach. At least, they think so. In fact, however, Black does remain a pawn down, but he has no trace of an initiative either. White takes what he can and returns to his main plan of building a kingside attack: 7 ... bS?! 7... ttJa6 is seldom seen. After 8 .td3 play transposes to lines from the next part, while 7... a6 8.td3!? bS requires more attention:

9 eS (The engines prefer White after 9 cxbS axbS 10 ttJxbS .ta6 11 ttJc3 ~aS 12 0-0 ttJbd7.) 9 ... dxeSlO fxeS ttJg411 ~e2 ttJd7 Game 11 Rodriguez VargasUhlmann, Dresden 2004, saw 11 ...b4 12 ttJa4 ~c7 13 .tf4 ttJd7 14 0-0 ttJgxeS 15 ~hl.tb716l"i:ael with strong initiative. 12 .tf4 ~c7 13 O-O! ttJdxeS 14 ~hl. White has full compensation for the pawn. 8 cxbS a6 9 a4

Black has two major plans now. He can play on elimination of the centre or seek typical Volga counterplay on the queenside. 9 ... ~aS? 10 .td2 ~b4 11 ~c2± hardly needs any attention. 72

1 d4 C2Jf6 2 c4 g6 3 C2Jc3 i,g7 4 e4 d6 5 f4 c5 6 d5 0-0 7 C2Jf3 The central plan with 9 ... e6 has a bad reputation due to 10 i,e2 axb511 .bb5 exd5 12 e5, but I have doubts about that. In my opinion, simplest is 10 dxe6 .be6 11 e5 dxe5 12 1Wxd8

11 i.d2 hb5 12 axb5 ~xa1 13 1l!fxa1 1l!fb6 13 ... C2Jbd714 0-0 C2Jb6 151Wa2, Diloudi-Zawadzka, Ag.Pelagia 2004. 140-0 lLlbd7

~xd8

13 C2Jxe5!? C2Jd514 C2Jxd5 .bd515 mf2 C2Jd7 16 C2Jxd7 ~xd7 17 bxa6 c4 18 i,e2, Stocek-Babula, Plzen 2000. White is a pawn up and his king is closer to the centre. 9 ... axb5 10 hb5 .ta6 Black has not time for slower plans since White is too active in the centre, e.g. 10 ... C2Ja6 11 e5!? C2Jd7 120-0.

151l!fe1! This is a key idea! White's play is on the kingside and he should not distract himself from his main objective. This move has been introduced in the game Lautier-Shirov, Belgrade 1997 and it is unanimously approved as best. Play might continue with 15 ...1Wb7161Wh4! C2Jb6 17 f5 gxf518 C2Jg5 ~a819 ~xf5 ~a1+ 20 mf2 with attack.

73

Part 3

1 d4 ltJf6 2 c4 g6 3 ltJc3 .1g7 4 e4 d6 5 f4 cS 6 dS 0-0 7 ltJf3 Rare Lines STEP BY STEP

A. 7... a6

B. 7... bS 7... tLla6 is a rare move. After 8 i.d3 play transposes to lines from the next part.

A.7 ... a6 This move is seldom seen in practice and it is completely unexplored. I'm afraid that if I were to analyse it in detail, the book could have taken another year to be completed. The whole problem is that it is not easy to weigh the pros and cons of the insertion of 7... a6 8 a4 in regard to the main lines. Most adepts at the FPA would say that it can be only in White's favour, but I'm not that sure. In some lines Black would be happy to have the bS-square covered. I suppose that White should seriously consider to refrain from the automatic 8 a4. Or he can link it with the fresh idea of 9 i.d3!?: 8 a4e6 Black has not full compensation for the pawn after 8 ... eS 9 fxeS tLlg4 10 exd6. For instance: lO .. .fS 11 i.d3 (or 11 exfS hfS 12 i.d3! :1l:e8+ 13 74

tLle2 hd3 14 'lWxd3 'lWxd6 IS a-a;!:;) 11 ... 'lWxd6

Golubev's suggestion 11...hc3+ 12 bxc3 fxe4 13 he4 :1l:e8 stumbles into Rybka's 14 a-a!! :1l:xe4 IS h3 tLlf6 (IS ... 'lWxd6 16 hxg4 hg4 17 'lWb3 tLld7 18 tLlgS±) 16 tLlgS :1l:e8 17 'lWf3 h6 18 'lWxf6±. 12 exfS±, Lautier-Shchekachev, Bad Zwesten 1999. 9 i.d3!? 9 i.e2 exdS 10 cxdS (10 exdS is completely dull, for instance, 1O ... :1l:e8 11 0-0 i.fS 12 i.d3 hd3! 13 'lWxd3 tLlbd7) 1O ... i.g4 should be nice for White who can follow the same strategy as described in Part 1. The critical test is 1O ... :1l:e8! with unknown position. 9 ... exdS 10 cxdS :1l:e8 11 0-0

1 d4ltJf6 2 c4 g6 3 ltJc3 ii.g7 4 e4 d6 S f4 cS 6 dS 0-0 7 ltJf3 This is a complex double-edged position where White's strong centre should ensure him good chances.

Vienna 1986: 14 ~aelltJxf3+ IS ~xf3 '!Nb6 16 '!Nxe7±, or Yrjola-Sigurjonsson, Reykjavik 1986: 14 whU.

Instead, I will examine another interesting plan: 8 Ad3!? bS

geS Personally I would not even think about offering the initiative to the opponent by grabbing a pawn: 9 cxbS axbS 10 ltJxbS ii.a6 llltJc3 '!NaS 120-0 ltJbd7. The engines, however, prefer White here. 9 ... dxeS 10 fxeS ltJg4 11 ~e2ltJd7

An interesting position arises after 11...b4 12 ltJa4 '!Nc7 13 ii.f4 ltJd7 14 0-0 ltJgxeS IS whl ii.b7 16 ~ael f6. Black is tied down and cannot defend against a direct attack, for instance, 17 h4! '!Nd6 18 hS--+. See game 11 Rodriguez Vargas-Uhlmann, Dresden 2004. 12 Af4 ~c7 13 O-O! see next diagram This capturing leaves the king's knight on an active position. Instead, 13 ... ltJgxeS was clearly better for White in Kortschnoj-Gheorghiu,

13... liJdxeS 14 ~h1 Maintaining the pin. 14 ~ael!? ltJxf3+ IS ~xf3! also gives White an initiative. 14•.• b41SliJb1! White's position is so good, that he can indulge in improving the placement of his only unemployed piece. Alternatives IS ltJa4 f6, IS ltJdl f6 16 h3ltJh6 17ltJf2oo or ISltJe4 '!Na716 ltJxeS ltJxeS 17ltJxcSltJxc4 18 ii.xc4 '!NxcS 19 ~ac1 ii.d7 20 d6 exd6 21'!Ndl= are not convincing. 1S...f6 Perhaps IS ... '!Nd6 16 ltJbd2 '!Nf6, which breaks the pin, would give more chances, although the endgame after 17 ltJxeS ltJxeS 18 ltJe4 '!Nb6 19 ltJxcS ii.g4 20 '!Nf2 ~ac8 21 ~ael ltJxc4 22 ii.xc4 '!Nxcs 23 '!NxcS ~xcS 24 b3 ii.f6 2S h3 ii.fS 26 d6± is unpleasant for Black. 16 liJbd2 liJh6 17 gae1 Ag4 18 h3 bf3 19 liJxf3 liJhf7 20 liJxeS liJxeS 21 h4 Black will be submitted to an opposite coloured bishops attack.

7S

Part 3

White can even keep his darksquared bishop when his great space advantage should provide a lasting pressure.

B.7 ••• bS?! This clumsy attempt to play in Volga style leads to a clear advantage for White. He keeps all his options for a kingside attack open while nurturing an extra pawn "as a compensation". 8 cxbS a6 9 a4

14 fxeS lDg4 15 lDdl! hbS 16 hb4 cxb4 17 ~e4 .id7 18 h3; 11 ... c4 12 eS! .if5 13 lDa2 ~cS 14 ~xc4 lDxdS 15 ~xcS±. 12 eS .if5 13 .id3 hd3 14 ~xd3 c4 15 ~bl± A.Schmidt-Benischek, Bayern 1999. b) 9 ....ib7 Black's bishop is not too useful on this diagonal. That gives White a wide choice. He can develop the bishop on e2, d3 or c4, for instance: 10 .ie2 e6 11 dxe6 fxe6 12 eS lDdS 13 lDxd5 exdS14 0-0 dxe51S fxeS;!;; 10 .ic4 axbS 11 hbS lDa6 12 0-0 lDb4 13 ~e2 .ia6 14 .id2 ~b6. However, 10 bxa6lDxa611.ic4;!; looks simplest. See game 12 Vaisser-Brito Garcia, Las Palmas 1993.

81. 9 ... e6 10 dxe6 You may read in many sources that 10 .ie2 axbS 11 hbS exdS 12 eS is in White's favour, but I do not believe it. Black should be fine after 12 ... lDe8 13 lDxdS .ib7 14 .ic4 lDc6 15 .ie3 dxeS16 hcSlDd600 17lDe7+ lDxe718 ~xd6 ~xd619 hd6 exf4oo. The text is more natural. 10 ....ixe6 11 e5 dxe5 12 VMxd8 ~xd8

Bl. 9 ... e6 B2.9 ... axbS Minor alternatives are: a) 9 .. :t;WaS? 10 .id2 ~b4 11 ~c2± 11 ....id7 Black can easily lose his queen: 11 ... axbS 12 hbS .ia6 13 eS+- dxeS 76

1 d4 ttJf6 2 c4 g6 3 ttJc3 ~g7 4 e4 d6 5 f4 cS 6 dS 0-0 7 ttJf3 13 ttJxe5 13 fxeS ttJg4 14 ~gS l"leS unnecessaryly returns the extra pawn. 13 ... llJd514llJxd5 .ixd515 i>f2 llJd7 16 llJxd7 1:'!xd7 17 bxa6 c4 18 .ie2

White is a pawn up and his king is closer to the centre. The game Stocek-Babula, PI zen 2000 saw further: 18 ....ie6 19 as .id4+ 20 .ie3 .ixb2 21 1:'!ab1 .id4 22 1:'!b7 .ixe3+ 23 i>xe3±.

B2. 9 ... axb5 10 .ixb5 .ia6 1O ... ttJa6 is too sluggish. White

can respond with 11 0-0 or the more aggressive 11 eS!? ttJe8 (11...ttJd7 12 0-0 dxeS 13 fxeS ttJxeS regains the pawn, with a gloom position for Black though: 14 ttJxeS heS IS ~h6 ~g7 16 hg7 <;t>xg7 17 Wf3±) 12 0-0 ttJb4 13 We2 ttJc714 ~c4 ~b71S l"ldl l"la7 16 ~e3 WaS 17 Wd2± with a healthy extra pawn, Banikas-Karner, Menorca 1996. 11 .id2 .ixb5 12 axb5 The only explanation of why Black would want to play this position is that he hopes to lure White

into the tempting, but drawing variation 12 ... l"lxaI13 Wxal Wb6 14 0-0 ttJbd7 IS Wa6?! l"lbS 16 l"lal ttJeS 17 Wxb6 ttJxb61Sl"la7 hc3! 19 hc3 fS 20 l"lxe7 fxe4= . However, in the Volga gambit White commonly wins on the kingside or in the centre. This position is no exception and the first player should always keep that in mind.

12 ... 1:'!xa1 Or 12 ... ttJbd7 13 0-0 ttJb6 14 We2±. 13 Wixa 1 Wib6 Practical experience has also seen 13 ... ttJbd7 14 0-0 ttJb6 15 Wa2 Wc7 16 l"lal c4 (16 ... l"laS 17 Wbl) 17 h3 WcS+ IS <;t>h2 ttJhS 19 Wbl ~h6 20 Wgl Wxgl+ 21 <;t>xgl hf4, Diloudi-Zawadzka, Aghia Pelagia 2004, 22l"la7!±. 140-0 llJbd7 15 Wie1! Very consistent! The queen goes toh4,orsupportse4-eS.lnthesource game Lautier-Shirov, Belgrade 1997 Black answered IS ...Wb7, when instead of 16 eS Lautier suggests 16 Wh4! with a better game, e.g. 16 ... e6? 17 dxe6 fxe61S ttJgSl"leSl9 eS+or 16 ... ttJb6 17 fS gxfS IS ttJgS l"laS 19 l"lxfS l"lal + 20 <;t>f2 with attack. 77

Part 3

1 d4 lDf6 2 c4 g6 3 lDc3 .ig7 4 e4 d6 5 f4 cS 6 dS 0-0 7 lDf3 Rare Lines COMPLETE GAMES

11. Rodriguez Vargas - Uhlmann EU-chT Seniors Dresden 2004 1 d4 lDf6 2 c4 g6 3 lDc3 .ig7 4 e4 d6 5 f4 c5 6 d5 0-0 7 lDf3 a6 8 .id3 b5 9 e5 dxe5 10 fxe5 lDg4 11 .if4 lDd7 12 ~e2 b4 13 lDa4 ~c7 140-0

sition: 14 ... l2JdxeSI5 Eiaell2Jxf3+ 16 ~xf3±.

15 ~h1 .ib716 ~ae1 f617 .ic2 ~ae8

Black correctly seeks counterplay with ... e6. Passive defence as 17... Eiac8 would face 18 h4! and it would be very difficult to neutralise White's attack. 18 .ixe5lDxe5 19lDxe5 fxe5 20 ~e3 ~xf1+ 21 ~xf1 ~c8 22 .id1

14... lDgxe5

Black wins a pawn indeed, but White's heavy pieces are much more active. The only problem is how to break the enemy's defensive line. One target is cS. We'll need to fix another one on the kingside with h4-hS. Capturing by the other knight does not affect the character of po78

Black's position is difficult. He has no counterplay at all, no matter with rooks on the board or without them. His strongest piece is tied down to defend c5, both bishops are unemployed.

1 d4 lDf6 2 c4 g6 3 lDc3 .ig7 4 e4 d6 5 f4 c5 6 d5 0-0 7 lDf3 22 .. JU8 23 gxf8+ .lxf8 24 Wxc5

12. Vaisser - Brito Garcia Las Palmas 1993

Perhaps White should have changed the light-squared bishops first: 24 .ig4 .ic8 25 .ixc8 Wlxc8 26 Wlxe5 wt7 27 g3 .ig7 28 Wle6+ Wlxe6 29 dxe6+ Wxe6 30 lDxc5+ wd6 31 lDxa6 .ixb2 32 lDxb4 e5 33 c5+ We6 34 Wg2, since the move played in the game allows Black to get oppositecoloured bishops.

1 d4 tilf6 2 c4 g6 3 tilc3 .lg7 4 e4 d6 5 f4 0-0 6 tilf3 c5 7 d5 b5 8 cxb5 a6 9 a4 .lb7 10 bxa6 tilxa6 11 .lc4 til b4 12 0-0

24 ...Wd6? 24 ...Wlxc5 25 lDxc5 .ic8 would

have given fair drawing chances in view of the threat e7-e6. 25 Wxd6?

White was obviously afraid to let the opponent activate his queen. Still, 25 .ig4 Wlf6 26 .ie6+ wh8 27 Wlgl would have given White a clear advantage, while now he would be even on the defensive. 25 ... exd6 26 .lg4 clJf7 26 ... e4 was better: 27.ie6+ Wg7 28 Wgl wf6 29lDb6=. 27 clJg1 e4 28 clJf2 clJf6 29 lDb6 .ih6 30 id7 White tries a last trick. Now he is threatening with 30 .ic6, but Black is on guard and the game peacefully steers to a draw. 30 ... a5 31 c5 dxc5 32 .lc6 .la6 33lDd7+ clJe7 34 tilxc5 e3+ 35 clJe1 .lc4 36 b3 .lxd5 37 .lxd5 clJd6 38 tilb7+ Wxd5 39 tilxa5 clJd4 40 clJe2 Wc3 41 tilc4 clJc2 42 tilxe3+ clJb2 43 tild5 clJxa2 44 tilxb4+ clJxb3 45 tild5 clJc4 46 tilf6 clJd4 47 clJf3 .lc1 48 tilxh7 clJe5 49 tilf8 clJf6 50 clJe4 .la3 51 lDd7+ We6 52 tile5 .ld6 53 tilf3 clJf6 54 g4 clJe6 55 h3 .lg3 56 tild4+ Wf6 %-%

Black has sacrificed a pawn, but his setup is mostly defensive. For now, he has discouraged e4-e5. To build up some counterplay, he must remove the blockade on the c4-square. In the game he tried to achieve that by lDf6-d7-b6. Alternative was 12 ....ia6 13 .ixa6 !%xa6 14 Wle2 Wlc8 15 whl !%a8 16 lDb5 lDd7. We see that Black has not progressed much while White is fully prepared to display activity on the kingside. Bach-Urban, Hamburg 1999 saw 17 f5!? lDb618 .ig5 Wlb719 fxg6 hxg6, when White could have underlined his advantage by 20 a5 lDc8 21 lDc3 Wla6 22 Wld2 Wld3 23 \Wcl±, intending .ih6. 12 ... tild7 13 .ld2 tilb6

Black is impatient to shift the bishop from c4, but probably he should have activated his queen first: 13 ...Wla5 14 !%a3 lDb6 15 b3 (15 .ib5.ia6ii5). 79

Part 3

14 b3

White completed his development and seemed to have restricted any Black counterplay. As a rule, he is clearly better in such structures. The concrete position, however, hides an excellent counterchance: 14 ... f5! It is not clear how White should react then. The trick is that 15 e5 would stumble into 15 ... e6!! 16 dxe6 ltJxc4 17 bxc4 i.xf3 IS ~xf3 dxe5 19 ltJd5 e4 t. 15 ltJgS .icS is not an option either. So White should probably choose IS ~e2 fxe4 16 ~xe4 ltJxc4 17 bxc4 .icS with double-edged play. This brings us back to White's 14th move. Apparently the light-squared bishop is important and should be preserved by 14

SO

.ibS!. Then 14.. .fS7 would face IS ltJgS .icS 16 as, so Black will have to play 14 ....ia6 IS i.xa6 l"lxa6 16 ~e2 ~aS 17 b3 l"lbS IS l"lac1, when IS ... c4 loses material to 19 ltJbS. In the game Black enters a similar position with a tempo down. 14 ....ia6? 15.ixa6 ~xa616 Wie2 ~a8 17 ~ac1 Wid7 18 ltJb5

18 ... ~fc8 After this move Black is lost, but lS ... ltJa619 fS is rather grim, too. 19 .ixb4 cxb4 20 Wie3 Wid8 21 ltJfd4 .ixd4 22 ltJxd4 ltJd7 23 ltJc6 Wie8 24 e5 ltJc5 25 exd6 exd6 26 ltJe7+ ~g7 27 f5 f6 28 fxg6 hxg6 29 ~xf6 ~xf6 30 ~f1+ Wg7 31 Wid4+ ~h7 32 Wih4+ Wg7 33 Wif6+ 1-0

Part 4

1 d4 ltJf6 2 c4 g6 3 ltJc3 .ig7 4 e4 d6 5 f4 0-0 6 ltJf3 Lines without 6 ... cS QUICK REPERTOIRE

Until now we have only dealt with Modern Benoni or Volga lines, where Black attacks the centre by ... c7-cS. In this part we finally reach the authentic KID waters. White players should be glad to face plans with ... eS. I feel that they are strategically dubious as they lead to positions with serious spatial advantage and a mobile centre for White. It is no chance that Black plays 6 ... CtJa6 only once in five games. At the same time, he scores decently and some authors fervently advocate it. I suppose that the main reason for that is because White happens to underestimate it and has not clear notion what the correct move order is. Let us examine first lines without 6 ... CtJa6. The general rule is: if Black delays too much ... e7-eS or deprives his f6knight of his best retreat square, d7, we thrust eS ourselves and launch a direct kingside attack:

1. 6 ... CtJbd7 7 eS CtJe8

8 h4 cS 9 hS cxd4 10 lMrxd4 CtJc7 11 hxg6 fxg6 12 .ie3 CtJe6 13 lMrdS CtJb6 14lMre4 .id71S O-O-Ot.

2. 6.•• CtJc6 7 eS!

Black's last move is a provocation. Do not fall for it by pushing 81

Part 4 7 dS. Our general plan is to strike on the right wing so we have to win space there and not on the other side. 7... ttlg4 8 !e2 f6 9 eM with an overwhelming space advantage. 3. 6... eS?! 7 dxeS! dxeS 8 l'!xd8 9 ttlxeS

~xd8

12 if2! This bishop is important to White as it supports the attack on the queenside. Now Black is helpless to stop it: 12 ... aS 13 0-0 ttla6 14 a3 ~e71S l'!bl b6 16 b4±, Moskalenko-Ermenkov, Wijk aan Zee 1992. These examples shed some light on the popularity of: 6 ••• c!Lla6

Black has not enough compensation: 9... ttla610 i.e3 l'!e811 !e2 ttlxe4 12 ttlxe4 f6 13 ttlxf6+;!;.

Black keeps all his options open. He waits for us to define our plans. Then he can choose to attack by either ... cS or ... eS. I recommend: 7 id31

4.6 ...i.g4 A consistent move. Black wants

to contest the central dark squares. However, he defines too early his plans so we can adjust our own actions accordingly: 7ie2 Against 6... ttla6 we will develop the bishop on d3, but now we have to ensure protection of d4. 7 ie3 pursues the same goal and it is also a good option: 7... ttlfd7 8 h3 .txf3 9 ~xf3

In practice Black chooses here 7... eS or 7...ig4, but White keeps an advantage due to his terrific centre.

7... ttlfd7 Again, we meet 7... ttlbd7by 8 eS ttle89 0-0. 8 i.e3 eS (8 ... ttlc6 9 eS) 9 fxeS dxeS 10 dS hi3 11 hi3 i.h6

5. 7... eS 8 fxeS dxeS 9 dS White has a clear plan for expanding on the queenside with a3, b3, l'!bl, b4, cS, while the opponent

82

5 f4 0-0 6 tt'lf3 Lines without 6 ... c5 cannot claim the same about his kingside pawn majority.

6. 7... e5 8 fxe5 dxe5 9 d5 tt'lc51O .tc2 as 11 0-0 tt'le8 12 i.g5 f6 13 i.e3

Black would be overrun if he allowed c4-c5 so his first task should be to prevent it. Only then he can think about counterplay with ...f7fS, but in most cases this thrust only would leave e5 hanging. Naturally, Black should not be able to build an attack on the wing when the centre is so mobile. The other big plus of White is the ability to quickly relocate his queen from one part of the board to the other with the manoeuvres ~dl-el-h4 or ~dl-a4-a3. 9 ... c51O 0-0 tt'le8 11 a3 tt'ld6 12 tt'lb5!?

We can also leave the bishop on the h4-d8 diagonal, but this is more efficient when Black's knight is on f6. In that case the pin is very awkward and we should maintain it. In the diagram position both sides have not much of a choice: 13 ...b6 14 a3 tt'ld6 15 tt'ld2 f5 16 ~e2 tt'ld7 17 b4 tt'lf6 18 c5 bxc5 19 bxc5;!;, Kahn-G. Pinter, Budapest 2000.

White is on top. He succeeds in removing all the blocking pieces the d6-knight and the c5-pawn, e.g. 12 ... tt'lxb5 13 cxb5 tt'lc714 .tg5 f6 15 .te3 b6 16 b4±.

7. 7... e5 8 fxe5 dxe5 9 d5 c6 cxd5 11 cxd5 tt'le8

Black prevented c4-c5 indeed, but he lost a tempo on ... c6 and opened the c-file. The diagonal a3f8 is also very weak and we should aim to put a bishop on c5. White obtains the better game 83

Part 4 with simple developing moves. 12 ie3lLld6 13 ~a4!t I have tested this novelty several times in blitz. The queen is heading for a3 from where it will pressurise d6. 8. 7... ig4 8 ie3lLld7

Pay special attention to this diagram as White's best move here is far from obvious: 9 ie2!! What is this, an admission of an earlier mistake?! Of course not. We have a tremendous centre and should we succeed in protecting it until we consolidate, we would enjoy a lasting edge. The text is aimed against 9 0-0 e5 10 fxe5 c5 with some chances for equalisation. Note that popular 9 h3? is a mistake. 9 ... c5 10 d5 hf3 11 hf3 ~a5 12 id2 ~b4 13 ie2t, Kahn-Rajlich, Budapest 2000. Black's pieces will soon be repelled back. 9. 7... ig4 8 ie3 c5 9 d5 e6 10 h3 exd5 11 cxd5 hf3 12 ~xf3

84

This position is rather topical as it is recommended for Black by both Golubev and Galagher. 12 ... lLlb4 13 ibl ~e8 I have also analysed 13 ... lLld7!? 14 0-0 ~a5 15 e5! dxe5 16 f5lLlf6 17 ie4 ~ad818 ~adl ~d619 a3lLla6 20 g4±. 14 a3 White has a clear advantage, see game 13 Banikas-Delithanasis, Kavala 1997. Golubev's suggestion 14 ... lLlfxd5 15 exd5 id4 fails to the simple 16 lLle4 he3 17 ~xe3 lLlxd5 18 ~f3 f5 19 O-O±.

Conclusion: The plans without 6 ... c5 are much less challenging. They offer White a great mobile centre. It is important to remember that we meet 6... lLlbd7 or 6 ... lLlc6 by 7 e5! and 6... lLla6 by 7 .td3. In the latter case critical are key lines 8 and 9. It is essential to avoid blockade on the central dark squares and keep our pawn chain flexible. I believe that everybody would like to play the FPA if he knew that Black would avoid transiting to the Modern Benoni.

Part 4

1 d4 ~f6 2 c4g63 ~c3.tg74e4 d6 5 f4 0-0 6 ~f3 Lines without 6 ... cS

STEP BY STEP

6 ... ~a6

This is the most fashionable alternative to 6 ... cS. The earliest example I see in the database is DannerHurtado, Thessaloniki 1984, but in this game Black did not actually show any new plan as he met 7 .id3 by 7... cS 8 dS CiJc7 with a passive position. The game which busted the interest towards the knight move was Arkhipov-Belov, Moscow 1987. The Russian master played 7....ig4 8 0-0 eS 9 fxeS and here he triggered the mine of his homework preparation: 9 ... CiJd7! 10 .ie3 cS

A picturesque position. Gradually the surprise effect has faded away and later many first players took

courage to test Black's idea more thoroughly. White may be even slightly better after 11 dS, but the structure is too static to my taste. I prefer to avoid it. Before focusing on Belov's plan, let me examine the alternatives: a) 6 ... eS?! 7 dxeS dxeS 8 ~xd8 l"1xd8 9 CiJxeS Black has not enough compensation: 9 ... CiJa6 10 .ie3 l"1e8 (or 1O ....ie6 11 .ie2 CiJb4 12 l"1cl CiJd7, MetgeRogers, Auckland 1992, 13 CiJxd7±) 11 .ie2 CiJxe4 12 CiJxe4 f6 13 CiJxf6+ hf6. Now simplest is 14 0-0 heS IS fxeS l"1xeS 16 .ih6 .ifS 17 .if3 c6 18l"1adU. b) 6 ... a6 This is a bad version of 6 ... cS 7 dS a6, since White can answer 7 .id3, when 7... cS is no longer possible due to the simple 8 dxcS dxcS 9 eS CiJfd7 (9 ... CiJhS 10 0-0 CiJc6 11 .ie4) 10 0-0 CiJb6 11.ie3±. So Black should switch to the idea with 7....ig4 8 .ie3 CiJfd7, but obviously ... a6 does not help him much. 8S

Part 4

9h3 This is my favourite treatment of such positions. I prefer to force the exchange in order to be able to meet ... e5 by dxe5 followed up by f5. The alternative 9 ~e2 allows 9 ... c5 (9 ... ~h6 10 0-0 e5 11 dxe5 dxe5 12 tt::lxe5 tt::lxe5 13 hg4 'lWxd114 hdl tt::lxc4 15 ~cU) 10 d5 b5 which would justify Black's 6 ... a6. 9 ... hf3 10 'lWxf3 tt::lc6 11 'lWf2 e5 12 dxe5 dxe513 f5 tt::ld4 Rybka likes to sacrifice the exchange so it contemplates 13 ... tt::lc5 14 hc5 'lWxd3, but White has a curious way to activate the hI-rook: 15 hf8 (15 Eldl 'lWxc4 16 hi8 Elxf8 17 'lWe2 'lWc5 18 Eld5 'lWe719 0-0 tt::ld4 20 'lWc4 c6 21 'lWc5 'lWc7 22 'lWd6 'lWc8 23 f6 hf6 24 'lWxf6 cxd5 25 tt::lxd5 tt::lc6=) 15 ... Elxf816 'lWe2 'lWg3+ 17 mfl tt::ld4 18 Wff2 'lWd3+ 19 mgl ~h6 20 Eldl 'lWxc4 21 h4 b5 22 Elh3t. 14 a-a-a!?

86

White's game is much easier. He will be advancing his kingside pawns while Black has not clear counterplay. His knight on d4 is mostly decorative and can be easily exchanged. As a rule, White is better. The game Avshalumov-Loginov, Budapest 1990 continued 14...b5 15 g4 c6 16 g5 f6 17 h4 tt::lc5 18 ~c2±. c) 6 ... tt::lbd7 Black wants to push 7 ... e5, but his last move deprived the f6-knight of good retreat squares. I propose to take advantage of that and stake even more space in the centre: 7 e5 tt::le8

8h4 8 c5!?, which restricts Black's counterplay, is much safer and ensures some edge: 8 ... dxc5 (or 8 ... c6 9 ~e3 b6 10 cxd6 exd6 11 ~d3 ~b7 12 a-at) 9 dxc5 c6 10 ~e3 tt::lc7 11 ~c4t. The extreme passivity of the enemy forces, however, suggests sharper means. 8 ... c5 9 h5 cxd4 10 'lWxd4 tt::lc7 11 hxg6 fxg6 12 ~e3 tt::le6 13 Wfd5 tt::lb6 14 'lWe4 ~d715 a-a-at. d) 6 ... tt::lc6 This move is really ugly. His only "value" is that it offers White too many options and the most obvious

5 f4 0-0 6 lLlf3 Lines without 6 ... c5 one, 7 d5, is probably worst. I like: 7 e5 lLlg4 8 ie2 f6 9 e6:t with an overwhelming space advantage, but 7 ie2 e5 8 dxe5 dxe5 9 '@'xd8 ~xd81O fxe5lLlg411 ig5 ~d712lLld5 lLlgxe5 13 O-O-O:t is also rather attractive, despite the simplification. e) 6 ...ig4 A consistent move. White trades his bishop in order to free room for the rest of his pieces. He stakes on dark-square strategy, hoping to exploit the weakness of d4. I shall examine two good plans for White: el) 7 ie3 lLlfd7 8 h3 ixf3 9 '@'xf3

9 ... c5 We already know from line c that 9 ... lLlc6 10 0-0-0 e5 11 dxe5 dxe5 12 f5 lLld4 13 '@'f2:t is in White's favour, for instance, 13 ... c6 14 g4 '@'a5 15 g5 ~fd8 16 h4 lLlc5 17 'it>bl b5 18 cxb5 cxb5 19 ig2 lLla4 20 f6 if8 21 lLld5+-, Glek-Damljanovic, Belgrade 1988. 10 d5 '@'b6 Alternatively, 1O ... lLla6 11 ie2 '@'b6 (11 ... lLlb4 12 ~c1) 12 id2 lLlb4 13idU. 11 '@'f2 11 O-O-O?! hc3 12 bxc3 '@'a5 13 'it>b2 b5 14 h4 occurred in the crazy game Gretarsson-Forster, Bermuda

1999, which went 14 ... lLlb6 15 id2 lLl8d7 16 'it>al ~fb8 17 h5 lLlxc4 18 hxg6 fxg6 19 '@'h3 lLlf8 20 f5oo. The text is more solid and allows White to consolidate. It is good for him as his spatial advantage would be working in his favour in the long run. 11 ... '@'b4 Or 11 ... lLla6 12 '@'d2 e6 13 dxe6 fxe6 14 ie2:t. 12 '@'c2:t. Everything is protected now and Black's pieces on the queenside will soon be repelled. e2) 7 ie2lLlfd7 We meet 7... lLlbd7 by 8 e5lLle8 9 0-0 (9 h3 hf3 10 hf3 ~b8 11 0-0 c512 dxc5 dxe513 ie3 ih6f±) 9 ... c5 10 dxc5 dxe5 l1lLlxe5 he2 12 '@'xe2 lLlxe5 13 fxe5 '@'d4+ 14 ie3 '@'xe5 15 '@'f3:twith tangible pressure. 8ie3

8 ... e5 If Black delays too long this move, we can deprive him of it at all: 8 ... lLlc6 9 e5 dxe5 10 fxe5 f6 11 e6lLlb6 12 d5±. 8 ... c5 definitely does not fit well with Black's setup: 9 d5 '@'a51Oid2 hf3 11 ixf3 lLla6 12 a3 lLlc7 13 0-0 '@'a6 14 ie2 id4+ 15 'kt>hl ~ae8 16 b4:t, Pesorda-Singer, Graz 1998. 9 fxe5 dxe5 10 d5 hf3 11 ixf3 87

Part 4 ~h6. Black got a chance to activate his dark-squared bishop, but at a price - he lost the battle for the centre. White should avoid the proposed trade, because even on f2 his bishop would be obviously more useful than the black one. 12 ~f2! as 13 0-0 lLla6 14 a3 fie7 15 ~b1 b6 16 b4±, Moskalenko-Ermenkov, Wijk aan Zee 1992.

Now it is time to deal with the main subject of this part - 6 ... lLla6. I prefer to meet it by: 7id3 I first met this variation in 1989. When Hebden put his knight to a6, I recalled the game of Belov and decided to radically cut across Black's idea: 7 eS?! lLld7 I wanted to launch an attack by h4, but the immediate 8 h4 cS 9 dS dxeS 10 hS did not look too convincing to me. Golubev suggests 1O ... fiaS and White will hardly prove sufficient compensation. So after having banned ... e7-eS, I chose to impede c7-cS, too! 8 cS?! dxcS 9 dSlLldb810 h4?! c6 1~ hSlLlb412 hxg6 hxg6

without a computer, Semkov-Hebden, Villeneuve Tolosane 1989, Informator 47/712. My idea looked so natural and the game, albeit drawn, was so wild and entertaining, that unexpectedly I found a surprising number of followers. Eventually it turned outthat 9... lLlb6! as in Videki-Gallagher, Kecskemet 1990, is in Black's favour. Have in mind though that the idea of both eS followed up by cS could be rather effective, as we saw in line c: 6... lLlbd7?! 7 eS lLle8 8 cS!? After 7 ~d3, Black is at a juncture.

In practice he chooses: A. 7... eS B. 7...~g4

13 e6 fxe6 14lLleS, when it is impossible to say what is happening 88

7... cS is also seen, but it has no theoretical significance. After 8 dS Black can transpose to line B by 8 ... ~g4. Alternatives are 8 ... e6 9 0-0 exdS 10 cxdS ~g4 11 h3 hf3 12 fixf3 lLlb4 13 ~bU or 8 ... lLlc7 9 0-0 a6 10 eS lLlfe8, when simplest is 11 h3 bS 12 b3t with an overwhelming space advantage.

5 f4 0-0 6 tDf3 Lines without 6 ... c5 A. 7 ••• e5 8 fxe5 dxe5 9 d5

The easiest way to kill the blocker. If Black attempts now 12 ...f5 13 i.g5 '!Wd7 14 tDxd6 W1xd6, White will use the absence of the a6-knight from the kingside with 15 '!WeI! f4 (15 ...h616 W1h4) 16 W1h4 gt717 g3! to crush the opponent on the allegedly "his" wing: 17... fxg3 18 hxg3 i.d719 i.h6±. Inkiov-Belkhodja, Evry 2005 saw instead: 12 ... tLlxb5 13 cxb5 tLlc7

Black tries to prevent c4-c5 by: A1.9 ... c5 A2.9 ... tDc5 A3.9 ... c6

Here simplest would have been 14 i.g5 f6 15 i.e3 b6 16 b4± with a lasting advantage.

A2. 9 ... tLlc5 10 .tc2 a5 11 0-0 tLle8

A1.9 ... c5

Now Black's pieces are passive and the only question is would White break through the blockade, or not. 10 0-0 tLle8 11 a3

White can also start with 11 i.g5 f6 12 i.h4 tDac7?! 13 a3 tDd6 14 b4 b6 15 bxc5 bxc5. One of my blitz games went on with 16 tDxe5! '!We817 tDc6 f5 18 i.g3 (18 i.e7+-) 18 .. .f4 19 gxf4±, Semi (2681)-Falstaf (2767) 3m + Is Playchess.com, 2008. 11 ... tLld612 tLlb5!?

Black aims to put his pieces on dark squares and thrust f5. Ifhe delays the knight retreat, the pin on g5 would be awkward, as in the game Zimmerman-Salai, Presov 1997: 11 ...W1e7 12 @h1 i.d7 13 i.g5 h6 14 i.h4 ga6 15 b3 g5 16 i.f2 i.g4 17 i.g1 tDh5 18 a3 tDd719 tDb5 c5 20 d6±. In the diagram position White should decide where to develop his bishop - on the h4-d8 diagonal, or on e3. I think that the latter option is more natural, but it would be 89

Part 4 good to provoke .. .f6 first: 12 ~g5 f6 13 ~e3 The game Ostenstad-Carlsen, Trondheim 2004 saw 13 ~h4 'Wffe7 14 'Wffe2 ~d7 IS b3 lLld6 16 a3 lLla6, when White should have shifted his knight to d3 via el. (17 lLle1!?) Instead, he chose 17 ~h1 ~h8 18lLlbS, but after 18... 'Wffe8 it turned out that 19lLlxd6 cxd6 20 'Wffd2 could be countered by 20 ...bS~. 13... b614 a3lLld615lLld2 f516 ~e2 lLld7 17 b4 lLlf6 18 c5 bxc5 19 bxc5

Both sides have completed their plans and White appears to be on top. For instance, 19 ... ~a6 20 ~d3 hd3 21 'Wffxd3 fxe4 22 'Wffe2 lLlfS 23 ~gS e3 24 lLlde4 'Wffc8 2S he3;!; regains the pawn with an edge. In the game Kahn-G.Pinter, Budapest 2000 Black opted for: 19 ... lLldxe4 20 lLldxe4 fxe4 Now 21 ~b3;!; ~h8 22 ~gS would have underlined White's edge.

A3.9 ... c6 Knaak puts an exclamation mark to this move in his 200S survey. It is hardly any better than the other

90

options though. White obtains the better game with simple developing moves. 100-0 The attempt to pin the f6-knight by 10 ~gS fails to 1O ... h6! 11 ~h4 'Wffb6! 12 'Wffe2 lLlcS. 10 ... cxd5 11 cxd5 lLle8

Black prevented c4-cS indeed, but he lost a tempo on ... c6 and opened the c-file. The diagonal a3f8 is also very weak and we should aim to put a bishop on cS. 11 ... 'Wffb6+ 12 ~h1lLlg4 would not have helped Black as 13 'Wffe2lLlcS14 ~c4 ~d7 IS ~gS fS 16 exfS± opens play in White's favour. 12 ~e3 lLld6 12 ... 'Wffe7 is similar: 13 ~xa6 bxa6 14 'Wffa4 lLld6 IS lLld2!±, BanikasGurcan, Yerevan 2000. 13 ~a4! I have tested this novelty several times in blitz. The queen is heading for a3 from where it will pressurise d6. 13 1':\c1 ~d7 14 'Wffe2 lLlc7 IS ~cS lLlce8 16 lLlbS hbS 17 hbS;!; is also better for him. 13 ... ~d7 13 .. .fS only serves to weaken Black's centre: 14 1':\ac1 ~d7 IS 'Wffa3

5 f4 0-0 6 tLlf3 Lines without 6 ... c5 fxe4 16 tLlxe4 tLlf5 17 .if2 tLld4 18 tLlfg5±. 14 ~a3 ~b8 15 i.xa6 bxa6 16 i.c5 llJb5 17 llJxb5 i.xb5 18 ~f2 ~d8 19 llJg5 ~d7 20 i.e7! This move fixes White's advantage. 20 1Wh3 h6 21 tLlxV1Wc8 221We6 1Wxc5 23 tLlxe5+ ~h8 24 a4 .ixe5 25 1Wxe5+ ~h7oo is pretty unclear. 20 ... ~b6 21 d6±.

B. 7 ...i.g4 8 i.e3 8 0-0 llJd7 9 .ie3 e5 10 fxe5 c5 should not be completely discarded as an option. I think that White has a pleasant game after both: 11 dxc5!? dxc512 .ie2 tLlc7, PeevSpasov, Tsarevo 2001, 131Wa4! .ixf3 14 gxf3 tLlxe5 15 f4 lLlc6 16 .ixc5 .id4+ 17.ixd4 tLlxd4 18 E:f2± or 11 d5 tLlxe5 12 .ie2 tLlxf3+ 13 .ixf3 .id7 14 1Wd2 1We7 15 .ig5 1We5 16 .if4, Golubev gives here 16 ...1We7? as in his blitz game Vaisser-Golubev 2004, but it stumbles into 17 e5! .ixe5 18 .ixe51Wxe5 19 E:ael1Wg7 20 lLle4±. Of course Black must play 16 ... 1Wd4+ 17E:f2 .ie5 18 .ixe51Wxe5 and he can hold the position, but this structure is known to be very passive for Black.

Bl. 8 ... e5 B2.8 ... tLld7 B3.8 ... c5 81. 8... e5?! fxe5 dxe5 10

d5~

This structure is better for White as a rule. The bishop on g4 makes Black's task even more difficult: 10 ... c6 1O ... lLlh5 is anti-positional, since Black needs this knight for attack and should not try to trade it for the d3-bishop. White has a serious advantage after 11 cS! (or 11 0-0 lLlf412 cS!) 11...lLlf4 12 0-0. 11 0-0 cxd5 12 cxd5 llJe8, (12 ... tLlhS 13 1Wa4±) 13 ~a4±, see game 14 Stocek-Folk, Ceska Trebova 2007. 82. 8... llJd7 9 i.e2!

91

Part 4 I remember that I analysed this move back in the 1980ies when Belov's plan with 6 ... tLla6 had just appeared. It is aimed against 9 0-0 eS 10 fxeS cS. It is true that White had just developed this piece on the previous move, but who can claim that the black knight stands better on d7 instead of f6! Note that 9 h3? (Knaak) 9 ....b:f3 10 ~xf3 cS! 11 dS is bad due to 11...tLlh4 12 ib1 ~aS (12 ...bS, Dorfanis-Spasov, Kavala 2001 is unclear) 13 0-0 tLlb6 and White loses material: 14 ~e2 ~a61S a3 tLlxc416 axb4 ~xa117~xc4 ~xb2+, ChytilekBabula, Czechia 1999. 9 ... c5 9 ... ixf31O .b:f3 eS is commonly faced with 11 dxeS dxeS 12 fS;!; while 9 ... eSlO fxeS dxeS11 dSleads to another pawn structure which is even more pleasant: 11... ixf3 (Or 11.. .fS 12 tLlgS f4 13 ixg4 fxe3 14 ie6+ ~ h8 1S tLlf7+±) 12 .b:f3 ih613 if2! fS14 O-O±, Jianu-Economescu, Bucharest 2008. 10 d5 .ixf3 11 .ixf3 ~a5 12 .id2 ~b4 13 .ie21;

rary. His pieces will soon be repelled back. In the game Kahn-Rajlich, Budapest 2000, Black attempted to complicate things by 13 ... ~xb2 14 1'!b1 ~a3, but 1S eS± came extremely unpleasant.

83. S... c5 Recommended by both Golubev and Galagher. 9 d5 e6 10 h3 exd5 11 cxd5 .ixf3 12 ~xf3

12 ... llJb4 12 ... 1'!e8 13 0-0 tLlb4 is hardly any better due to the tempo 14 ibS! 1'!e7 1S eS dxeS 16 ixcS;!;. 13.ib1 1'!eS I have also analysed 13 ... tLld7!? 14 0-0 ~aS 1S eS! This thematic thrust is very good here. 1S a3 misses the opportunity and Black takes over the initiative after 1S ... 1'!ae8 16 ~f2 tLla6 17 eS dxeS 18 fS e4 19 tLlxe4 ~b6CX) 20 d6 ~xb2.

It looks like Black obtained some counterplay, but it is only tempo-

92

1S ... dxeS 16 fS tLlf6 17 ie4 1'!ad8 181'!ad11'!d6 19 a3 tLla6 20 g4± Remember this position! It illustrates White's main strategical goal in the Modem Benoni structure.

5 f4 0-0 6 tDf3 Lines without 6 ... c5 tDxd5 17 ~xd5 !'1xe3+ 18 @d2;!;:) but instead of the passive 16 tDdl, White has the simple 16 tDe4 he3 17 ~xe3 tDxd5 18 ~f3 f5 19 O-O±. 15 0-0 ll)d7 16 eS!

Black might continue further with 20 ... !'1fd8 (20 ... c4 21 g5 tDh5 22 f6 ~h8 23 ~g4 tDc5 24 ~c2±) 21 fxg6 fxg6 22 ~e2 @h8 23 ~g2. The threat of ~g5 is very awkward for Black. 14 a3 WaS Golubev advocates here as an improvement "the bizarre" 14 ... tDfxd5 15 exd5 ~d4, (or 15 ... hc3+ 16 bxc3

White has a clear advantage, see game 13 Banikas-Delithanasis, Kavala 1997.

93

Part 4

1 d4 ttlf6 2 c4 g6 3 ttlc3 J.g7 4 e4 d6 5 f4 0-0 6 ttlf3 Lines without 6 ••• cS COMPLETE GAMES

13. Banikas - Delithanasis Kavala 1997 1 c4 g6 2 ttlc3 J.g7 3 d4 d6 4 e4 ttlf6 5 f4 0-0 6 ttlf3ttla6 7 J.d3 J.g4 8 J.e3 c5 9 d5 e6 10 h3 exd5 11 cxd5 J.xf3 12 ~xf3ttlb4 13 J.b1 ~e8 This is one of the most testing line against White's setup. The text move is a double-edged decision since it weakens the f-file and especially the V-square. 13 ... liJd7!? is more restrained, but it allows White to castle and follow up with his main plan: 14 0-0 ~aS1S eS! 14 a3 ~a5 14 ... liJfxdS fails to 1S exdS .td4 16liJe4±. 150-0

94

15 ... ttld7 White is fully developed now and he is aiming to open the f-file. Another good idea is the blockading pawn sacrifice eS, followed up by fS. Even the prophylaxis Wh1, aimed againsttactical tricks along the g1-a7 diagonal, would be very unpleasant. Black has no adequate counterplay. For instance, the thematic 1S ... c4 16 fS! liJd3 only loses a pawn after 17 bd3 cxd3 18 fxg6 fxg6 19 .td4l'!e7 20 ~xd3±. Black decides to discourage fS, but then the other central breakthrough gains in strength: 16 e5! c4 16 ... dxeS 17 fS is thematic ... and awful for Black:

5 f4 0-0 6 ttJf3 Lines without 6 ... c5 17... e4 18 ixe4 ttJd3 19 fxg6 fxg6 20 ~f7+ It>h8 21 ~xd7 ttJxb2 22 :1l:ael ttJd3 23 :1l:c2 :1l:ad8 24 ~xb7 :1l:b8 25 ~d7 :1l:bd8 26 ~g4 ixc3 27 ~h4+­ or 17... ttJf6 18 ie4 ttJa6 19 fxg6 fxg6 20 d6+-. 17 e6

White has an overwhelming advantage in the centre so he only needs to open up play. 17.. .fxe6 18 dxe6 lLlf6

Or 18 ... :1l:xe6 19 f5. 19 Wlxb7 lLlbd5

Black is beyond salvation, for 19 ... :1l:ab8 loses just another pawn after 20 ~f7+ It>h8 21 ~xa7±. 20 .id4 Wlc7 21 Wlxc7 lLlxc7 22 .ia2

The best move order was 22 f5 gxf5 23 :1l:xf5 :1l:xe6 24 ia2+-. 22 ... lLlxe6

The point is that 22 ... d5 fails to 23 f5+-. 23 .ixc4 @h8

Black's last chance to put up some resistance was 23 ... d5! 24 ixf6 dxc4± with "only" one pawn down.

Such positions with doubled apawns arise often in this variation. As a rule, White is much better because he has two excellent squares for his minor pieces - c5 and c4. Right now he could underline that by 15 b4 E1c8 16 ~a3 f5 17 ic5 ~f6 18 E1ael!± - there is no reason to "give" the bishop for the rook as it will hardly run away. This is a recent game, databases are abundant with practical examples, and I really cannot understand what Black wanted to achieve when choosing this variation. 15

~ac1 ~b8

16 b3

24 .ixe6 ~xe6 25 ~ae1 ~xe1 26 ~xe1 lLlg8 27 ~d1 ~b8 28 .ixg7+ @xg7 29 ~xd6 ~xb2 30 ~d7+ @f6

White demonstrates that Black has no play along the b-file, but 16 ttJd2± would have been fine, too.

31 g4 ~b3 32 lLle4+ @e6 33 lLlc5+ 1-0

16 ... lLlc7 A grim alternative is 16 ... ixf3 17

:1l:xf3 E1c8 18 b4 f5 19 ic5 E1xc5 20 bxc5 ~xc5+ 21 It>hl ttJd6 22 ~c6±. 14. Stocek - Folk Ceska Trebova 11.08.2007 1 d4 lLlf6 2 c4 g6 3 lLlc3 .ig7 4 e4 d6 5 f4 0-0 6 lLlf3 lLla6 7 .id3 .ig4 8 .ie3 e5 9 fxe5 dxe5 10 d5 c6 11 0-0 cxd512 cxd5 lLle8 13 Wla4 Wld614 .ixa6 bxa6

17 lLld2 .id7 18 lLlc4

The opening is over and White directly simplifies to a winning endgame. 18 ....ixa419 lLlxd6 .id7 20 .ixa7 ~a8 21 .ib6 lLlb5 22 lLlcxb5 axb5 23 ~c7 ~a6 24 .ic5 .ig4 25 ~f2 1-0

95

BLACK DEVIATIONS FROM THE MAIN MOVE ORDER

The best part of having the FPA in one's repertoire is that Black will often try to outsmart you with tricky move orders and second grade systems which commonly delay ... ttlf6 or ... g6. They are called "Modern", but perhaps "Dubious" would have been a more appropriate name. I do not claim that White can refute them. I only assert that these offbeat systems are less challenging and offer White a much wider choice of good plans. As a rule, White can pick between solid development that should ensure him a slight edge, and more testing and sharper, though doubled-edged variations. I will propose a repertoire based on my own preferences, but unlike the first part of the book, here the variations are seldom forceful and on many occasions White will have other worthy continuations. So if you already have your own favourite systems against the Modern Defence, I do not see any reason for you to renounce them. Any centre-oriented natural development should give White a pleasant game. The paramount difference between the various offbeat Black's setups is the position of the darksquared bishop. It can go to g7 or e7. The latter option is rather pas96

sive and its popularity has faded away. I deal with it in Part 7. The fianchetto with an early ... ttlc6 is subject of Part 5, while Part 6 considers the Classical King's Indian with ... ttld7. I think that White cannot, and should not, avoid this specific type of the KID since it hides no venom. White should obtain a stable advantage in the opening without risking the dreadful KID attacks from the most topical lines with ... ttlc6. There are a number of other curious systems for Black, but I cannot encompass all of them. For instance, 1 d4 d6 is best met with 2 e4!, followed by ttlc3. Chess Stars has published a whole 400-pages -thick book on these lines - An Opening for White According to Anand, volume 4. I'd like to add that many players (and I myself) love the fianchetto system against the KID, but they refrain from trying it often because of the possibility of Black switching to Gruenfeld setups. This is not a problem after 1 d4 d6 which allows 2 ttlf3 followed up by g3. This move order also enables interesting setups without ttlc3. They are of independent significance and Black is usually less prepared to face them adequately. In the next parts I will only focus on setups where White plays c4 and an early ttlc3.

PartS

1 d4 g6 2 e4 ~g7 3 c4

QUICK REPERTOIRE

In these systems Black tries to attack early the d4-square, thus preventing White from building a broad pawn centre. He can achieve that by putting his knight on c6 immediately, or after the preliminary 3 ... d6. 1. 3... lDc6 4lDf3! e5 (4 ... d6 5 d5 lDe5 6 lDxe5) 5 dxe5 lDxe5 6 lDc3 d6

7... lDxf3+ 8 gxf3! lDf6 (or 8 ... lDe7 9 'l&d2 0-0 10 0-0-0 lDc6 11 h4---+) 9 \&d2 with attack. 2. 3 ... d6 4 lDc3 lDc6 (4 ... lDd7 5 lDf3 is subject of the next part.) 5 d5! (5 ~e3 lDf6! would face us with the main KID lines) 5... lDd4 6 ~e3 c5 (or 6 ... e5 7lDge2t) 7lDge2

7 ~e3!?

We have the better centre and good prospects of opening files on the kingside. Our own king will be safe on the other wing: 7... lDe7 8 lDxe5 he5 9 'l&d2 0-0 10 0-0-0 ~e6 11 h4 f5 12 ~g5 f4 13 h5±, Piasetski-Lacasta Palacios, Sants 2006;

White is slightly better due to his space advantage and good prospects on the kingside. Black's attempt to complicate things by 7... ~g4 8 f3 hf3? 9 gxf3 lDxf3+ 10 It>f2 lDe5 is easily refuted by 11 'l&a4+. 7... lDxe2 8 he2 lDf6 9 0-0 0-0 allows him to trade one minor piece and complete development, but his future is 97

PartS not bright at all. We can expand on the kingside by f4, and then, g4-gS orf4-fS. In that case it would be better for us to keep the centre closed. Black's problem in that structure is the lack of counterplay. 10 f4 e6?! was bad for him in Vaisser-Turner, rapid Athens, 1997: 11 dxe6 he6 (11 ...fxe6 12 eS±) 12 fS±. Staying passively is also not too attractive, as it can be seen in game 15 Vaisser-Barlov, Las Palmas 1995 which went on 10 f4 lLle8 11 fS !eS 12 ~d2 lLlf6 13 !h6 ~e8 14 g4. Remains: 7... ~b6 8lLla4 ~aS+ 9 !d2

Lately Black has been experiencing problems here after both 9 ... ~d8 10 !c3 eS 11 dxe6lLlxe6 12 hg7 lLlxg7 13 ~d2 lLlf6 14 f3 !e6 1S lLlf4 ~e7 16 O-O-O±, Gelfand-Azmaiparashvili, Dortmund 1990, and 9... ~a6 10 lLlxd4 hd4 l1lLlc3. See

98

game 16 Belov-Bologan, Plovdiv 2008 and 17 Zs.Polgar-Todorcevic, Pamplona 1990. In these two examples Black attacked the centre by his knight. He might also use his e-pawn, but the resulting endgame is rather unpleasant for him: 3. 3 ... d6 4lLlc3 eS S dxeS!? (This move is the main reason for most players to avoid 4 ... eS.) S... dxeS 6 ~xd8+ \!fxd8 7 f4 lLlc6 8 lLlf3

White achieves a lasting pull here with simple moves. He attacks the centre and tries to fix a target, commonly on the queenside. 8 ...!e6 9 fxeSlLlxeS1 0 !f4lLlxf3 + (1O ... lLlxc4 l1lLlbSlLlxb2 12 ~c1t) 11 gxf3 c6 12 0-0-0+ \!fe8 13 \!fc2 lLlf6 14 b3 ~d8. This position arose in Donner-Ivkov, Wijk aan Zee 1972. White is clearly better.

PartS

1 d4 g6 2 e4 ig7 3 c4 STEP BY STEP

It is a matter of taste whether to play 3 ttJc3 or 3 e4. The downside of 3 e4 is that it allows 3 ... ttJc6. As we shall see, this variation hardly equalises though. Now let us examine the move order with: 3 ttJc3!? Now 3 ... ttJc6 does not make much sense as White has the pleasant choice between 4 d5 ttJe5 5 e4 d6 6 ie2 ttJf6 7 f4 ttJed7 8 ttJf3 0-0 9 0-0 c6 10 ie3 cxd5 11 cxd5 ttJg4 12 id4 ixd4+ 13 ~xd4 ~b614 WTxb6;t, Mista-Knaak, Zinnowitz 1969 and 4 ttJf3 d6 5 e4 ig4 6 ie3 ttJf6 (6 ...ixf3 7 gxf3;t) 71e2 with an edge. The only reasonable deviation from my main line is 3 ... c5 4 d5 ixc3+ 5 bxc3 f5.

I can only say that it is not so easy to checkmate Black here as it seems at first glance: a) 6 e4 fxe4 7 f3!? e5; b) 6 ~a4 ttJf6 7 g3 0-0 8 ig2 d6 9 ttJh3 ttJa6 10 0-0 ttJc7 11 ~b3;t, Delchev-Seret, St. Affrique 2002; c) 6 h4!? ttJf6 (6 ... d6 led to a miniature in Karpov-William, Parsippany 1998: 7h5 ~a5 8 id2 gxh5 9 ttJh3 ttJf6 10 ttJf4 ttJbd7 11 e3 ttJe5 12 ttJxh5 ttJxh513 WTxh5+ ttJg614 f4 wd8 15 i.d3 id7 16 g4 ie8 17 WTxf5 ttJe518 fxe5 ,ig6 19 ~xg61-0) 7 h5 l:!g8!? (7... ttJxh5 lets in 8 e4 WTa5 9 exf5 ~xc3+ 10 id2 ~g7 11 id3±, Borchgrevink-Hersvik, Hamburg 1999) 8 hxg6 hxg6

99

PartS White has not shown a clear path to advantage, for instance, 9 1!h'a4 1!h'b6 10 ct:Jf3 ct:Je4 11 ct:Jg5 1!h'f6 12 ct:Jxe4 fxe4131!h'c2 d614 l"1b1 <;t>d8 15 i.e3 <;t>c7oo, Sakaev-Reinderman, Cappelle la Grande, 1995. I suppose that White should be better after 5 .. .f5, but this variation needs more testing. 3 ... d6 Alternatively: a) 3 ... ct:Jc6 is ten times less popular than the text, but it deserves attention. You should resist the temptation to push 4 d5 since the line 4 ... ct:Jd45 ct:Je2 c5 6 ct:Jxd4 cxd4 gives Black counterplay: 7 i.d3 d6 8 0-0 ct:Jf6 9 ct:Jd2 0-0 10 b4 as 11 bxa5 1!h'xa5 12 ct:Jb31!h'b6oo. I prefer 4 ct:Jf3.

Now 4 ... d6 5 d5 ct:Je5 6 ct:Jxe5 he5 should be pleasant to White due to his impressive pawn centre and the clumsy stand of the e5-bishop. The game B.Kovacevic-Biliskov, Zadar 2007 went on 7 i.d3 ct:Jf6 8 0-0 0-0 9 ct:Jd2 ct:Jh5 10 l"1e1 c6 11 h3 e6 12 ct:Jf3;!:;. 4 ... e5 is more interesting: 5 dxe5 ct:Jxe56 ct:Jc3 d6 7 i.e3!? The fine point is to wait for Black to define the position of his king's 100

knight. For instance, after 7... ct:Je7 we take on e5 and castle long: 8 ct:Jxe5 he5 91!h'd2 0-0 10 0-0-0 i.e6 11 h4 f5 12 i.g5 f4 13 h5± as in Piasetski-Lacasta Palacios, Sants 2006. If Black himself captures on f3, we open a file against his king: 7... ct:Jxf3+

8gxf3! ct:Jf6(or8 ... ct:Je791!h'd2 0-0 10 0-0-0 ct:Jc611 h4---+) 91!h'd2. This is more natural than 91!h'c2, intending to play l"1g1 and thrust f4-f5 which also deserves attention though. 9 ... 0-0 10 0-0-0 i.e6 11 h4 as 12 i.e2t. It is clear that White stays better all over the board. b) 3 ... c6 4 ct:Jc3 d5 5 cxd5 cxd5 is seldom seen. Of course, White can play 6 e5 with a stable space advantage, but I like 6 ct:Jxd5 e6 7 ct:Jc3 hd4. In this symmetrical pawn structure White is way ahead in development. He can use it by 8 1!h'a4 or 8 i.b5+ ct:Jc6 9 ct:Jge2 i.f6 10 0-0 1!h'xd111 l"1xd1 i.d712 f4 i.e713 i.e3±, Bareev-Seoev, Tallinn 1988. c) 3 ... c5 4 d5 e6 is an original treatment of Modern Benoni ideas. The fine point is that Black might develop his knight on e7 instead of

1 d4 g6 2 c4 i.g7 3 e4 the common f6-square. We should not insist on our repertoire with f4, since the peculiarity of the position gives Black extra options, e.g.: 5 f4 exd5 6 cxd5 d6 7 4:Jf3 4:Je7 S i.d3 0-0 9 0-0 Wb6!? 10 st>h1 i,xb2 11 i,xb2 Wxb2 with unclear consequences: 12 4:Jbd2 Wf613 e5 Wxf414 4:Je4 4:Jf5 15 4:Jf6+ st>h816 ~5 i,xf5. It is better to wait and see where the gSknight will go. 5 4:Jc3 exd5 6 cxd5 In general, I do not like symmetry, so I would not recommend 6 exd5 d6 7 i.d3 4:Je7 8 4:Jf3 0-0 9 0-0. 6 ... d67f4 The "normal" move here is 7 i.d3, but I do not see any reason to abandon the f4-plan. 7... 4:Je7 S h4!?

You will notice that I advocate this advance in all the variations where Black puts his knight on e7. His plan of pushing .. .f5 is not to be underestimated. At least, then the e7-knight will not be so awkward. The text cuts across Black's intention. S ... h5 This may weaken the kingside, but allowing White to open the h-

file would be a mistake: S ... O-O 9 h5 :ge8 (9 .. .f5 10 e5 dxe5 11 i.c4±) 10 i.e2 gxh5 (1O .. .f5 11 e5; 1O ... 4:Jd7 11 hxg6 4:Jxg6 12 f5 4:Jge5 13 i.h6 i.f6 14 Wd2±) 11 f5±. White's attack is running by itself in these examples. 9 i.e2 I toyed for a while with the seemingly attractive idea 9 4:Jf3 4:Jd71O g4 hxg4 11 4:Jg5 4:Jf6 12 e5 dxe5 13 fxe5 4:Jfxd5 14 i.c4 i.e6 15 4:Jxe6 fxe6 16 Wxg4 Wb617 i.b5+ st>dS 18 i.g5 st>c7 190-0-0, but Black is holding after 19 ... :gafS (or 19 ... :gad8 20 4:Jxd5+ 4:Jxd5 21 i,xdS+ :gxdS 22 Wxg6 i,xe5 23 a4 WaS 24 Wxe6 i,xb2+ 25 st>xb2 Wc3+=) 20 i,xe7 i.h6+ 21 i.g5 4:Jxc3 22 bxc3 i,xg5+ 23 hxg5 :gxh1 24 :gxh1 Wxb5 25 Wxe6 :gfl + 26 :gxfl Wxfl+=. 9 ... 4:Jd7 10 4:Jh3 4:Jf6 114:Jf2 The hole on g4 is well covered and White can go on with his main strategic idea of breaking through with e4-e5: 11 ... 0-0 12 0-0 a6 13 a4 Wc7 14 :gel :gbS 15 i.f3 b5 16 axb5 axb5 17 e5. White has the initiative, e.g. 17... 4:Jg4 IS i,xg4 i,xg419 4:Jxg4 hxg4 20 :ga6 :gfdS 21 4:Je4 4:Jf5 22 :gc6 Wa7 23 Wxg4 dxe5 24 h5---+. 4~c3

101

PartS 4 ... lLlc6

The other major option 4 ... ttJd7 5 ttJf3 is subject of the next part. I do not think that 4 ... e5 deserves much of attention. White can try to steer the game into the main line by 5 ttJge2 when Black might deviate not too successfully with 5 ... ttJe7 6 ~e3 (6 dxe5 dxe5 7 ~xd8+ \t>xd8 8 ~g5 f6 9 0-0-0+ ttJd7 10 ~e3 c6 11 h4 h5 12 g3t) 6 ... 0-0 7 h4 f5 8 dxe5 dxe5 9 ~g5 ~e81O h5t. The most testing continuation however is 5 dxe5!? dxe5 6 ~xd8+ \t>xd8 7 f4 ttJc6 8 ttJf3

8 ... ~e6 Or 8 ... ttJd4 9 \t>f2! ttJxf3 10 gxf3 ~e6 11 ~e3 ttJe7 12 Efd1 + ~c8 13 ttJd5 Efe8 14 Efg1 (14 ttJxe7+ Efxe7 15 f5 gxf5 16 Efg1 ~f8 17 ~h3 fxe4 18 he6+ Efxe619 E!:g8 E!:e8 20 fxe4 b6 21 E!:h8±) 14 .. .f5 15 ~h3 b6 16 b3±, Halkias-Krum Georgiev, Athens 1998. 9 fxe5 ttJxe5 10 ~f4 ttJxf3+ (1O ... ttJxc411 ttJb5 ttJxb2 12 E!:c1t) 11 gxf3 c6 12 0-0-0+ ~e8 13 ~c2 ttJf6 14 b3 E!:d8. This position arose in Donner-Ivkov, Wijk aan Zee 1972. White is clearly better. He has a mobile pawn centre, connected rooks and he is able to quickly relocate 102

his pieces for attack on both wings. Black's queenside is left without defence and that could be underlined by 15 ~g2!t ~e716 ~g5 h617 ~e3 b6 18 e5 ~f5+ 19 ~b2 ttJd7 20 f4 f6 21 hc6 with an edge. The possibility of 5 dxe5!? is the main reason for most players to avoid 4 ... e5.

5 d5! Perhaps the exclamation mark will surprise some readers. White often uses 5 ~e3 or 5 ttJge2. These moves may not be any worse than the text, but they have a major drawback - they allow the opponent to throw us outside of our repertoire. Suppose we choose: 5~e3

The overwhelming majority of games see then 5 ... e5?! 6 ttJge2

1 d4 g6 2 c4 ~g7 3 e4 In my opinion, White is already better. It is late for 6 ... ttJf6 in view of 7 d5 ttJe7 8 f3 0-0 9 g4 and Black is cramped in the back two lines: 9... ttJe8 10 h4 f5 11 g5;!; (11 gxf5±) 11.. .f4 12 ~f2, Balashov-Potapov, Vladimir 2008. That is why he commonly chooses: a) 6... ttJh6 7 f3 f5 8 d5 ttJe7 9\Wd2 ttJf7 (or 9 .. .f4 10 ~f2 g5 11 c5) 10 0-0-0 with a clear plan for collecting something on the queenside. I have pleasant memories in this position from a tournament where I achieved my second GM norm. My opponent was one of the most devoted KID adepts GM Krum Georgiev. The game did not last long: 1O .. .f4 In the past years Black has not come up with a reasonable plan. For example, the game Malakhatko-Simutowe, Panevezys 2008 saw instead 10 ... a6 11 @bl ~d7 12 c5 0-0 13 ttJc1 dxc514 hc5 b615 ~a3 ttJd6 16 ~d3 and White went on to win. 11 ~f2 g512 h3 (12 c5±) 12 ...h513 @bl ttJg6 14 c5

White's spatial advantage makes his attack much more efficient. 14... g4 15 hxg4 hxg4 16 gxh8+ hh8 17 lDb5!± a6 18 \WaS @d7 19 c6++-, Semkov-Kr.Georgiev, Plov-

div 1988. b) 6... exd4 7 ttJxd4 ttJge7

8h4! Beware the trap 8 ~e2 0-0 9 0-0 f5 10 exf5 hd4! 11 hd4 ttJxf5=. I first witnessed it in the game TalAzmaiparashvili, Albena 1984. I also played in that tournament and watched the postmortem analysis. Then I decided that White cannot aspire to the advantage by castling short. I'll repeat here something I had already said before: When Black develops his knight on e7 in the Modern Defence, White should always contemplate h4 as the most testing answer. 8 .. .f5 Black fairs a little better after 8 ...h5, if the tiny 31% can be called better at all: 9 \Wd2 ttJe5 (or 9 ... 0-0 10 0-0-0 ttJxd4 11 hd4 hd4 12 \Wxd4 ttJc6 13 \We3 ~e6 14 ~e2 ttJe5 15 c5±, Stangl-B. Schneider, Germany 1989) 10 0-0-0 a6 11 @bl ~d712 ttJf3 \Wc8, Norris-McNab, Glasgow 2000, when 13 c5± would have fixed White's edge. 9h5 gf8 From the top players only Mamedyarov tries to defend this variations lately. In his previous game he 103

Part 5 chose 9 ...fxe41O hxg6 hxg611 Ei:xh8+ hh8 12 CLlxe4 d5 (12,..CLlf5 13 ~g5 ~d714 CLlxf5 ~xf5 15 ~e2 ~e6 16 0O-Ot) 13 CLlxc6 bxc6 14 ~d4 ~g4 15 ~xg4 hd4 16 cxd5 hb2 (16,..cxd5 17 CLld6+ ~xd6 18 ~xd4t) 17 Ei:dl (17 Ei:bl!? Ei:b8 18 dxc6±) 17,..cxd5 18 ~b5+ mf8 19 ~e6 with strong initiative, Kramnik-Mamedyarov, Moscow 2007. Undoubtedly Anand expected this opening and he came well prepared 3 days later, even for a blitz game! 10 hxg6 hxg611 Ei:h7 Ei:t7

This is not a problem if you like the Saemisch with ... CLlc6 or the main line Classical with 6 ~e2 0-0 7 CLlf3 CLlg4, but I assume that you read my book to find a way to avoid the big theory. Of course White can play 7 d5 CLle5 8 h3, but I'm not sure I like this. The same reasoning applies to 5 CLlge2. Black should answer 5,..CLlf6! when the Saemisch would be White's best choice. Now let us return to the main line. 5 ..• tL'ld4 6 i.e3

12 Ei:xg7!! Ei:xg7 13 ~g5~ ~d7 14 CLlxc6 (or 14 ~d2 CLlxd415 ~xd4 CLlc6 16 ~e3 ~e6 17 O-O-O~) 14,..CLlxc6 15 CLld5 and White went on to win, Anand-Mamedyarov, Moscow 2007. c) 6,..f5 7 exf5 gxf5 (7,..hf5 8 d5 CLlce7 9 CLlg3 CLlf6 10 ~e2 0-0 11 0-0 c5 12 ~d2 ~d7 13 f3 Ei:ae8 14 a3 a6, Csom-Sax, Budapest 1977 is positionally grim for Black following 15 b4 b6 16 CLlge4t) 8 dxe5 dxe5 9 ~xd8+ CLlxd8 10 f4 CLle6 11 0-0-0 CLlf6 12 h3 ~h6 13 g3t, B. Kovacevic-Jurkovic, Nova Gorica 2002. White's pieces are clearly better coordinated. Then why I do not recommend 5 ~e3 for our repertoire?! Because Black can answer 5 •.. CLlf6! 104

6 .•. c5 6,..e5 plugs the diagonal to the KID's bishop on g7 so it is obviously bad. White follows up with: 7 CLlge2 CLlxe2 The piece sac 7,..~g4 hf3? 9 gxf3 CLlxf3+ 10 mf2 ~f6 is purely speculative. White easily neutralised the attack in Volkov-Montero Martinez, Dos Hermanas 2004 after 11 CLlg3 CLld4+ 12 mg2 h5 13 h4 ~h6 14 hd4 exd4 15 CLlb5 ~e3 16 CLlxc7+ md7 17 CLlxa8 ~f2+ 18 mh3 ~f419 ~b3+-.

1 d4 g6 2 c4 i.g7 3 e4 7 ... cS 8 dxc6 CLlxc6 9 CLlbS i.f8 10 iWd2;!; also looks awful for Black. 8 i.xe2 White is clearly better due to his spatial advantage and the inevitable breakthrough c4-cS with pressure on the queenside. Even the exchange of the bad bishop does not help Black much after 8 ... i.h6 9 iWd2 he3 10 iWxe3 CLlf6 11 cS±. The other attempt to make use of the rare move order, 8 .. .fS 9 exfS gxfS 10 i.hS+ @f8 11 f4±, also favours White. Perhaps the lesser evil is 8 ... CLle79 cS fS, but then the simple 10 f3± underlined White's edge in Polugaevsky-Gojak, Sarajevo 1964.

his dark-squared bishop. At least he would achieve some structural benefits and eliminate White's potentially most dangerous piece: 8 ... hc3+ 9 bxc3 \WaS 10 iWc2 CLlf6 (10 .. .fS opens up play in favour of Whites bishop pair. White was clearly better in the game TimmanHennig, Breda 2001 after 11 f3 CLlf6 12 0-0 i.d7 13 i.h6 0-0-0 14 exfS hfS 15 iWd2±.) 11 0-0 i.d7 12 a4 0-0-0 13 Eifb1 @b8, Lazarev-Horn, Biel 1994. While White's advantage is beyond doubt, he will have to work hard to press it home. An alternative is 8 .. .fS?! 9 exfS hfS 10 i.d3 iWd7 11 O-O;!;. 90-00-0

7c!lJge2

A. 7 ... CLlxe2 B. 7... iWb6 7 ... i.g4 8 f3 i.xf3? 9 gxf3 CLlxf3+ 10 @f2 CLleS is easily refuted by 11 iWa4+, but 11 CLlf4 CLlf612 h3±, should also be good enough, Brodsky-Mihajlovskij, St. Petersburg 2006.

A. 7 ... c!lJxe2 8 .b:e2 c!lJf6 I suppose that should Black yield so much space, he should also take more strategic risks by parting with

White has a stable spatial advantage. Obviously he should expand on the kingside by f4, and then, g4gS or f4-fS. In that case he should like to keep the centre closed so variations like 10 eS CLld711 exd6 exd6 are not too consistent. Black will hardly generate efficient counterplay. He might try to open a file himself by ... e6, but White will take on e6 and the d6pawn would become a nice target. In that case Black misses his queen's 105

Part 5 knight, which is very important in such structures, for instance, 10 f4 e6?! 11 dxe6 he6 (11 ... fxe6 12 e5±) 12 f5 ic8, Vaisser-Turner, rapid Athens, 1997, 13 fxg6 fxg6 14 e5!±. If Black trades his knight, he would also land into a murky position: 10 'Wd2 LtJg4 11 hg4 hg4 12 ih6 a6 13 hg7 cJ1xg7 14 'Wf4 id7 15 e5±, Cs.Horvath-Strozewski, Velden 1995. Now I will examine White's most straightforward plan: 10 f4 LtJe8 In the game Kozhuharov-Kukov, Sunny Beach 2007, Black tried to anticipate f4-f5 by 1O ... LtJd7 11 'Wd2 f5?! and was much worse following 12 exf5 gxf5 13 E1ael±. 11 f5 ie5 12 §'d2 LtJf6 13 .ih6 ~e8 14 g4 See game 15 Vaisser-Barlov, Las Palmas 1995 in the "Complete Games" chapter.

9 •••§'a6 Alternatively: a) 9 ... 'Wd8 10 ic3 (10 LtJxd4 cxd4 11 c5 only gives Black counterplay in the centre, for example, 11 ... id7 12 E1c1 E1c8 13 b4 LtJf6 14 f3 0-0 15 id3 e6, Lazarev-Abolianin, Belgium 1995) 1O ... e5 11 dxe6 LtJxe6 12 hg7 LtJxg7 13 'Wd2 LtJf6 14 f3 ie6 15 LtJf4 'We716 O-O-O±, Gelfand-Azmaiparashvili, Dortmund 1990. b) 9 ... 'Wc7 10 ic3 e5 (As in other similar positions, 1O ... ig4 11 f3 ixf3 12 gxf3 LtJxf3+ 13 cJ1f2 LtJe5 is not satisfactory for Black. I faced it in the game Semkov-Gratseas, Heraklio 1993 and went on to win after 14 LtJg3 LtJh6 15 ie2±. Rybka claims that 14 ih3± is even better. 10 LtJxd4 .ixd4 11 LtJc3 §'b6

B. 7 ...§'b6 8 LtJa4 §'a5+ 9 .id2

12 ~b1 White has the better prospects. See game 16 Belov-Bologan, Plovdiv 2008 and 17 Zs.PolgarTodorcevic, Pamplona 1990.

106

Part 5

1 d4 g6 2 e4 i.g7 3 c4

COMPLETE GAMES

15. Vaisser-Barlov Las Palmas 1995 1 d4 d6 2 e4 g6 3 c4 .ig7 4ltJc3 ltJc6 5 d5 ltJd4 6 .ie3 c5 7 ltJge2 ltJxe2 8 .ixe2 ltJf6 9 0-0 0-0 10 f4 ltJe8

11 f5!

I chose to comment this game because it illustrates one of the most fundamental plans in the FPA. Inexperienced FPA players might think that White should try to prepare e5, but this advance is essential in the Modern Benoni pawn structure where Black has not an e-pawn. In the current situation, it is correct to gain more space on the kingside

by f5. Thus White limits the impact of the black light-squared bishop, opens the route to his own one to h6, and enables an attack on h7 by ~e1-h4 and the rook lift Elf3h3. Black is practically deprived of counterplay. White can eventually use his g-pawn to repel the knight from f6. Barlov decides to redeploys his pieces by putting the bishop before the knight. In the game Kozhuharov-Kukov, Sunny Beach 2007 Black chose 1O ... ttJd7. Then 11 f5 should also be good, but White counted on the fact that the opponent cannot continue developing and preferred the useful move 11 ~d2. Black answered with 11...f5?! and was much worse following 12 exf5 gxf5 13 Elael±. All the examples show Black struggling without any counterplay. However, the attempt to undermine White's centre is not too helpful either: 1O ... e6?! 11 dxe6 he6 12 f5! (again this move!) 12 ... i.c8, Vaisser-Turner, rapid Athens, 1997, when simplest would have been 13 fxg6 fxg6 14 e5!±. 107

PartS 11 ... ie512 ~d2 White could have forced Black's knight to take a passive stand on g7 by 12 .th6liJg713 g4, but Vaisser apparently prefers to see it on f6. 12 ... liJf6 13 ih6 ge8 14 g4 'itlh8 Black's position is already very dubious. (But my computer still likes it!) His king looks defenceless and it is unclear how to protect it. Queenside counterplay is not effective. For instance, 14 ... a6 15 .tf4 .td4+ 16 @hl WfaS 17 Wfel bS 18 gS hc319 bxc3liJxe4 20 fxg6 hxg6 21 .td3. 15 'itlh1 id7

do. The rest is more or less clear: 21 ... bxc4 22 ixc4 ~b6 23 liJe2 ie5 24 ic3 geb8 25 b3 ~d8 26 ixe5+ dxe5 27 ~h4 ie8 28 fxg6 fxg6 29 ~h3

29 ... h5 30 gxh6 ~d6 31 h7 liJf6 ~h6 id7 33 ~g3 liJg4 34 ~h4 ~f6 35 ~xg4 ixg4 36 ~xg4 ~f8 37 1'!g 1 'itlxh 7 1-0 32

The computer overestimates the threat on e4. Indeed, Black can win that pawn with IS ... WfaSI6 .tf4 hc3 17 bxc3liJxe4, but after 18 Wfc2 gxfS 19 gxfS liJf6 20 l"Iael.td7 21.td3 bS 22 .tgS± all the open files work in White's favour. 16 if4 id4 17 g5 liJg8 18 id3 ~a5 19 gf3 a6 20 ~e1 b5 21 id2 Black is already lost. My impression from the very persistent and straightforward play of Vaisser is that he had won a fair number of blitz games in this structure and had a very clear notion of what to 108

16. VI.Belov-Bologan Plovdiv 22.04.2008 1 d4 g6 2 e4 ig7 3 c4 d6 4 liJc3 liJc6 5 d5 liJd4 6 ie3 c5 7 liJge2 ~b6 8 liJa4 ~a5+ 9 id2 ~a6 10 liJxd4 ixd4 11 liJc3 ~b6 12 1'!b1

Simple and good. In the next commented game White chose 12 Wfb3?!. 12 ... liJf6 13 id3 liJd7

1 d4 g6 2 c4 ~g7 3 e4 13 ... 0-0 gives White more options: 14 ttJe2 ttJd7 IS ttJxd4 (IS f4;);) IS ... cxd4 16 0-0 ttJeS 17 ~h6 l"&d8 18 h3;);. White's pawn structure is more flexible. 140-00-015 ttJe2 ~g716 ~c3 e5 17 dxe6 fxe6 18 ~xg7 @xg7

19 VNd2 The diagram position is critical for this variation. In my opinion, White would have much better chances after 19 f4! eS 20 fS±. His main advantage is not the stranded d6-pawn, but the weak black king. His attack would be stronger with a pawn on fS. In the game White allowed 19 ... eS, obviously counting on his pressure on d6. It is possible though that with a stubborn defence Black might be able to hold on. On many occasions later White was rather hesitant and avoided critical decisions. No wonder he did not beat a player like Bologan. 19 ... e5 20 ttJc3 Here was the second chance for White to use his better development with 20 f4 exf4 (20 ... ttJf6 21 fS ~d7 22 ttJg3±) 21 ttJxf4 ttJeS (21...ttJf6 22 ttJdS 1lNd8 23 b4±) 22 ttJhS+!? gxhS 23 1lNgS+ ttJg6 24 l"&xf8 rJtxf8 2S eS

dxeS (2S ... rJtg7? 26 hg6 hxg6 27 1lNf6+ rJth7 28 1lNf7+ rJth6 29 1lNf8+ rJth7 30 l"&f1) 26 l"&f1 + rJte8 27 1lNxhS 1lNd6 28 hg6+ 1lNxg6 291lNxeS+ ~e6 30 1lNxcs rJtd7. White has full compensation for the piece and he can force a draw whenever he wants. A computer should be able to hold Black's position though. 20 ... ttJf6 21 .ie2 Again 21 f4 would have posed tactical problems to Black: 21... exf4 22 l"&xf4 ~e6 23 l"&h4. Now the natural 23 ...1lNc7 would face 24 eS! dxeS 2S hg6! hxg6 261lNh6+ rJtf7 27 1lNh7+ ttJxh7 28 l"&xh7+ rJtg8 29 l"&xc7 b6 30 b3 with a stable advantage, but 23 ... rJtg8! should neutralise White's initiative. 21 ....ie6 22 ~fd1 ~ad8 23 VNe3 ~d7 24 h3 VNd8 25 ~d2 VNe7 26 l"&bd1 b6 27 b3 ~fd8 28 ~d3 ttJg8!

29.ig4 Black has devised an excellent manoeuvre: ttJf6-g8-h6-f7 from where the knight would not only defend d6, but it would be trying to continue his route to d4 via gS-e6. I think that White should have regrouped, too: 29 1lNc1 ttJh6 30 l"&1d2 ttJf7 311lNd11lNh4. Now 32 ttJbS1lNxe4

109

Part 5 33 i.f3 iMfh4 34 i.c6 i.fS 35 i.xd7 i.xd3 36 fi:xd3 fi:xd7 37 ttJxd6 ttJxd6 38 fi:xd6 fi:xd6 39 iMfxd6 would be close to a draw, 32 fi:g3 ttJgS 33 i.g4 hS 34 i.xe6 ttJxe6 would be fine for Black, but 32 i.g4 i.xg4 33 hxg4 ttJh6 34 fi:g3 iMfgS 35 ttJbS ttJf7 36 fi:gd3 iMfe7 would be similar to the game with some extra tempi for White. He can try then to open files on the queens ide with 37 a4 h6 38 as. Belov however shows that he is content with the half point: 29 ...i.xg4 30 hxg4 ttJh6 31 fle2 flh4 32 ~g3 flg5 33 ttJd5 ~g8 34 ~f3 ~f8 35 ~xf8 ~xf8 36 g3 %-% 17. Zsuzsa Polgar-Todorcevic Pamplona, 1990 1d4g62c4i.g73~c3d64e4 ~c65d5~d46i.e3c5 7 ~ge2f1b6 8 ~a4 fla5+ 9 i.d2 fla6 10 ~xd4 i.xd4 11 ~c3 flb6 12 flb3?!

It is a positional mistake to trade queens here. White's space advantage can be exploited to build a kingside attack or maintain tension in the centre. Perhaps Todorcevic wanted to keep more pieces on the board, hoping to outplay his young opponent by tactical means. 12".fld8 13 i.d3 i.g7 14 0-0 ~f6

15

~ae1

It is not sure that this rook be-

longs to el. In some cases White can need it on d1, in others, as in the game Vaisser-Barlov, she may be able to win even without its help. In short, 15 f4 deserved attention, e.g. 1S,..ttJg4 16 eSt. 15,..0-016 f4 e6 We had seen that Black has not good chances to survive with a passive stand in the centre. In case of 16,..ttJd717 ~h1 White should adopt the plan of Vaisser and shift the queen and a rook to the kingside, for instance, 17,..a6 18 iMfd1 fi:b8 19 fi:e3 iMfc7 20 iMfe1 bS 21 b3. 17 dxe6 In general, this exchange is correct. However White has a better option, based on his lead in development. She should have opened the centre by 17 eS! dxeS (17,..ttJe8 18 i.e3 b6 19 i.e4±) 18 fxeS ttJg4 19 i.f4 f6 (19,..fS 20 h3 ttJh6 21 i.c2±) 20 ttJe4 ttJxeS 21 ttJxf6+ with a clear edge. 17 ... fxe6 18 fld1

White's main advantage in this structure is that she has a mobile centre that could move forward at 110

1 d4 g6 2 c4 .ig7 3 e4 any moment. Black should be ready to meet not only the obvious threat e4-eS, but f4-fS as well. The latter is usually more dangerous since it weakens Black's castling position and may lead to a crushing attack. However, if Black gets enough time to complete his development and consolidate, it would be difficult to break his defence. Therefore White should play concretely in order to convert his dynamic advantage, while Black must aim to anticipate the opponent's threats. Now 18 .. .tLld7 looks best against eS or fS, but White has additional resources. I like 19 ~g4 ~e7 20 h4! ~f7 21 ~h3, threatening hS, e.g. 21...b6 22 eS dxeS 23 hS. This idea is easy to find if we are aware that our goal should be the black king and not the d6-pawn! 18 ... a6 19 .ie3 b6 This move misses the hit 20 eS!, but I do not know what to suggest instead: 19 ... lLld7 20 .ic2lLlb8 21 eS± and 19 ... eS 20 fS are grim enough. 20 ~U2?! It looks like Zsusza overestimates her positional advantage and hopes to convert it by simply doubling on the d-file. It was a fine moment for 20 eS! which would have faced Black with a difficult choice between 20 ... lLle8 21 .ic2 ~e7 22 lLle4 dxeS 23 fxeS .txeS 24 ~xf8+ ~xf8 2S ~fl ~e7 26 .igS ~c7 27 ~f3± and 20 ... dxeS 21 fxeS lLld7 22 .ie4 ~a7 23 ~d6 lLlxeS 24 ~xf8+ ~xf8 2S ~xb6 ~f7 26 .txcS lLlxc4 27 .ixf8lLlxb6 28 .txgn. In both examples White would transform his dy-

namic advantage into more stable achievements. 20 ....ib7 This move is not precise. 20 ... ~e7! would have allowed Black to finally consolidate. Then 21 fS ( or 21.ic2 ~b8 22 ~d2lLle8) 21...lLld7! 22 fxe6 ~xe6 23 eS .ib7 24 exd6 would not favour White due to 24 ... lLleS 2S .ic2 ~ad8 26 ~xf8+ ~xf8 27 c;t>hllLlf3 with a counterattack.

21 gd2?! This is a conceptual mistake. White's rook was better on the ffile! 21 fS!, threatening with 22 eS, was consistent and good. 22 ... lLld7 22 fxe6 ~xf2 23 .ixf2 lLlf8 24 lLldS lLlxe6 would not equalise in view of the weak f6-square: 2S ~g4 .ic8 (2S ... lLld4 26 .ih4±) 26 .ih4 gS 27 eS dxeS (27 ... gxh4 28 lLlf6+) 28 ~hS h6 29 .ig3±. 21 ...V!1c7 22 .ic2?! Time and again Zsuzsa misses 22 fS! ~f7 (22 ... gxfS 23 exfS dS 24 ~f2±) 23 fxg6 hxg6 24 .ic2±. 22 ... gad8 23 f5 Played at the most inappropriate moment! Now Black had the counterthrust 24 ... dS 2S cxdS (2S fxe6 d4 26 e7! ~xe7 27 .txd4 ~f7oo) 2S ...

111

PartS exdS 26 i.b3 c4 27 i.c2 EJ:de8 with double-edged play. 23 ... gxf5?! 24 exf5 exf5 25 .ixf5 i>h8

White still has some advantage and now she should take some measures against the counterplay with ...bS. Perhaps simplest would be 26 a4;!;, while 26 i.e6 bS 27 cxbS axbS 28 tLlxbS 1M'c6 29 a4 dS 30 i.gS;!; is more testing. However she was in time trouble and allowed the opponent to alter the pawn structure in his favour: 26 gf1?! b5 27 tLld5 tLlxd5 28 cxd5 V!Je7 (28 ... i.c8!) 29 ge2 f;lfh4

112

The rest of the game is quite chaotic. White plugged the d-file and finally turned her attention to Black's king. After mutual mistakes Black lost the tactical clash and the game. 30 if2 V!Jc4 31 .ie6 f;lfxa2 32 gfe1 V!Ja5 33 .ih4 gde8 34 .ig3 c4 35 ge4 gf6 36 gh4 gef8 37 h3 gg6? (37 ...1M'c7) 38 f;lfc2 gh6 39 gg4 f;lfc7 40 i>h2 ghf6 41 f;lfd2 as 42 gee4 .ia6 43 V!Je3 b4

44 .if5 (44 EJ:xg7!) 44 ... g6f7 45 ge6 f;lfa7 46 f;lfe4 .ic8 47 .ixd6 gxf5? (47... he6 48 dxe61M'b6=) 48 V!Jxf5! gxf5 49 ge8+ gf8 50 .ixf8 ie5+ 51 i>h1 1-0

Part 6

1 d4 g6 2 e4 ig7 3 c4 d6 4 liJc3 liJd7SliJf3 QUICK REPERTOIRE

In this part I consider the Classical KID with ... ttJd7. We have no reason to avoid it since it is the most passive variant of the KID and White's game is clearly better. Black also has a limited choice. Sometimes he puts his knight on e7, In all such positions I like plans with h2-h4: 5 ... e6 6 .te2 ttJe7. h6 8 .te3 fS 9 exfS ttJxfS 10 g4!± or S... eS 6 .te2 ttJe7 7 h4 h6 8 .te3 fS 9 dxeS ttJxeS 10 ttJxeS heS 11 \Mfd2±. S ... eS 6 ie2 ttJgf6 7 ie3 0-0 80-0

Black has two main plans here: He either opens the centre by taking on d4 and following up with ... c6 and ... dS, or he tries to close it and

push .. .fS. They lead to totally different structures. 1. 8 ... exd4 9 ttJxd4 Eie8 10 f3 c6 11 \Mfd2

l1 ... dS If Black misses the moment for this breakthrough, White answers 12 ttJc2, protecting the e3-bishop. Then his grip on the centre assures him oflasting pressure. 12 cxdS! (Do not allow the exchange sac on e3!) 12 ... cxdS 13 ttJdbS ttJb6 14 a4!? The threat of as forces play. Black has nothing better but to look for salvation in the endgame with an extra exchange for White: 113

Part 6 14 ... dxe4 IS lLlc7±.

~xdB

gxdB 16

2. B... c6 9 dS cS 10 lLlel lLleB 11 lLld3 11 g4 also gives White somewhat better chances, but play is more double-edged. l1 .. .fS 12 exfS gxfS 13 f4 e4 14 lLlf2 hc3 IS bxc3 lLldf6

I think that White is clearly better because the opponent cannot prevent g4 and lacks any counterplay. The best setup of White's pieces is probably: 16 ~el!i.d717h3lLlg71BlLldl ~aS 19 i.f2;!;, Obodchuk-Korobov, Moscow 2007. White can activate the f2-bishop from h4, the knight will support g4 from e3. White should not be in a hurry with this breakthrough though. He can patiently improve the position of his rooks first. See the commented game 18 Tisdall-Sutovsky, Gausdal 1995 which illustrates White's plan. Black also tries to maintain the tension with moves like B... aS, B... h6. In these cases White opens the d-file and expands on the queenside with cS: 114

3. B... h6 (threatening with 9 ... lLlg4 followed up by ... fS) 9 dxeS! dxeSlO ~c1 ~h711 gdl geB12 cS c6 13lLld2 ~e7

White has successfully redeployed his pieces. Now he can defend the cS-pawn with 14 b4± or offer Black the option of sacrificing the exchange with 14 lLlc4 lLlxcS IS lLld6±. Note that had Black weakened his queenside with B... aS, this plan would have been even more efficient. 4. B...geB 9 dS lLlg4 10 i.gS f6 11

i.d2

It is possible to adopt a passive stand on the kingside with 11 i.h4 hS 12 lLld2 lLlh6 13 f3;!;. However I prefer the method of example 2. We will take on fS and wedge our pawn 01} f4. Then we can stage c4-cS or hit

1 d4 g6 2 e4 tg7 3 c4 d6 4 lLlc3 lLld7 5 lLlf3 the black pawn centre: 11...f5 (Or 11 ... lLlh612 ~c2 lLli713 b4) 12 exf5 gxf5 13 lLlg5 lLldf6 14 h3 lLlh6 15 f4 ~e7 16 ~c2 @hS 17 !l:ael a6, Einarsson-McNab, Reykjavik 1993.

White can shatter Black's defence by sacrificing a piece: IS fxe5! dxe5 19 td3 e4 20 lLlcxe4!! fxe4 21 lLlxe4 lLlxe4 22 !l:xe4±. Finally we shall examine the only case when White should maintain the tension: 5. S ... lLlg4 9 tg5 f6 10 td2 lLlh6

The plan from example 3 11 dxe5 dxe5 12 b4 c6 13 c5 is possible here, too, but since Black is not threatening anything, we can make a couple of useful moves before defining our plans: 11 ~c2 c6 12 !l:adl ~e7 13 b4 lLli7 14 tc1 !l:eS 15 !l:fel lLlfS

We have a free hand on the left side of the board. Our main task now is to open the centre. That is typically achieved by the breakthrough c5: 16 b5 tg4 17 bxc6 bxc6 IS c5! dxc519 d5±. We can conclude that White achieves some advantage in this version of the KID with simple natural moves. We have nothing obligatory for memorising except for the moves in example 1. It seems that Black has been running short of new ideas lately.

115

Part 6

1 d4 g6 2 e4 .ig7 3 c4 d6 4 ttJc3 ttJd7 5 ttJf3 STEP BY STEP

S ... eS Sometimes Black tries to perplex the opponent by S... e6 6 ~e2 CfJe7. When I see a knight on e7, I always thrust 7 h4 to provoke weaknesses. After 7...h6 8 ~e3 fS 9 exfS CfJxfS 10 g4! (10 Wic2 CfJf6 11 CfJd2 slows down White's development and gives Black fair counterchances, e.g. 11 ... 0-0 12 g4 CfJxe3 13 fxe3 eS 14 Wixg6 exd4 1S exd4 Wie8 16 Wixe8 ~xe8 17 gS CfJhS with compensation, Skalik-Markowski, Suwalki 1999) 1O ... CfJxe3 11 fxe3±. Black is cramped and the g6-pawn is very sensitive. 6 .ie2 ttJgf6 We meet 6 ... CfJe7by7h4 as usual:

116

Now 7... h6 8 ~e3 fS 9 dxeS CfJxeS 10 CfJxeS heS11 Wid2 clearly favours White, as in Garcia Vicente-Timon Piote, St. Feliu 1994: 11 ... fxe4 12 hh6 CfJfS 13 ~gS e3 14 he3 CfJxe3 1S Wixe3±, so he chooses 7... exd4 8 CfJxd4 CfJc6, but then 9 ~gS is extremely awkward to meet. 9 ... CfJf6 (9 ... ~f6 10 CfJdS±) 10 CfJxc6 bxc6 stumbled into 11 eS! dxeS 12 Wixd8+ ~xd8 13 0-0-0+ ~e7 14 ~f3±, Rusev-Popchev, Plovdiv 2006. Popchev has been playing this rare variation for decades, but even he proves unable to show any decent idea against White's natural play. 7 .ie3 0-0 8 0-0

1 d4 g6 2 e4 .tg7 3 c4 d6 4 tLlc3 tLld7 5 tLlf3 This position is rather unpleasant for Black so I'm not surprised that he is trying nearly all possible moves. I will examine 6 of them: A. S ... exd4 B. S... a5 C. S ... h6 D. S ... tLlg4 E. S .. JleS F. S ... c6 You can find much more details on that variation in the Chess Stars book "Opening for White According to Kramnik", volume Ib by Khalifman. My favourite treatment is similar, except for a few lines which I attempted to fix. A. 8 ... exd4 9 tLlxd4 ge8 10 f3 c611 §'d2

This pawn structure is grim for Black if he misses the chance to open the centre, so the most popular continuation is: 11 ... d5 Alternatively: a) 11 .. :~e7 12 tLlc2 (protecting the e3-bishop) 12 ... tLle513 E1adl.te6 14 b3 E1adS 15 .tg5;l;, De Souza HaroVescovi, Registro 1997; b) 11 ... tLle5 12 E1adl

This is a natural setup of the white rooks, but 12 E1fdl, intending to retreat the e2-bishop to fl, also favours White: 12 ... a613 tLlc2!? (The game Lautier-Shchekachev, Dordrecht 2001 saw 13 .tfl Wie7 14 tLlb3 .te6, when instead of 15 c5?!, 15 Wixd6 Wixd6 16 E1xd6 .tfS 17 E1ddl hc4 IS tLla5 hf119 ~xfl E1abS 20 .ta7 would have been slightly better for White.) 13 ....te6 14 Wixd6 Wixd6 15 E1xd6 .tfS 16 E1d4±. 12 ...Wie713 tLlc2 .te6 14 b3 E1adS 15 .tg5;l;. White consolidated and retained his grip. 12 cxd5 Khalifman considers 12 exd5 cxd5 13 tLldb5 a614 tLld6, but the exchange sacrifice 14 ... E1xe3! 15 Wixe3 .tfS 16 c5 tLlxc5 17 Wixc5 hd6 IS Wid4 Wic7 would give Black significant compensation. It is better to keep the e-file closed. 12... cxd5 13 ltJdb5 ltJb6 14 a4!? 14 E1adl a6 15 tLla3 .td7, Watanabe-Diaz Hollemaert, Sao Paulo 200S, 16 .td4t also looks attractive, but the text is more forceful, since 14 ... a6 would fail to 15 a5. 14... dxe4 15 §'xd8 E1xd8 16 ltJc7 exf3 Or 16 .. JlbS 17 a5 tLlbd51S ha7±. 17 i.xf3 ltJc4 18 i.g5 IS .tc5 might be insufficient for winning after IS ....te619 tLlxaS E1xaS 20 hb7 E1bS 21 .ta6 tLlxb2 22 ha7 E1aS 23 .td4 E1xa6 24 E1fbl tLleS 25 hg7 tLlxa4 26 tLlxa4 ~xg7 27 tLlc5;l;. 18...i.f5 19 ltJxa8 gxa8 20 i.xb7 gb8 21 ltJd5 ltJxd5 22 i.xd5±. 117

Part 6 B. 8 .•• aS 9 dxeS!

White's plan of opening the dfile is quite good in most branches of the ttJd7-variation, but it is especially efficient when Black's queenside is weakened. 9 ... dxeS 10 1Mfc2 c6 11 lLla4 V!!e7 12 cS lLle8 13 ~ad1 lLlc7

the main line. 11 ~d1 lLld21Mfe7

~e8

12 cS c6 13

14 .ic4 lLle6 15 a3±

This position of the recent game Ivanisevic-Vajda, Herceg Novi 2008 hardly needs any comment. White will double his rooks on the d-file and will look around to collect material.

White discouraged the kingside attack and made good use of the weakness on h6 to achieve a clear advantage. Now he can defend the cS-pawn with 14 b4± or offer Black the option of sacrificing the exchange with 14 ttJc4 ttJxcS 15 ttJd6±. D. 8 .•• lLlg4 9 .igS f6 10 .id2 lLlh6

C.8 ... h6 This seemingly insipid move is actually very aggressive. Black is threatening with 9 ... ttJg4 followed up by .. .fS. It is time to open the centre: 9 dxeSdxeS

Khalifman-Yurtaev, Novosibirsk 1989 saw 9... ttJg4 10 exd6 ttJxe3, when best would have been 11 dxc7±, according to Khalifman. 10 1Mfc1 'it>h7

1O ... ttJg411 i.d2 hS12 h3 ttJgf613 i.e3 c614 cS V!1e71S b4;l; is similar to 118

White is faced with the dilemma: to maintain the tension in the centre or fix the pawn structure with 11 dxeS dxeS 12 b4 c6 13 cS. I definitely prefer White's game here,

1 d4 g6 2 e4 ig7 3 c4 d6 4 lLlc3 lLld7 5 IDf3 but Black has no weaknesses on the queenside and he is able to cover all the invasion squares along the d-file. The game Banikas-Damljanovic, Panormo 1998 saw further 13 ...'.We7 14 '.Wb3+ lLlfl IS ~adl ~e8 16 ie3lLlf8 17lLld2 ie618lLlc4 ~ad8 19 '.Wa3 ~xd120 ~xdl a6 21 '.WaS ~d8 22 ~xd8 '.Wxd8 23 '.Wxd8t. Perhaps we should take into account that the manoeuvre lLlf6-g4-h6-fl is passive and Black has no threats. Thus we can make a couple of useful moves like: 11 YNc2 c6 12 ~ad1 YNe7 13 b4 CiJf7 14 ic1 ~e8 15 ~fe1 CiJf8

11 id2

The retreat to h4 looks safer as it restricts Black's attack and leaves the bishop on the kings ide for defence: 11 ih4 hS 12 lLld2 lLlh6 13 f3t. However I prefer another approach. Instead of embracing a defensive setup on the right wing, we will counterattack there by f4. 11 ... CiJh6

The alternative is 11...fS 12 exfS gxfS 13 IDgS IDdf6 14 h3 IDh6 IS f4 '.We716 '.Wc2 l!ih817 ~ael a6, Einarsson-McNab, Reykjavik 1993.

We have a free hand on the left side of the board. Our main task now is to open the centre. That is typically achieved by the breakthrough cS: 16 b5 ig417 bxc6 bxc618 c5! dxc5 19 d5±.

E. 8 ••• ~e8 Now Black is seriously threatening to take on d4, so we should play: 9 d5 CiJg4 10 ig5 f6

White is fully mobilised and he can switch to fifth gear by sacrificing a piece: 18 fxeS! dxeS 19 id3 e4 20 lLlcxe4!! fxe4 21 IDxe4 lLlxe4 22 ~xe4 '.WcS+ C22 ... '.Wd7 23 ~h4+-) 23 ie3 '.Wxe3+ 24 ~xe3 ~xe3 2S ixh7 id726ig6±. 119

Part 6 12 YlYc2 It is always good to develop pieces, but 12 b4 is also possible. After 12 ... a5 13 a3 f5 14 ltJg5 ltJf6 15 c5 fxe416 cxd6 cxd617ltJcxe4ltJxe418 ltJxe4 i.f5, Evdokimov-Gorovykh, St. Petersburg 2007, White had 19 i.g5! Wb6 20 bxa5l'!xa5 21 i.e3 Wd8 22 ltJg3 with a huge advantage. 12.•• tDf7 Black cannot prevent b4 anyway: 12 ... a5 13 a3 ltJf8 14 b4 We7 15 l'!acl g5 16 c5 axb4 17 axb4 ltJg6 18 ltJb5 g4 19 ltJell'!d8 20 i.e3 ltJf4 21 cxd6 cxd6 22 i.b6 l'!f8 23 ltJc7 l'!a4 24ltJd3ltJxe2+ 25 Wxe2 f5, Zakharstov-Yanvarjov, Korolev 2000, 26 Wc2t. 13b4 White is way ahead with his play. In the game Harestad-Mikh.Ivanov, Gausdal 1997, Black bluffed here with 13 ...f5, when White should have taken the gift: 14 exf5 gxf5 15 Wxf5±, e.g. 15 ... e4 16 ltJg5 ltJf6 17 Wf4 ltJxg5 18 Wxg5.

F. 8 ••• c6 9 d5 White has completed development. The next stage of his plan is an expansion on the queenside so he should first define the situation in the centre. 9 ... c5 This unexpected turn of events forces White to reconsider his plans. Now he will have minimal chances for a quick queenside attack, while the opponent is only two moves away from taking over the initiative by .. .f5. Instead, 9 ... cxd5 10 cxd5 120

ltJg411 i.g5 f612 i.d2ltJh613 'lWc2 a6 14 a4 f5 15 ltJg5 ltJc5 16 i.c4 'lWe7 17 b4ltJxe418ltJcxe4 fxe419ltJe6 he6 20 dxe6 offered White rich play in the centre in the game KrasenkowArbakov, Moscow 1989. 10 tDe1 tDe8

11 tDd3 11 g4 (Khalifman's recommendation) is more topical. It also gives White somewhat better chances. Main answers are: a) 11... @h812 @hl f5 13 gxf5 gxf5 14 exf5, transposing to line c; b) 11 ... i.f6 12 Wd2 ltJg7 13 @hl i.e7 14 l'!gl @h8 15 ltJd3t, FaragoKaratorossian, Budapest 2004. c) 11 ... f512 exf5 gxf513 gxf5ltJb6 14ltJf3 i.xf5 15ltJg5

15 ... We7 15 ... ltJf6 is premature as White

1 d4 g6 2 e4 ig7 3 c4 d6 4 'i:lc3 'i:ld7 S 'i:lf3 can repel the b6-knight with 16 a4!? Wfe717 as 'i:lbd718 a6 b619 'i:lbS l"lfc8

20 f4 (20 whl±) 20 ... l"le8 (20 ... e4) 21 l"la3 'i:lf8 22 fxeS+-, Wells-Areshchenko, Coventry 200S. 16 Whl 'i:lf6 17 l"lgl wh8 18 Wfd2 l"lg8 In the game Kramnik-Knaak, Dortmund 1992, Black attempted to occupy the f4-square by 18 ... ig6 19 l"lafl 'i:lhS, but he was much worse after 20 'i:le6±. 19 l"lg3 ih6 20 l"lagl l"laf8 21 a4 (or 21 a3 'i:lbd7 22 b4 b6) 21...'i:lbd7 22 as a6 23 h4 'i:le4 24 'i:lcxe4 ixe4+ 2S f3 ifS 26 b4;!;, Paragua-Lin Chen Singapore 2007. White may be somewhat better in these examples, but his kings ide is weaken and play is double-edged as a whole. That's why I prefer the text since it is linked with a clear straightforward plan where Black is doomed to passivity. 11 ...fs12 exfs

This is more precise than 12 f4 fxe413 'i:lxe4 Wfe7!=, Semkov-Radulov, Bul ch. 1992. 12... gxfs 13 f4 e4 14 lLlf2 ~xc3

This bishop is useless as it is directed to the left wing where nothing is going to happen. On the contrary, White's knight is hitting e4 and it is a potentially dangerous piece. After 14 ... 'i:ldf6 IS whl a6 16 a4 b6 17 l"lgl l"la7 18 g4 wh8 19 gS White had a comfortable edge in Loncar-Bonetti, Genova 1999. 15 bxc3 'i:ldf6

This is the critical position for White's plan with 11 'i:ld3.

It is clearly better for White because Black is the passive side and he cannot generate any counterplay against the imminent g4. The only question is which is the best setup of White's pieces. In the first known to me game, Cafure-Panno, Buenos Aires 1965, White was too impatient to execute the thematic thrust g4 and was worse after 16 whl 'i:lg7 17 g4? fxg418 'i:lxg4 'i:lfS 19 l"lgl wh8't. Then in Filip-Langeweg, Amsterdam 1965, White showed the best plan. He retreated the e3-bishop to f2 and shifted the f2-knight to e3 via dl. The only flaw of his redeployment was that GM FIlip put his queen to d2. In our stem game Obodchuk-Korobov, Moscow 2007, White put it on el from where it helps in activating the f2-bishop from h4. See also the commented game 18 Tisdall-Sutovsky, Gausdal 1995 which illustrates how White can prepare and achieve g4 with natural, though not best, moves.

16 ~e1! ~d7 17 h3 lLlg7 18 lLld1 ~as 19 ~f2;t. Here Black at-

tempted to get some counterplay by the pawn sacrifice 19 ... e3, but 20 ih4!? 'i:le4 21 l"lf3! 'i:lhS 22 g3± completely disillusioned him. 121

Part 6

1 d4 g6 2 e4 ~g7 3 c4 d6 4 liJc3 liJd7 5 liJf3 COMPLETE GAMES

18. Tisdall-Sutovsky Gausdal 1995 1 d4 lLlf6 2 c4 d6 3 lLlc3 g6 4 e4 5 lLlf3 0-0 6 ~e2 lLlbd7 7 0-0 e5 8 ~e3 c6 9 d5 c5 10 lLle1 lLle8 11 lLld3 11 g4 is fashionable, but I prefer a more clear plan. 11 .. .f5 12 exf5 gxf5 13 f4 e4 14 lLlf2 ~xc3 15 bxc3 lLldf6

~g7

16 h3 Of course White's only active plan is linked with g2-g4-g5 so the text cannot be bad. However, the game Obodchuk-Korobov, Moscow 2007, showed a better implementa122

tion of that idea: 16 'lWeI! .id7 17 h3 lLlg718 't:Jd1'IWa5 19 .if2;J; intending to activate the dark-squared bishop fromh4. 16 ... ~h8 17 ~h2 ~g8 18 g3! Preventing any tactics. In the game Belunek-Sergeev, Czechia 2003 White played the reckless 18 8:g1, but Black missed his chance: 18 ... 't:Jg4+! 19 hg4 (18 ... hxg410ses to 19 .. 'lWh4++) 19 ... fxg4 20 't:Jxe4 'lWe7 21 8:e1 't:Jf6 22 't:Jg5 'lWg7 with serious counterplay. 18 ...'IlWe7 19 ~g1 'IlWg7 20 ~g2 'IlWh6 21 ~g1 tLlg7 22 'IlWd2 White could have already played 22 g4, but 22 ... 'lWh4 23 g5 't:Jfh5 24 'lWd2 't:Jg3 25 .id1 't:J7h5 would allow Black to find active places for his knights. Tisdall wisely mobilises his second rook. 22 ...'IlWg6 23 <j;>h1 'IlWh6 24 ~ag1 ~d7 25 ~h2 There was no need for this prophylaxis. 25 g4 was possible right away as 25 ... fxg4? loses material to 26 f5 'lWh5 27 't:Jxg4 't:Jxg4 28 hg4 'lWf! 29 8:fl 't:Jxf5 30 8:xf5 hf5 31

1 d4 g6 2 e4 .ig7 3 c4 d6 4 4:Jc3 tLd7 5 ~f3 8:f2±. The defence from the previous comment 25 ... iWh4 26 g5 4:Jfh5 also does not work in view of 278:h2 8:af8 (27 ... 4:Jg3+ 28 8:xg3 iWxg3 29 h4) 28 4:Jg4. Remains 25 ... iWg6 26 g5 4:Jfh5 27 4:Jg4! as in the game, but with important extra tempi. 25 .. .'!Wg6 26 g4 a6 27 g5 tDfh5 28 tDg4! b5 29 ~hg2 bxc4

30.ixc4? White achieved everything he was aiming for, but suddenly he commits a terrible positional mistake. After 30 4:Jh6 8:gb8 31 hc4 .ib5 32 .if1 (32 .ib3 as 33 c4 a4 34 cxb5 axb3 35 axb3 8:xb5 36 iWc3 is not too clear) 32 ....ixf133 8:xf1 8:b7 34 c4 White would have been clearly better. He needs only to shift his bishop on the main diagonal to finish the combat. Black would be unable to oppose that, e.g. 34... a5

35 iWc3 a4 36 <j;Jh2 a3 37 .id2 gb2 38 .icl±. Now Black's knights become very active, ensuring an edge. 30 ... fxg4 31 hxg4 .ixg4 32 ~xg4 tDf5 33 .ie2

33 ... tDf6! 34 ~4g2 tDxe3 35 'lWxe3 tDxd5 36 'lWd2 'lWf5 37 .ig4 'lWf7 38 c4 tDxf4 39 ~f2 e3 40 'lWxe3 ~af8 41 ~gf1 'lWb7+ 42 c.!?g1 ~xg5 43 ~xf4 ~fg8 44 c.!?h2 ~xg4 45 ~xg4 ~xg4 46 ~f8+ c.!?g7 47 ~g8+ c.!?xg8 48 'lWe6+ c.!?f8 49 'lWxg4 'lWb2+ 50 c.!?h3 'lWc3+ 51 c.!?h2 'lWe5+ 52 c.!?h3 c.!?e7 53 'lWh4+ c.!?e6 54 'lWh6+ 'lWf6 55 'lWe3+ c.!?d7 56 'lWb3 c.!?c7 57 a4 'lWf1 + 58 c.!?h4 'lWe1 + 59 c.!?g5 'lWg1 + 60 c.!?f6 'lWg6+ 61 c.!?e7 'lWg7+ 62 c.!?e8 'lWg8+ 63 c.!?e7 'lWb8 64 'lWd3 'lWd8+ 65 c.!?f7 'lWd7+ 66 c.!?f8 'lWxa4 67 'lWxh7+ 'lWd7 68 'lWd3 a5 69 'lWb3 a4 70 'lWc3 c.!?b6 71 'lWd3 c.!?a6 72 'lWc2 a3 73 'lWa2 'lWc8+ 0-1

123

Part 7

1 d4 &LJf6 2 c4 d6 3 &LJc3 QUICK REPERTOIRE

In this part I examine positions where Black refrains from g6. That allows him to complete quickly development, but restricts his counterplay.

8 d5!

1. 3 ....if5 4 g3!

It is not easy to achieve e4 without significant concessions which would justify Black's play, e.g. 4 f3 e5 5 e4 exd4 6 ~xd4 .ie6. 4 ... e5 (4 ... tiJe4 5 ~d3) 5 tiJf3 tiJbd7 6 .ig2 c6 7 0-0 h6 Practice has seen Black struggling after 7....ie7 8 tiJh4 .ig4 9 h3 exd4 10 hxg4 dxc3 11 g5 tiJh5 12 tiJf5±, Christiansen-Shirazi, New York 1989.

124

White has more space and he is ahead in development. The game Uhlmann-Gheorghiu, Sofia 1967 went on 8 ... cxd5 9 tiJh4 .ie4 10 tiJxd5 hg2 11 @xg2 tiJxd5 12 ~xd5 ~c7 13 E1d1 tiJb6 14 ~b5+ ~c6+ 15 ~xc6+ bxc6 16 b3 E1d8 17 .ib2 tiJd7 18 f4t with lasting pressure. 2. 3 ... e5 4 dxe5 4 tiJf3 is also in White's favour, but why not punish the opponent by dooming him to a difficult defence with only two possible results! 4 ... dxe5 5 ~xd8+ Wxd8 6 tiJf3 tiJfd';i' 8 g4!

1 d4 ttJf6 2 c4 d6 3 -tc3

A typically human move. It is connected with a clear plan of opening a file and fixing targets on the kingside. 7... c6 8 b3 (preparing to meet 8 ...!b4 by 9 !b2) 8. ..f6 9 g5 as 10 h4 ttJc5 11 h5 ttJ ba6

Now simplest would have been 12 !e3 mc7 13 h6 !f5 14 !h3±, starting to collect Black's kingside pawns.

Strong players like Malaniuk proved that without a well premeditated plan White was gradually losing his initiative and that gave the better player fair chances to win. The icing on the cake was the trifling learning overhead - Black follows his scheme, develops pieces and castles no matter White might play. This most attractive feature is, however, Black's biggest problem. Once White got used to this setup, it turned out that Black has little to oppose to best play from the opponent. White learned to seize space in the centre by d5 and immediately attack on the queenside with b4 anda4: 3. 7... 0-0 (7... a6!? 8 ttJh4! g6 9 !h6) 8 !e3 a6 9 d5! cxd5 10 cxd5 b511 ttJd2

About 25 years ago Black achieved some success with the following passive setup: 3 ••. c6 4 e4 e5 5 ~f3 ~bd7 6 .ie2 Ae7 7 0-0 His play was not a revelation, of course, but as simple as it looked, it was not easy to pinpoint the drawbacks of Black's strategy. 125

Part 7 White aims to play b4, followed up by a4. This plan is based on the zwischenzug 11 ... lLlb6 12 a4 b4 13 a5!±. Instead, 12 ...bxa4 13 lLlxa4 lLlxa4 14 E1xa4 i.d7 15 E1a3 is hardly an improvement since the a3-rook will go to the c-file with a total domination of White. In fact, Black can impede a4, but look at what cost: 11..:~c7 12 E1c1 Wfb8 13 b4 lLlb6 14lLlb3 i.d715lLla5 E1c816 Wfd2 i.d8 17f3

The gaping hole on c6 totally binds Black down. At the end, let us examine an even more passive setup: 4. 7... 0-0 8 i.e3 E1e8 9 d5 c5

White must open the b-file: 10 E1b1 lLlf8 11 b4 b6 12 bxc5 bxc5 13 lLld2 lLlg6 14 g3;!;. Play through game 19 Sakaev-Vukic, Saint Vincent 2005 for clarifying the further plans of both sides.

126

Part 7

1 d4 tiJf6 2 c4 d6 3 tiJc3

STEP BY STEP

Sometimes Black players use this move order to sidestep variations with g3 without CLlc3. 3 CLlf3 is outside the scope of our repertoire. Anyway, it has no serious advantages over 3 CLlc3. After the text main branches, beside 3 ... g6, are: A. 3 ... .ifS B. 3 ... eS C.3 ... c6

A.3 ....if5

4g3! I have tried many times 4 f3 eS

S e4 exd4 6 Wfxd4 .ie6 7 b3, only to convince myself that Black obtains sufficient counterplay on the dark squares. Another known variation is 4 e4 he4 S CLlxe4 CLlxe4 6 Wff3 dS=. 4 ... e5 4 ... CLle4Ioses the battle for e4 together with the bishop pair after S Wfd3 dS 6 cxdS CLlxc3 7 WfxfS CLlxdS 8 a3 CLlc6 9 CLlf3 g6 10 Wfd3 .ig711.ig2 CLlb6 12 0-0 0-0 13 ~dl±, KhenkinScalcione, Saint Vincent 2002. 5 tiJf3 tiJbd7 Or S... CLlc6 6 .ig2t CLle4?! 7 0-0 CLlxc3 8 bxc3 .ie4 9 Wfa4±. 6 .ig2 c6 7 0-0 7 CLlh4 exd4 8 Wfxd4t is also good. Note that 8 CLlxfS?! dxc3 9 b3 Wfc7 10 0-0 g6 11 CLle3 .ig7 12 .ia3 CLlcS 13 b4 CLle614 ~c1 CLld71S bS cS+ turned well for Black in D. Gurevich - Gheorghiu, New York 1986. 7 ... h6 Practice has seen Black struggling after 7....ie7 8 CLlh4: 8 ....ig4 9 h3 exd4 10 hxg4 dxc3 11 gS CLlhS 12 CLlfS± Christiansen127

Part 7 Shirazi, New York 1989 or 8 ... ~g6 9 dSl"1c81O dxc6 bxc611 b4ltJb612 ~b3. White has some pressure. The game Koneru-Krasenkow, Wijk aan Zee 2008 saw further 12 ... dS when best would have been 13 l"1d1! ltJxc4 14 bS.

8 d5! White was slightly better following 8ltJh4 ~h7 9 e4 ~e71O ~e3 0-0 11 ltJfS hfS 12 exfS exd4 13 hd4 ~aS 14 ~c2;!;, Ligterink-Miles, London 1981, but the text is better as it gains space in the centre. The game Uhlmann-Gheorghiu, Sofia 1967 went on 8 ... cxdS 9 ltJh4 ~e4 10 ltJxdS hg2 11 'tt>xg2 ltJxdS 12 ~xdS ~c7 13 l"1d1ltJb6 14 ~bS+ ~c6+ 1S ~xc6+ bxc6 16 b3 l"1d8 17 ~b2 ltJd7 18 f4;!; with lasting pressure.

B. 3 ... e5 4 dxe5

We saw in the previous part that White obtains a lasting initiative by trading queens in this pawn structure. White also often plays 4 ltJf3. Then 4 ... exd4 SltJxd4 ~e7 6 e4 0-0 7 ~e2 l"1e8 is rather passive, while 4 ... e4 S ltJd2 ~fS 6 g3 c6 7 ~g2 dS 8 0-0 128

~e7 (Or 8 ...hS 9 cxdS cxdS10 f3

exf3 l1ltJxf3±, Khismatullin-M.Nikolov, Oropesa del Mar 2000) 9 cxdS cxdS 10 f3 destroys the black centre: 10 ... exf3 l1ltJxf3 ltJc6 12 ~b3;!;. 4 ... dxe5 5 VMxd8+ @xd8 6 lDf3 lDfd7

The eS-pawn needs protection so Black has to play sooner or later .. .f6. He can try another setup though: 6 ... ltJbd7 7 l"1g1! c6 (7... h6 8 g4 e4 9 gS exf3 10 gxf6 ltJxf6 11 exf3 c6 12 ~e3 ~e6 13 0-0-0+ 'tt>c7 14 ~f4+ 'tt>c8 1S ltJe4;!;, RowsonM.Ivanov, Verona 2006.) 8 g4 ltJe8 9 ~e3 f6 10 0-0-0 'tt>c7 11 gS ltJd6 12 gxf6 gxf6 13 b3 ltJfS 14 ~d2 as 1S ltJe4;!; G.Grigorov-Rusev, Sunny Beach 2004. Play is similar to the main line.

894 ! The computer does not include this move between his first 4 choices, but human practice is strongly in its favour. The advance of the gpawn marks the beginning of a clear plan - to push gS, eventually open the g-file. Then White's rook will be constantly threatening to invade g7, while the split pawn formation will be easier to attack, especially the f6-

1 d4 ltJf6 2 c4 d6 3 ltJc3 pawn (by liJe4). 7 g3 f6 8 ig2 c6 9 0-0 rJlc710 ie3 as 11l"1adU is comfortable for playing, but White lacks targets here. 7 ... c68 b3

This is the modern treatment. White avoids doubled pawns in the case of 8 ie3 ib4 9 0-0-0 ixc3 10 bxc3 rJle7. However, I thrust the intuition of possibly the finest positional player ever, Smyslov, who was clearly better after 11 gs liJa6 12 ih3liJb613 ixc8l"1hxc814liJxeS cs 15 l"1hg1!±, Smyslov-Lombardy, Monte Carlo 1969. 8 ...f6 8 ... ib4 is well met by 9 ib2 as 10 0-0-0 f6 11 gs liJa6 12 gxf6 gxf6 13 liJe4 rJle7 14 l"1gl±, Rowson-Sanchez Guirado, Palma de Mallorca 2008.

C.3 ... c6

Black commonly prefers this move order in order to sidestep White's plans with long castle as in the line 3 ... liJbd7 4 igs es sliJf3 ie7 6 e3 0-0 7 ~c2 c6 8 0-0-0. It is arguable that the bishop is better placed on gs though. For instance, Black can use it to gain tempi for his development by 4 ... h6 5 ih4 gs 6 ig3 liJhs. 4 e4 e5 5 liJf3 liJbd7 6 .ie2 .ie7 7 0-0

9 g5 as 10 h4 liJc5 11 h5 liJba6 7 ... 0-0

We have been following the game Epishin-Llorente Zaro, Benasque 2007. White has realised his plan and he needs only a couple of precise moves to fix his advantage. Now simplest would have been 12 ie3 rJlc7 13 h6 if5 14 ih3±, starting to collect Black's kingside pawns.

7... a6!? is seldom seen, but it hides some venom. This move allows Black to escape from the main line since White's bishop is not on e3 yet and his standard answer to ... a6 - 8 ds, does not work. (8 ds cxds 9 cxds bs 10 b4 liJb6 11 ie3 id7 and a4 is impossible) White can underline the drawback of delaying castling only by 8 liJh4! g6 9 ih6. Then 9 ... ~b6 would face White with a choice: He can either win a pawn, risking to find it impossible to convert it after 10 dxes liJxeS 11 ig7 l"1g8 12 i.xf6 i.xf6 13 ~xd6 ~d8 14 ~xd8+ ixd8 15 liJf3 liJxf3+ 16 ixf3 ie6, or sacrifice one with 10 liJf3! 129

Part 7 'Wxb211 'Wd3!? (Khalifman suggests 11 ttJa4 'Wa3 12 c5 exd4 13 .ic1 'Wb4 14 a3 'Wa5 15 cxd6 hd6 16 'Wxd4~) 11 ... 'Wb6 12 flab1 'Wa5 13 flfd1 exd4 14 ttJxd4 ttJe5 15 'Wd2t. In any case White has ample compensation. 7... 'Wc7 S .ie3 ttJfS is an ambitious plan to concentrate all minor pieces on the kingside and organise an attack. Our common antidote is to gain space in the centre by 9 d5

Now Black cannot prevent us from opening files by 9 ... c5 due to 10 :gb1 ttJg6 11 b4 b6 12 bxc5 bxc5 13 'Wa4+ .id714 'Wa6 0-0 15 flb7.icS16 flxc7 ha6 17 flbU with a pleasant initiative, so he continues his main idea: 9 ... ttJg6 10 ttJel h5 11 ttJd3 ttJg4 12 .id2 f5. Here we should open the centre by 13 dxc6 (or 13 c5!? dxc5 14 dxc6t - Khalifman) 13 ... bxc6 14 'Wa4. Black is hardly happy to have his pawn on h5. For instance, 14 ... fxe4 15 ttJxe4 0-0 16 h3 ttJf6 17 ttJxf6+ hf6 IS .ia5± and he has two pawns hanging.

Si.e3 S fle1 and S 'Wc2 are at least as popular as the text, but these moves are linked with maintaining tension in the centre. I prefer a clear plan 130

which is easy to follow. White completes development and gains space by d5. Then he attacks the queenside using the power of the bishop pair on e2-e3. This plan has proved to be the most effective in practice. S ... a6 Alternatively: a) S ... ttJg4 9 .id2 does not change considerably the situation. For instance:

9 ... 'Wb6 Or 9 ... a6 10 d5; 9 ... ttJgf6 10 d5 a5 11 'Wc2 ttJeS 12 a3 g6 13 b4 f5 14 dxc6 bxc6 15 exf5 gxf5 16 b5±, Banas-Slovineanu, Panormo 200l. 10h3 Black's threat on b2 could be also neutralised with 10 a3 exd4 (or 10 ... ttJgf6 11 c5 'Wc7 12 cxd6 hd6 13 flc1 exd4 14 ttJb5 'WbS 15 ttJxd6 'Wxd6 16 .ib4 c5 17 e5±) 11 ttJa4 'Wc7 12 ttJxd4t. This variation has been proposed by Khalifman. The text opens the centre and changes the pawn structure in White's favour: 1O ... ttJgf611 c5 'Wc712 cxd6 hd6 13 .ig5 exd4 Black cannot wait any longer as 13 ... fleS 14 d5 would leave his eS"rook without prospects: 14 ... h6 15

1 d4 tiJf6 2 c4 d6 3 -tc3 i.h4 i.b4 16 dxc6 bxc6 17 tiJd2 i.e7 IS Elcl. White has a pleasant game due to his better pawn formation, e.g.: IS ... tiJc5 19 b4 tiJe6 20 b5:t or IS ... ElbS 19 tiJc4 tiJc5 20 i.g3 i.fS 21 'lMfc2:t. 14'1Mfxd4 EleS 15 Eladl i.c516'1Mfd2 tiJxe4 17 tiJxe4 Elxe4 IS i.d3 EleS 19 Elfel±. It turns out that White is not obliged to defend the central pawn. b) S... EleS 9 d5 c5

This is a very dull and passive setup for Black who will be staying cramped for ever. In a tournament such a strategy does not seem attractive, but in blitz the active side can run out of time while trying to figure out how to break through. White must open the b-file and he can start with either 10 Elbl or 10 tiJel. I prefer the rook move only because it keeps open the option for the f3-knight to go to d2. Khalifman gives another reason - to avoid 10 tiJel tiJb6, which is irrelevant in my opinion. In that case the knight on b6 is obviously out of place and White obtains a comfortable edge by simple natural means: 11 Elbl h6 12 b4 cxb413 Elxb4 tiJbd714 tiJd3 b615 'lMfa4 tiJc5 16 tiJxc5:t, D.Kosic-Vukic, Neum 200S. Let us return to 10 Elbl tiJfS 11 b4 b6 12 bxc5 bxc5 13 tiJd2

tiJg6 14 g3:t. This is yet another typical position. See game 19 SakaevVukic, Saint Vincent 2005. 9d5! Now that Black has weakened his wing, we can close the centre. 9 ... cxd5 10 cxd5 b5 1Q... tiJg4 11 i.d2 b5 gives White an extra option - 12 tiJel tiJgf6 13 tiJc2:t, as noted by Khalifman. However, I do not see any reason to renounce our main plan with b4, followed up by a4: 12 b4! tiJb6 13 a4 bxa4 14 tiJxa4 f5 15 exf5 hf5 16 tiJc3. The a6-pawn is doomed, for example, 16 ... 'lMfcS (Or 16 ... mhS 17 Ela5 i.f6 IS h3 tiJxf2 19 Elxf2 e4, Gavrikov-Hickl, Switzerland 2006, 20 tiJh2! i.d4 21 i.el 'lMff6 22 'lMfb3 hh3 23 tiJdl±) 17'1Mfb3 mhS IS Ela5 i.dS (IS ... e4 19 tiJd4 i.f6 20 tiJe6 EleS, Ionov-Sturua, Ohrid 2001, 21 Elcl+-) 19 tiJg5! tiJd7 20 Ela3 hg5 21 hg5'1MfeS 22'1Mfc4'1Mfh5 23 h3±. 11 ttJd2

In this position Black has tried unsuccessfully many moves which proves that his game is rather bad. He cannot effectively prevent White's two major plans: 131

Part 7 1. To fix the b5-pawn by b4 and then attack it by a4 or: 2. To instal a knight on c6 via the route d2-b3-a5-c6.

C1. 11 ... tt)xe4 12 tt)cxe4 f5 13 a4 bxa4 14 \!!Vxa4 fxe4 15 tt)xe4 tt)f6 16 tt)xf6+ i.xf6 17 \!!Vc6 ~b8 18 i.xa6~

This is a variant of the previous line. The pawn on b4 becomes a target: 14 \!!Vb3 14 f3 fxe4 15 fxe4 :B:xf1 + 16 ~xf1 is also somewhat better for White: 16 ... i.g5 17 bg5 ~xg5 18 ~cl± or 16 ... a5 17 :B:el i.g5 18 bg5 ~xg5 19 :B:c6 i.b7 20 ~c1 ~xel+ 21 :B:xel :B:c8 22 :B:xc8+ bc8 23 i.g4;l;. 14.. .fxe4 15 \!!Vxb4 ~b8 16 \!!Vxe4 tt)f6 17 \!!Vc4 i.d7 18 tt)c3 ~xb2 19 \!!Vxa6±.

C3.11 ...i.b712 b4 tt)b613 a4±.

Play is similar to the next lines.

18 ...\!!Ve7 18 ... e4 does not give counterplay du to 19 bc8 :B:xc8 20 ~a4 ~e7 21 i.d4±. 19 b4±. White is clearly better, FaragoB.Stein, Dortmund 1986.

C4. 11 ... tt)e8 12 b4! i.g5 13 i.xg5 \!!Vxg5 14 a4 bxa4

C2. 11 ... b4 12 tt)a4 tt)xe4 13 tt)xe4 f5 15 tt)c4 White can also keep more pieces on the board by 15 ttJf3 ~e716 ~xa4 f5 (16 ... :B:b8 17 i.d3 ttJb6 18 ~c2 ttJc7 19 :B:fcl±) 17 exf5 :B:xf518 ttJd2 :B:f419 g3 :B:f8 20 ttJde4;l;. 15 ... ~b8 16 \!!Vxa4 tt)b6 17 tt)xb6 ~xb6 18 ~fc1 f5 19 tt)d1 ~b8 20 exf5 i.xf5 21 tt)e3~ i.g6 22 ~c4 ~f4~

132

1 d4lLlf6 2 c4 d6 3 -':..c3 This occured in the game StohlA.Jones, Moscow 1994. Now simplest would have been 23 g3±, threatening to win the rook by h4.

C5.11 ... 'lWc7121'!c1 'lWb813 b4 ltJb6 14 ltJb3 .id7 15 ltJa5 1'!c8 16 'lWd2 .id8 17 f3

In this line Black prevented a4, but let our knight to a5. Now 17... lLlc4 would lose a pawn to 18 hc4 bxc4 19 lLldl .ib5 20 1'!f2±, so in the game Antic-Aronian, Yerevan 2000 Black followed up with: 17 ... h6 18 ltJb1 1'!a7 19 1'!xc8 'lWxc8 Here 20 lLlc6± would have been clearly better for White.

C6. 11...ltJb6 12 a4 bxa4 Our whole strategy is based on the variation 12 ...b4 13 a5!±. As a result, Black is unable to keep the queenside closed and the week a6pawn is a constant source of concern. 13 ltJxa4 ltJxa4 14 1'!xa4 .id7 Or 14 ... a5 15 b4 .id716 b5±. 15 1'!a3

15 ... a5 After 15 ....ib5 we can restrict Black's play on the kings ide by 16 f3 lLlh517 g3± when 17....ig5 would fail to 18 f4+-. 16 'lWc2 a4 17 ltJc4 1'!b8 18 1'!c1 The a3-rook will also go to the c-file with a total domination of White.

133

Part 7

1 d4 c!l)f6 2 c4 d6 3 c!l)c3 QUICK REPERTOIRE

19. Sakaev-Vukic Saint Vincent 20.09.2005 1 d4 c!l)f6 2 c4 d6 3 c!l)f3 c!l)bd7 4 c!l)c3 c6 5 e4 e5 6 .ie2 ie7 7 0-0 0-0 8 ie3 ~e8 9 d5 c5

10 ~b1

I do not understand grandmasters like Vikic who voluntarily accept to defend such positions. Three years after the current game he became victim of yet another model realisation of White's queenside plan with b4: Kosic-Vukic Neum 10.06.2008 10 ttJe1 ttJb611 :1'!b1 h612 b4 cxb4 13 :1'!xb4 ttJbd7 14 ttJd3 b6 15 ~a4' 134

ttJc5 16 ttJxc5 bxc5 17 :1'!b3 .id7 18 .ic8 19 ~a3 a6 20 :1'!tb1 :1'!a7 21 .id2± ~a6

Black has managed to trade one minor piece in comparison to the commented game, but he is no less cramped: 21. .. ~c7 22 :1'!b6 .ib7 23 .id1 :1'!ea8 24 .ia4 ttJd7 25 :1'!6b3 <;t>h7 26 ~b2 :1'!b8 27 ttJd1 .ig5 28 .ixg5 hxg5 29 .ixd7 ~xd7 30 ttJe3 g6 31 h3 <;t>g7 32 :1'!b6 f6 33 ttJg4 and White eventually won on the kingside. My conclusion is that the whole structure is bad for Black. 10 ... c!l)f811 b4 b612 bxc5 bxc5 13 c!l)d2 c!l)g6 14 g3 ih3 15 ~e1 Wd7 16 f3 ~eb8

1 d4 lLIf6 2 c4 d6 3 '::'c3 22 .id3 lLIe7 23 f4 f6 24 f5 ~c8 25 1:!b1 ~c7 26 .id2 ~a6 27 a3 ~c7 28 cj;>g2 a6 29 cj;>f3 ~e8 30 g4 cj;>f7 31 h3 h6 32 cj;>e2

17~a4

Sakaev is conducting this game as if he wanted to present me with a classical example for my book. White is not attacking anything until the very last moment, when he is already obviously winning. There is nothing to calculate, every decision is based on long-term considerations. Computers prefer here 17 i.f1 Ml18 lLIxf1, assuming that the white light-squared bishop is "bad". Sakaev has another opinion on that subject. White has spatial advantage, so he should keep as many minor pieces as possible. Conversely, it would be good to trade queens, since this piece defends the weak d6- and a7-pawns. Most inexperienced players are afraid to swap queens fearing that it would be difficult to win later. Sakaev however knows exactly what he is doing: he exchanges one pair of rooks to avoid counterplay along the b-file, and expands on the kingside to gain even more space. While Black will be trying to cover all the invasion squares on the b-file, he would probably miss a breakthrough on the other wing. 17 ...~xa418lL1xa4 i.d719lL1c3 .id8 20 1:!xb8 1:!xb8 21 lLIb3 lLIe8

32 ... ~b6? An obvious concession as the white knight finally gets the chance to reach the decisive outpost on c6. Let us consider the more precise defence 32 ... lLIc7. Then the computer suggests 33 ~a5, but it would be a serious mistake. As cramped as Black looks, without rooks he has all the chances to survive. Instead, White should aim to open another file and to use his more active rook. He can make progress by the breakthrough g4-g5, or by sacrificing a piece on cS. He can try two interesting setups. 1) 33 ~d1lL1e8 34 ~c2lL1e7 35 h4 lLIc8 36 i.e2 lLIc7 37 E:hl!

135

Part 7 Now 37... ltJa8 would stumble into 38 g5! fxg5 39 hxg5 hg5 40 hg5 hxg5 41 l'!h8 'it>f6 42 ltJa5± (42 i.h5 g4) 42 ...g6 43ltJc6 l'!b7 44 l'!g8 gxf5 45 i.h5+-, but 37... 'it>g8! saves the day as 38 ltJa5 ltJa8 39 ltJc6 does not work in view of 39 ... hc6 40 dxc6 ltJa7. White can attempt a piece sac on c5: 38 l'!gl 'it>f8 39 i.e3 ltJe8 40 ltJxc5 dxc5 41 hc5+ i.e7 42 he7 + ltJxe7 43 c5 l'!c8 44 c6 ltJxc6 45 dxc6 hc6=. We see that somehow Black is holding here, mainly because White's king is not well placed on the c-file in the last variation. So, we should consider: 2) 33 i.c2ltJe7 34 'it>d3ltJc8 35 h4 ltJe8 36 l'!hl 'it>g8 37 i.e3 ltJc7, when 38 ltJxc5 dxc5 39 hc5 'it>f7 40 i.e3;!; is already quite appealing. Black's defence is not trivial in these lines. He must be constant-

136

lyon alert and a tiny mistake could turn to be decisive. After his mistake, he is lost: 33 ltJa5! fi.e7 34 ltJe6 ~e8 35 a4 f8 36 a5 ltJa8 37 ltJa4

Total domination! 37 ...fi.xe6 38 dxe6 e7 39 ~b7 gb8 40 tiJe3 d8 41 tiJd5 e8 42 tiJe7+ d8 43 tiJd5 e8 44 .ie2 .id8 45 .ia4 .ie7 46 f3 .id8 47 g3 1-0

137

Index of Variations Modern Benoni Pawn Structures Part 1 1 d4 tLlf6 2 c4 g6 3 tLlc3 il.g7 4 e4 d6 5 f4 c5 6 d5 0-0 7 tLlf3 e6 8 il.e2 exd5 9 cxd5 (9 ... a6, 9 ... tLla6) .......................................................... 16 9 ... b5?! (9 ... a6, 9 ... tLla6) ............................................................................ .17 10 e5 dxe5 11 fxe5 tLlg4 12 il.g5 ~b6 13 0-0 c4+ 14 ~hl tLld7 ........ 17 14... tLlxe5 ...... 18 9 ... tLlbd7 .................................................................................................... 18 9 ...il.g4 10 0-0 tLlbd7 ................................................................................. 20 11 h3!? hf3 12 hf3 !%e8 ........................................ 21 12 ... tLle8 ........................................ 25 13 g4!? ............................ 26

13 il.e3 ............................. 29 12 ... a6 13 g4 .................................. 31 13 ... tLle8 ............... 31 13 ... h6 .................. 33 12 ... !%b8 ......................................... 34 12 ... c4 ............................................ 34

Part 2 7 tLlf3 e6 8 il.e2 exd5 9 cxd5!%e8 10 e5 dxe5 (10 ... tLlfd7) ............ 51 11 fxe5 tLlg412 e6! (12 il.g5) ....................................................................... 53 12 ... fxe613 d6 (13i1.g5) ...................................................... 56 13 ... ~b6 ......................................................... 57 14 tLlg5 tLlh6 15 0-0 tLlc6 .................. 59 15 ... tLlf5 .................. 60 13 .... tLle5 ........................................................ 62 13 ... tLlc614 0-0 !%f8 ....................................... 63 14... tLlf6 ........................................ 63 14 ... tLld4 ....................................... 64 13 ...il.d7 (14 il.g5, 14 tLlg5) .............................. 66 14 0-0 ~b6 ...................................... 67 14 ... il.c6 ....................................... 67 King's Indian Defence and Volga Pawn Structures Part 3 5 f4 c5 6 d5 0-0 7 tLlf3 Rare Lines without 7 ... e6 ........................ 74 7... a6 .......................................................................................................... 74 7... b5?! ....................................................................................................... 76 8 cxb5 a6 9 a4 e6 (9 ... ~a5?!, 9 ... il.b7) .......................................... 76 9 ... axb5 ...................................................................... 77

138

Part 4 S f4 0-0 6lLlf3 ............................................................................. 8S 6 ... eS?!, 6 ... a6 ............................................................. 8S 6 ... lLlbd7, 6 ... lLlc6 ....................................................... 86 6 ....ig4 ....................................................................... 87 6 ... lLla6 7 .id3 .............................................................88 7 ... eS 8 fxeS dxeS 9 dS ............................. 89 9 ... cS ....................... 89 9 ... lLlcS .................... 89 9 ... c6 ....................... 90 7 ...ig4 ..................................................... 91

Modern Defence and Other Rare Move Orders Part 5 1 d4 g6 2 e4 .ig7 3 c4 (3 lLlc3) ..................................................... 99 3 ... d6 (3 ... lLlc6, 3 ... c6, 3 ... cS) .................................................................. 100 4lLlc3lLlc6 (4... eS) S dS (S .ie3) ................................................... 102 S... lLld4 6.ie3 c5 (6 ... eS) ......................... 104 7 lLlge2 lLlxe2 ................. lOS 7... V9b6 ................... 106 Part 6 1 d4 g6 2 e4 .ig7 3 c4 d6 4lLlc3lLld7 SlLlf3 ............................... 116 S...eS (S ...e6) 6.ie2 lLlgf6 (6 ...lLle7) 7.ie3 0-0 8 0-0 ............................... 116 8 ... exd4 .................... 117 8 ... aS ........................ 118 8 ...h6 ........................ 118 8 ... lLlg4 ..................... 118 8 .. J:1e8 ...................... 119 8 ... c6 ....................... 120

Part 7 1 d4 lLlf6 2 c4 d6 3 lLlc3 Lines without ...g6 ............................. 127 3 ....ifS ...................................................................................................... 127 3 ... eS ........................................................................................................ 128 3 ... c6 ........................................................................................................ 129 4 e4 eS SlLlf3lLlbd7 6 .ie2 ie7 7 0-0 0-0 8 ie3 a6 9 dS cxdS 10 cxdS bSlllLld2 .................................................................................... 131 11... lLlxe4 ......................................................................... 132 11...b4 .............................................................................. 132 11 ...ib7 ........................................................................... 132 11... lLle8 ........................................................................... 132 11. ..V9c7 ........................................................................... 133

139

Other CHESS STARS Books Repertoire books: Opening for White According to Kramnik l.Nf3 by A. Khalifman Volume 1a: Old Indian, rare lines in the Classical Variation, 2006 Volume 1b: The Classical Variation, 2006 Volume 2: Anti-Nimzo-Indian, Anti-QID, English, Knight Tango, 2008 Volume 3-5: second editions coming in 2009-2010 Opening for White According to Anand l.e4 by A. Khalifman Volume1: Petroff, Ruy Lopez without 3 ... a6 Volume2: Ruy Lopez with 3 ... a6 Volume3: Caro-Kann, 1...c6, 2 ... g6 Volume4: 1...d6, 1...g6 and others VolumeS: Alekhine's Defence, 1...b6 and other rare lines Volume 6: The French Defence 3.Nc3 dxe4, 3 ... Nf6, 2006 Volume 7: The French Defence 3.Nc3 Bb4, 2006 Volume 8: The Sicilian, Paulsen-Kan and rare lines, 2006 Volume 9: The Sicilian, Paulsen-Taimanov and other lines, 2007 Volume 10: The Sicilian, Sveshnikov, 2007 Volume 11: The Dragon, 2009 Opening for Black According to Karpov by Khalifman

Current theory and practice series: Challenging the Sicilian with 2.a3! by Bezgodov, 2004 An Expert's Guide to the 7.Bc4 Gruenfeld by Sakaev, 2006 The Sharpest Sicilian by Kiril Georgiev and At. Kolev, 2007 The Safest Sicilian by Delchev and Semkov, 2nd rev.ed. 2008 The Queen's Gambit Accepted by Sakaev and Semkov, 3rd. rev. ed., 2008 The Easiest Sicilian by Kolev and Nedev, 2008 The Petrosian System Against the QID by Beliavsky and Mikhalchishin, 2008 Games collections My One Hundred Best Games by Alexey Dreev, 2007 Bogoljubow. The Fate of a Chess Player by S. Soloviov, 2004 Super Tournaments 2003, 456 pages + colour photos Super Tournaments 2002, 556 pages + colour photos Shirov's 100 Wins by Soloviov 316 p., interviews, biography, colour photos, 2003 Leko's 100 Wins by Soloviov 340 pages, biography, colour photos, 2003 More details at www.chess-stars.~om 140