Winning With the Spanish

WINNINC WITH· THE SPANISH ~NATOLY

KARPOV

Winning With the Spanish Anatoly Karpov Translated by John Sugden

B. T. Batsford Ltd, London

First published 1994 © Anatoly Karpov 1994 ISBN 07134 64712 British Library Cataloguing-in-Publication Data. A CIP catalogue record for this book is available from the British Library.

All rights reserved. No part of this book may be reproduced, by any means, without prior permission of the publisher.

Typeset by John Nuon GM and printed in Great Britain by Redwood Books, Trowbridge, Wilts for the publishers, B. T. Batsford Ltd, 4 Fitzhardinge Street, London W1H OAH

A BATSFORD CHESS BOOK Adviser: R. D. Keene GM, OBE Technical Editor: Graham Burgess

Contents Preface 1 Closed Spanish: Zaitsev Variation Game 1 Kasparov-Karpov 215, Moscow 1985 Game 2 Timman-Karpov, Kuala Lumpur Ct (7) 1990 Game 3 Kasparov-Karpov 219, Moscow 1985 Game 4 Kasparov-Karpov 3114, LondonlLeningrad 1986 Game 5 Kasparov-Karpov 3116, LondonlLeningrad 1986 Game 6 Kasparov-Karpov 5/4, New YorkILyons 1990 Game 7 Kasparov-Karpov 5/22, New YorkILyons 1990 Game 8 Kasparov-Karpov, Amsterdam 1990 Game 9 Timman-Karpov, Kuala Lumpur Ct (9) 1990 Game 10 Khalifman-Karpov, Reggio Emilia 1991192 Game 11 Hjartarson-Karpov, Seattle Ct (5) 1989 Game 12 Timman-Karpov, Kuala LumpurCt (5) 1990 Game 13 Hjartarson-Karpov, Seattle Ct (3) 1989

72

2 Closed Spanish: 9 ...~7 Game 14 Kasparov-Karpov 5112, New YorkILyon 1990 Game 15 Kasparov-Karpov 5/6, New YorkILyon 1990

80 87

3 Open Spanish Game 16 Short-Beliavsky, Barcelona 1989 Game 17 Van der Wiel-Hjartarson, Rotterdam 1989 Game 18 Speelman-Timman, London Ct (4) 1990 Game 19 Dolmatov-Yusupov, Wijk aan Zee Ct (3) 1991

5

7 13 18 22 29 37 42 45 52 57 61 68

96 106 112

121

4 Marshall Attack Game 20 Short-Pinter, Rotterdam 1988 Game 21 Beliavsky-Malaniuk, Minsk 1987 Game 22 Andrijevic-Pavlovic, Yugoslavia 1988

125 137 143

5 Bird, Schliemann and Berlin Defences Game 23 Kamsky-Ivanchuk, Tilburg 1990 Game 24 Timman-Korchnoi, Brussels Ct (3) 1991 Game 25 Short-Gelfand, Brussels Ct (5) 1991 Index of Variations

151 159 166 175

Preface This book is the third in a series of publications which I have devoted to individual openings. Those written previously were Beating the GrUnfeld and Winning with the Petroff, to continue the series, here is Winning with the Spanish. Like its two predecessors, the book is based on the 'round figure' of 25 - this being the number of 'main' or 'primary' games which it incorporates. However, the overall number of examples, taking every reference into account, must be at least ten times that figure. Nearly all the material (with the exception of a few encounters with Kasparov from our earlier matches) dates from the late 1980s or the beginning of the nineties, which is to say that it is highly topical. It derives, moreover, from contests which lack nothing in calibre: world title matches, Candidates matches and top-level grandmaster tournaments. I believe that the book reflects all the most interesting and fruitful controversies about the Spanish Game in the last few years. Two thirds of the primary games in the book were played by the author. I hope the reader will not object to this; after all, the Spanish was thoroughly examined in a number of World Championship matches in which I took part - first in Baguio and Merano, and later in almost every match with Kasparov. It follows that the author's own games are indispensable for an account of recent advances in the Spanish. The fact is that the Spanish Game has featured in my opening repertoire throughout my entire chess career. Of course, if this book had been written just after the matches with Korchnoi, I should have been constantly referring to myself as playing the White side. But times change, and in my marathon contests with Kasparov, I repeatedly adopted this opening with Black - a fact which is, naturally, reflected in the book. Incidentally, for convenience, games from the Kasparov matches are mostly designated (as in other publications) by a pair of numbers; the first is the number of the match, the second is the gamenumber within the match. Which variations and systems in the Spanish Game occupy the forefront here? In the first place, there is the variation which is now named after my long-standing second, Igor Zaitsev. This is the line with 10...:e8, in which Black temporarily dispenses with the prophylactic

6 Preface ... h7-h6. This variation, like the system with 9 ...lLld7, regularly occurred in my duels with Kasparov. As to other branches of the Spanish, I must mention the Open Variation and the Marshall Attack, which never cease to attract interest and are used quite frequently by many prominent grandmasters. Turning to less popular lines, one game each is devoted to Bird's Defence (3 ... lLld4), the Schliemann (3 .. .f5) and the Berlin Defence (3 ...lLlf6 40-0 lLlxe4). As in my four-volume work The Open Game (Semi-Open Game / Closed Openings / Semi-Closed Openings) in Action, each of the 'primary' games basically amounts to an extensive essay (or, if you like, a lecture) on a currently fashionable branch of the opening, embracing all the latest theoretical and practical developments. Unfortunately, many variations of the Spanish have had to be left out; but then, virtually every one of them has had a book (in some cases more than one book) written specially about it, and to embrace all the Spanish trends within a single publication would be quite out of the question. The reader may be astonished that the Chigorin System, which headed the popularity table for decades, is not to be found here. But precisely because this system has had a vast quantity of literature devoted to it, I decided to 'give it a rest', all the more since the interest in it has somewhat declined in recent years. Of the more rarely seen systems, I have excluded for example those in which the queen goes to e2. We recall that the variation 1 e4 e5 2lLlf3 lLlc6 3 .tb5 a6 4 .ta4 lLlf6 5 0-0 .te7 6 'ii'e2 was employed twice by Nigel Short in the 1992 Candidates Semi-Final against the author. Both games ended in a win for Short, and may even have been decisive in turning the match in his favour. But it is clear that the queen move to e2 fails to refute Black's play, and these losses are to be attributed not to my choice of opening but to my poor performance in the match as a whole. Some of the World Championship games given here will doubtless already be familiar to the reader from The Open Game in Action. I have repeated them in the present volume for the sake of completeness, but in so doing I have supplemented them with new material to illustrate how the debates begun by Kasparov and myself in the title matches have been carried on subsequently. Finally I must express my gratitude to Evgeny Gik, chess master and writer, for his help in assembling the material for this book. Anatoly Karpov

1 Closed Spanish: Zaitsev Variation Game No. 1 Kasparov-Karpov World Championship Match game 215 Moscow 1985 Kasparov and I are both great adherents of the Spanish Game, and it occurred in all our matches except for the one in Seville. Although both of us readily play this opening for either colour, it happens that in our World Championship games with it, Kasparov always had the white pieces. The Spanish was played twice in each of the first three matches, and no less than eight times in the last match! In the first contest, in 1984, the opening made its appearance at the very end - in games 44 and 46. Although both games ended peacefully, I was not satisfied with the way they went, and prepared myself better for our next contest. The present game turned out to be one of the most successful of the second match. In the notes, I shall insert some explanations of the opening refinements in games 1144 and 1146. 1 e4 eS 2 lOfJ lOc6 3 .*.b5 a6 4 .*.a4 lDf6 5 0-0 .*.e7 6 :el b5 7 .*.b3 d6 8 c3 0-0 9 h3 .*.b7 9...lOaS, the characteristic move of the Chigorin System, one of the most time-honoured variations in chess theory, is seen much more rarely now than it used to be - either because it is simply not in fashion or because clear-cut methods of obtaining the initiative have been discovered for White. The Breyer System with 9...lOb8 is not in vogue nowadays either. The 9...lOd7line will feature in Games 14 and 15 of this book. As for the Smyslov System - 9...h6 10 d4 %le8 11 lObd2

8 Closed Spanish: Zaitsev Variation .tfB 12lt)f1, with 12....tb7 or 12 ....td7 to follow - it has been completely superseded by the immediate 9 ....tb7 which aims to save a tempo, since in some cases Black can do without ...h7-h6. 10 d4 :e8(J)

The 9 ... .tb71ine was introduced into practice by grandmasters Flohr and Lilienthal as long ago as the 1940s. Their analysis focused on the position arising after 10 d4 ed 11 cd d5 12 e5 ~4. In our own day, it is thanks to the move 10...:e8 that the line has risen again in popularity. The rook move was introduced by Igor Zaitsev, my second for many years; hence the variation now bears his name. Black is playing a kind of improved Smyslov System. On move 10, some other continuations have been seen - 1O.....d7, 10... ~, 10... ed and 1O... h6 - but they are less reliable, although there are of course possibilities of transposition. Today the Zaitsev Variation figures in the repertoires of numerous grandmasters, and I myself employ it regularly. Black fortifies the centre without wasting time on the prophylactic ... h7-h6. True, there is the danger that White may repeat moves with 11lt)g5:f8 12lt)f3 :e8 13 It)g5. It is rare for grandmaster games to end like this, and yet I have been the victim of this very drawing manoeuvre on about ten occasions. It follows that if a win is essential, Black must choose something different. It goes without saying that from the theoretical standpoint, the possibility of forcing a draw cannot count as an achievement for White. On the other hand, the repetition 11lt)g5 :f8 12lt)f3 is sometimes carried out merely in preparation for time-trouble. In this case, on 12...:e8, White switches to the 'normal' 13lt)bd2. I should add that in several games quoted in this book, including some of the primary ones, these extra moves actually occurred (in one case they even occurred twice); but I have always excluded them from the game scores, so that

Closed Spanish: Zaitsev Variation 9 in comparing variations and positions the reader will not be confused by the move-numbering. Incidentally, after 11 lDg5 :f8, White has nothing better than the knight retreat 12lDf3. It is well known, for instance, that he achieves nothing with 12f4efI3j,xf4~aSI4j,c200! 15edj,xg5161VhS h6 17 .tg3 g6 18 -.f3lDc4 19 ~ ~b6 20 j,b3 hS 21l&2 -.f6 22 as 23 a3 lDc4, when Black has the advantage (Ljubojevi~-Gli gori~, match 1979).

-.e4

11

~bd2

In game 1144 Kasparov played the immediate 11 a4, obtaining a substantial plus after 11...h6 12lDbd2 ed 13 cd ~b4 14 -.e2 j,f8 IS eS j,c6. Black's last move is an error; the annotators suggested IS ... de 16 lDxeS ~fd5 !1. But I think that the pawn exchange in the centre was also premature. A more precise continuation is 12....tf8. This position arose in the 9th game of the present (198S) match, twice in the return match (1986), and four times in the 1990 match. I shall therefore have a good deal to say about it later. 11 ... .tfS

12 a4 Game No. 13 is devoted to the more modest 12 a3. Note that at this point the standard manoeuvre 12 M is unavailable to White on account of 12...ed 13 cd lDaS, when the threat to win the e-pawn forces the exchange of the important light-squared bishop. This is one of the assets of9 ....tb7. 12 ... 'iVd7(2) This comparatively rare queen move led to victory in the present game, but afterwards I constantly opted for the traditional 12...h6, giving a position mentioned in the notes to White's 11th move.

10 Closed Spanish: Zaitsev Variation The move played transposes to a form of the Smyslov System which is normally reached by ...•d7 on move nine.

13

ab

In Hiibner-Portisch, Brussels 1987, White advanced his d-pawn at once, before the queenside exchanges: 13 dS ~7 14 c4lLlg6 (l4 ...h6 and 14...g6 have also been seen) IS .tc2, obtaining no more than the usual Spanish initiative. The retreats 12.ta2 and 13 .tc2 are also well known, but the exchange on bS is regarded as strongest.

13 14

...

ab

haS .tnS If now IS/OgS, Black has IS ...:e7 with ...h7-h6 to follow, whereas after 14...:xaS IS/OgS the black knight would be forced to retreat to the back rank, and White would seize the centre: Is ...1Od8 161Odf3 ed (l6...h6 17 lLlxf7! lLlxf7 18 de) 17 cS!. Ivanchuk-Portisch, Linares 1990, varied with 16...cS 17 de de 18 .xd7 /Oxd7 19/Oxm c420 lLlxd8 :xd8 21 .ta2 :e8 22 .te3, and White acquired a large plus. 15 cIS (3)

The first time we had reached this position was in game 1146, in which I retreated with IS ...1Od8 and White obtained a substantial advantage after 16lLln h6 17lLl3h2! lLlb7 18 .tc21Oc5 19 b41Oa6 20 lLlg41Oh7 21/Og3 c6 22 dc .txc6 23 .tb31Oc7 24 .f3lLle6 2S h4 .d8 26 :dl. At this point 26 ...•xh4? is bad in view of 27lLlfS .d8 28 lLlxeS IOhgS (28 ...lLlegS 29 .txt7+!) 29 .g4 .txe4 30 lLlxf7!, and White comes out on top. I played instead 26 ...•a8, and after 27 .tdS?! .txdS 28 ed fiX7 291Oe4 .c8! I gradually extricated myself. However, White could have gained a decisive plus with 27 .txh6! gh 28 hd6! .txd6 (or 28....tg7 2910fs .txe4 30 lLlfxh6+ .txh6 31lLlxh6+ Wg7 32 .xt7+ Wxh6 33 .txe611fb8 34 .tfS+ .xd6 3S .xh7 mate; Editor's

Closed Spanish: Zaitsev Variation 11 note: Speelman and Tisdall gave 2s ... llli4! as far from clear) 29lilxh6+ ~hS

30 lilxf7+ ~g7 31lilxd6 :e7 32 "ii'g4+ etc. There are other ways of withdrawing the black knight. One of them, 15 ...lilaS, will be examined in the present game. A further alternative is 15 ... lile7; I twned to this a few years later (see Game No.2). In Kasparov-Smejkal, Dubai OL 19S6, Black played 15 ...lilbS, and things did not go especially well for him: 16lllillila6 17 .i.g5! .i.e7 ISlilg3 g6 19 "ii'd2 .i.b7 20 :al :as 21 .i.c2 c6 22 dc .i.xc6 23 :dl :dS 24 "ii'e3 "ii'b7 25 .i.h6lLlc7 26lilfS, and White had a big advantage. 15 ••• lLla5 Winning a tempo in comparison with other moves - with the 15 ...lilds of game 1146, at any rate. 16.i.a2 c6 17 b4 lilb7 (4) The manoeuvre 17... lLlc4 had already been investigated by grandmasters analysing game 1146. In that case White has 18 lilxc4 be 19 .i.g5! (19 .i.xc4 cd 20 ed ':cS 21 "ii'b3 "ii'c7 22lild2 g6! gives a level game; the black bishop aims to jump out to h6) 19...cd 20 J.xf6 de 21 lilxe5 "ii'a7 22 lilg4 "ii'xa2 23 lilh6+ ~hS 24 "ii'h5 (24 .i.d4! is also good) 24 ... gf 25 "ii'xf7 .i.xh6 26 "ii'xeS+ ~g7 27 "ii'e7+ with the advantage, since the opponent's bishops have no prospects.

After the knight has gone to b7 it looks as if Black is virtually suffocating for lack of living space. But this impression is deceptive. Nimzowitsch once said that a piece kept confined for a long time can acquire incredible strength when it breaks free. And indeed, in the present game, my queen's bishop, hidden on as and walled in for the time being by my own knight on.b7 and pawn on c6, will soon break its bonds, develop vigorous activity and become a key factor in Black's victory.

12 Closed Spanish: Zaitsev Variation

18 c4 The standard IS lOfl is evidently more reliable; after lS ...cd (or lS ...c5 19 J.g5 J.e7 20 ll)g3 g6 21 .d2, preparing M-h2 and f2-f4) 19 ed h6 20 ll)g3 (20 ll)3h2ll)dS 21ll)g4ll)h7, aiming for counterplay with ...f7-f5) 20...:cS 21 .d3ltldS 22ll)h2! g6 23 h4!, the initiative is with White. At move 19, there is also danger for Black in 19...:cS 20 J.g5! ll)es 21 .d3 g6 22ll)g3 h6 23 J.d2 f5 24ll)h4! (analysis by Gutman). 18

•••

:c8!

It makes sense to occupy a more promising file with the rook before withdrawing the knight to dS. 19 de SUITendering the centre allows Black to obtain a fully equal game at once. Some fascinating play would result from 19 .e2 (alternatives are 19 J.b2 and 19 c5) 19...ltldS! 20 J.b2 be 21ll)xc4 .a7! 22 :al cd 23 ed J.xd5 24ll)xd6 J.xd6 (or 24 ... .txf3 25ll)xcS .xa2 26 gf, with unclear play) 25 J.xdS .xa1+ 26 .txal :cl+ 27 ~h2 (27ll)el J.xb4) 27 ...:xal, and Black has adequate compensation for the slight material deficit. He also obtains good chances with 21...:bS!' 19 ... .xc6 20 eS? Rather recklessly played; 20 .tb2 at once was better. The threat against f7 (20...dc 21 J.xf7+) is easily parried, and Black seizes the initiative. 20 ~ 21 J.b2 de! 22 be Black also has good chances after 22ll)xe5 .a6! 23 .al c4 24 J.c3 .a3 25 :e3 J.xb4 26ll)exc4 be 27 J.xf6 c3! 2S :g3 ~6, or 22 J.xe5 ll)d7 23 J.b2 c4. 22 .xeS 23 J.xe5 ll)d7 24 J.b2 'ifb4! (5) More precise than 24 ...•c2 25 .al .a4 26ll)d4ll)c5 27 :e3, with sharp play. 25 ll)b3? Mter 25 9bl White's position would still have been defensible. His wish to play actively lands him in trouble. 2S ll)c5! 26 J.al

Closed Spanish: Zaitsev Variation 13

.a3

A forced sacrifice of a pawn. After 26 ~xc5 .xb2! 27 :e2 ~d3 .i.xe4 29 .i.xt7+ ~xfT 30 :xe4 b4, Black has a clear plus.

28

26 ... .txe4 27 ~d4 If 27 ~g5, Black wins with the striking 27 ....i.c2! 28 .xc2 .xel+ 29 ~h2 .td6+ 30 g3 lOe4. 27 ~b7 28 .e2 ~6 29 ~eS .xeS

30 31 32

33

.g4

:eS

:dl .f4

.i.g6

.eI .as .as :e3 .eI

Wb4!

.i.e4

34 :el 3S .tb3 b4 36 Wb2 .tg6 37 .xe8 38 :XeS 39 ~ ~ 40 .tdS 41 ~b3 ~ The last move was sealed. In the adjourned position Black's passed pawn guarantees the win. Kasparov decided he had no saving chances, and resigned without resuming.

Game No. 2 Timman-Karpov Candidates Final. 7th game Kuala Lumpur 1990


1 e4 eS 2 00 lOc6 3 J.b5 a6 4 J.a4 ~ 5 0-0 J.e7 6 :el b5 7 J.b3 d6 8 c3 0-0 9 h3 J.b7 10 d4 11 a4 'iVd712 o!Llbdl J.f8 13 ab ab 14 ]haS J.n8 15 dS 15 o!Lle7 (6)

:e8

6 W

In Game No.1 we examined IS ... o!Lla5, but on e7 the knight is not so badly placed either. 16 o!Lln After 16 c4 :b8 16 :e3 c6! 18 dc o!Llxc6 19 cb o!Lld4 20 J.c4 o!LlxbS 21 b3 'iVa7 22 J.b2 h6 23 :el J.c6 Black has a good game (Popovic-Smejkal, Zagreb 1985). 16 ••. h6 The bishop sortie to gS cannot be allowed. 17 o!Llg3 A new move. The game A.Sokolov-Karpov, Linares 1989, went 17 o!Ll3h2, and after 17 ... cS (17 ... c6 is also possible) 18 dc o!Llxc6 19 o!Llg4 o!Llxg4 20 hg b4 21 gS hg 22 J.xgS o!LlaS 23 J.a4 J.c6 24 J.xc6 o!Llxc6 2S o!Lle3 be 26 bc J.e7 27 J.xe7 a draw was agreed. 17 c6 18 de J.xc6 19 o!Llh2 If 19 o!Llh4?!, then 19 ...dS is good. 19 ... d5 19 ... g6?! is too passive: 20 'iV£3 J.g7 21 o!Llg4. 20 %5! o!Llxe4 Black is more or less forced to accept the pawn sacrifice; after 20... o!LlxhS 21 'iVxhS g6 22 'iV£3! (not 22 'iVxeS? o!LlfS! with equality, but 22 'iVg4!? is also dangerous for Black) 22 ...hS (22... J.g7 23 o!Llg4) 23 J.gS, White has the advantage. Perhaps at move 22 Black has to go in


for the risky 22 ...fS!?

21 ~g4! Apart from the simple 22 ~xeS, White threatens 22 ':'xe4 and ~f6+. 21 ••• (7) The queen excursion to f5 may well be Black's only defence. On 2l....c7, White easily mounts a decisive attack: 22 ':'xe4 de 23 ~xh6+ gh 24 ~f6+ ~h8 (or 24 ...~g7 251iVhs ~g6 26 .txh6+ ~xf6 27 .gS mate) 2s1iVhs ~g8 26 ~xg8 ~xg8 27 .g6+ ~h8 28 .txfl.tg7 29 .txh6 ':'g8 (or 29 ...•xfl30 .xf7 .txh6 36 .f6+) 30 .txg7+ ':'xg7 31 1iVh6+ :b7 32 mate.

.rs

.£8

7 W

Diagram 7 is a critical position in this game and perhaps in the match as a whole. Timman now played 22 0, and I managed to get myself out of trouble with 22 ...~xc3!' However, the annotators considered that by continuing 22 ':'xe4!?, White could virtually have won by force. True. this move would have been very dangerous for me to meet; nevertheless Black could have held out. Let us look at it in detail; the starting point is the position after 22...de 23 ~g3 .c8 (23 ...•d7 24 .txfl+ ~h7 25 .xd7 .txd7 26 .txe8 .txe8 27 ~xe4 is bad for Black) 24 ~xe5 (8).

Black has two defences: 24 ...':'d8 and 24 ...~h7. After 24 ...':'d8 2s1iVh5 .tdS 26 .txdS ':'xdS 27 .xf7+ ~h7 28 .tf4 ':'d6 29 ~g4, many commentators ended their analysis, judging 30 .txh6 to be an irresistible threat. However, at move 28, Zaitsev unearthed some quite interesting variations starting with 28 ...b4!. Play can continue 29 cb JIbS. and it is not simple for White to develop his initiative: 30 lDc4 (he gains nothing from 30 ~e4 .fS; or 30 ~ .fS 31 .e8?! ':'xeS 32 .txeS ~g6!; or 30 ~g4 ~g8, and now 31 .teS ':'b7! or 31 ~xe4 ':'fS) 30... ~g6, with quite sharp play. 30... ~g8 is also


8 B

possible. On the other hand, 30...:xb4 31tai6 "d8 32 tDit5 :xb2 (or 32... lilg8 33 "f5+ g6 34 "f7+ ~h8 35 "xg6, and wins) leads to a time-honoured mate with 33lilf6+ ~h8 34 "g8+ lLlxg8 35 M. It is also unlikely that Black can hold out with 31.....a8 (instead of 31.....d8); White has 32 W, 32lile8, 32lilgxe4 or 32 .*.xh6. From diagram 8, a more subtle defence is 24 ...~h7! 25 .*.xf7 (if 25 lilxf7, the reply 25 ...lLlf5 obliges White to force a draw at once with 26 lilxf5 "xfS 27lilg5+ ~h8 28lilf7+ ~h7 29lilg5+ etc.; while 25 'iVh5 fails to 25 ... g6) 25 ...:d8, and now: (a) 26 "c2 g6!, and the straightforward 27 lilxe4.*.xe4 28 "xe4 l%dl+ 29 ~h2 :xcl (not 29.....fS? 30 .*.xg6+!) 30 lilxg6 "fS 31 lilxf8+ ~g7 32 "xf5 (32 "xe7? "f4+) 32...lilxfS 33 .*.e6leads to complete equality. If instead 27 .*.f4, then after 27 ....*.g7 28 lilxg6 (28 lilxe4 .*.xe5 29 .*.xe5 .*.xe4 30 "xe4 "fS is also harmless) 28 ...lilxg6 29lilxe4 .*.xe4 30 "xe4 "a6 31 h4 :dl+ 32 ~h2 "f6 Black is already on top. (b) 26 'iVh5 :d6 27 .*.f4 (27 .*.g5? g6!) 27 ...g6 (27 ...lilg8? 28 "g6+! forcing mate) 28 "e2 l%f6 29 .i.e3 .*.g7 (29 ... lilf5? 30.*.xg6+ :xg6 31lilxg6 ~xg6 32 "g4+ ~6 33 .i.d4+) 30 .*.d4 (9). The situation remains fairly tense, but Black is hardly risking defeat. 22 f3 This allows Black to rescue himself at once. In addition to 22 l%xe4, it was worth considering 22lilg3!? lilxg3 23 :xe5, or 22 "O!? 22 ••• lilxc3! The only move, but adequate. Black does badly with 22.....xh5 23 fe'iVh4 24lilxe5 "g3 25 :e3 "xe5 26 ed "d6 (or 26 ...'ifb8 27 dc lilxc6 28 .*.xf7+! ~xf7 29 "d5+ and wins) 27 dc "xc6 28 :f3, and White has an obvious plus. 23 be "xh5


9 B

24 lheS White also gains little from 24 tDxe5 :d8 25 .i.a3. 'iVb4 24 ••• 2S .i.e3 An equal game results from 25 .d4lOg6 26 :xe8 .i.xe8 27 .i.d2. 2S lOg6 26.i.tl 1Ms 27 lheS .xeS 28 .i.xd5 hS! 29 lOe3 lOr4! 30 .i.xc6 .xc6 31 .i.g3 (10) 10 B

White is already compelled to defend accurately; 31 .d2 is met by 32....xc3, with advantage to Black. 31 ••• lOe6 32 lOds Or 32...lOd4!? 33 .i.f2 (33 .xd4? is met by 33 ....i.c5, 33 cd?! by

.c4

18 Closed Spanish: Zaitsev Variation 33 .....xd5, and 33 i.e5 by 33 .....xd5 34 i.xd4 i.c5) 33 .....xd5 34 "xd4 "xd4 35 i.xd4 f6 36 g4! hg 37 fg q;f7 38 q;t2 q;e6 39 q;e3 q;d5 40 h4, with a draw. 33 "d2 ""3 34 q;h2 ""1 35 "e1 "xel 36 i.xe1 i.d6+ 37 i.g3 i.cS 38 i.e1 ~ 39 g4 hg q;e8 40 fg q;d7 41 q;g2 q;c6 42 i.g3 43 ~b4+ i.xb4 l/z_l/z

After his failure to win this game Timman completely went to pieces, and played the next two games way below par. I succeeded in winning both of them, and thus the way was opened to a further duel with Kasparov! Game No. 3 Kasparov-Karpov World Championship Match game 219 Moscow 1985 1 e4 eS 2 ~f3 ~ 3 i.b5 a6 4 i.a4 ~f6 5 0-0 i.e7 6 .:tel b5 7 i.b3 d6 8 c3 0-0 9 h3 i.b7 10 d4 .:te8 11 ~bd2 i.f8 12 a4 12 ... h6 So I decided to vary from the fifth game of the match (No.1 in this volume), in which I had played 12.....d7. It is interesting that shortly before this match, in a training match against Timman (Hilversum 1985), Kasparov had twice played this same opening line (up to move 11) with Black. In both games the Dutch grandmaster refrained from a2-a4, preferring the quieter a2-a3 with the idea ofb2-b4, i.bl-b2 and c3-c4. Although the result was one win and one loss, Kasparov obtained an opening advantage in both these games. 13 i.c2 Another possibility is 13 d5, but the bishop retreat looks more solid. 13 ••• ~b8 (11)


11 W

As already mentioned, the exchange on d4 and the excursion ...lDc6b4 occurred in both Spanish games of the return match (1986) and in four games of our last match (1990), so the detailed material on that line will come later. The transfer of the knight from c6 via b8 to d7 is characteristic of the Breyer System. True, in that system Black plays an immediate 9 ...~b8 and usually dispenses with ...h7-h6, but perhaps the difference is of minor importance. 14 i.d3 c6 15 ~bd7 16 ~g3 'fIe7 In Sax-Rivas, Rome 1984, Black played 16...g6 at once, and after 17 .i.d2~7! 18 'fIcl h519'f1dl i.g7 20.i.c2~b621 b3 ba22 baa5 23 .i.d3 i.a6 24 i.xa6 :xa6 25 de de the position was completely level. A more energetic line is 17 h4 h5 (otherwise h4-h5 is unpleasant), and in view of the weakness of g5, White's game is to be preferred. 17.i.d2 g6 18 'fIc1 ~h7 19 b3 (12) If 19 'fIc2 (in the game White postpones this manoeuvre by one move), then 19...c5! 20 abc4100ks quite a good reply. But here again it was worth considering 19 h4, forcing 19...h5. Counterplay in the centre is inadequate: 19... c5 20 ab c4 21 b6! (21 ba cd 22 ab lIxal 23 'fIxal 'fIxb7 gives Black good play for the pawn) 2l...~xb6 22 i.c2 with advantage. So I am prepared to admit that Black's opening experiment in the present game is none too effective. However, in the final phase of the game, Kasparov and I succeed in creating something rather like a study, and it is for that reason that I have decided to include this among the 'primary' games. 19 i.g7

m


12 B

20 'ii'c2 Preventing ... d6-dS, for example: 20 ...dS 21 t[)xe5 t[)xe5 22 de t[)xe4 23 t[)xe4 de 241be4 c5 25 :g4 'ii'xe5 26 J..xg6+! fg 27 'ii'xg6+ ~g8 28 :el and Black is in a bad way. 20 21 J..e3 22 :adl :ac8 23 J..n J..f8

24

:d2

White's basic plan involves advancing his b-pawn and then his cpawn too, exerting pressure in the centre. He could also have carried it out immediately, with 24 b4 'ii'b6 25 'ii'a2 J..g7 26 de de 27 c4.

'ifb8

24 2S

'iVbl

30

:001

J..aS 26 b4 J..b7 27 ab Remarkably, the first exchange of the game occurs only on move 27. ab 27 28 :edl 'ii'c7 29:c1 J..g7 Here too White could have played 30 de de 31 c4. But again he postpones the advance of his c-pawn, and I manage to consolidate. 30 :OOS 31 de de 32:XdS :xdS

33 34 3S

:XdS

~dS

c4 J..xc4

be t[)eS

Closed Spanish: Zaitsev Variation 21 Black could equalise more simply with 35 ....i.c8 36 "al ~8 37 i.e6.

"as

36"82

ltid6

37 .i.b3 c!£Jb5 If now 37...i.c8, then 38"aS is unpleasant. 38 b4 This manoeuvre looked more dangerous twenty moves earlier. Now Black has sufficient counterplay. ltid4 38 39 .i.xd4 eel 40 b5 "e7 41 White wastes an important tempo, and suddenly the play becomes sharp. The right move was 41 "c2, retaining the initiative after 41.. ..i.a6 42.i.c4 .i.xc4 43 "xc4. 41 ... cS 42"c2 cb The sealed move. The adjourned position looks more pleasant for Black, but Kasparov succeeds in finding a study-like draw. 43 bg+ 43 gh 44 c!£Jxd4 would give Black the advantage. 43 Ig 44"c4 b5 (13)

"d2

"c4

"e5

13 W

At this point, after 45 c!£Jxd4 .i.xe4 46 c!£Jxe4 "xe4 47 ~6 "el+ 48

..n "d2, Black would retain somewhat the better chances. But Kasparov has prepared a surprise, and I shall now have to attend to my own safety. 45 eS!

22 Closed Spanish: Zaitsev Variation By opening the bl-h7 diagonal, White sets up the unpleasant threat of 46 ~xh5! gh 471i'g8+ ~h6 48 i.c2. H 45 ...h4, the knight sacrifice is still possible: 46 ~h5! gh 471i'g8+ ~h6 48 i.g2 with unavoidable mate, or 46...i.xf3 47 1i'g8+ ~h6 48 ~xg71i'xg7 (48 ... i.g4 49 f4 'ifxg7 50 'ifxd8) 49 'ifxd8, with a large endgame advantage for White. However, in his notes to the game Kasparov has shown how Black could have saved himself: 46 ...i.xe5! 47 ~xe5 1i'xe5 48 1i'g8+ ~h6 49 f4 1i'el+ 50 ~h2 gh! (50... ~xh5 51 i.dl+! 1i'xdI521i'xd8) 511i'g5+ ~h7 52 'ifxh5+ ~g7, with a draw. 45 i.xf3 46 gf i.xeS 47 f4! i.:xf4! After 47 ...i.g7 48 f5! the black king is in trouble. ~h6 48 1i'g8+ 49 i.el 'ifg7! Black loses with 49 ...1i'f6 50 ~e4, or 49 ... d3 50 i.xd3 'iff6 51 ~ .al+ 52 ~g2 i.c7 (52...~6 531i'f8+) 53 ~3! 1i'a8+ 54 ~51i'c6 55 .f8+ ~h7 56 i.e4, and it is all over. 50 1i'xd8 i.xg3 51 fg 1i'eS 52 1i'f8+ ~gS 53 ~g2 The storm that has swept across the board has abated, and we agreed a draw; 53 ...1i'e2+ 54 ~h31i'g4+ (54 ...1i'xc2?? 551i'f4 mate) 55 ~g2 1i'e2+ gives perpetual check.

Game No. 4 ~parov-~ov

World Championship Match game 3114 London/Leningrad 1986

:e1

1 e4 eS 2 ~f3 ~c6 3 i.b5 a6 4 i.a4 ~f6 5 0-0 i.e7 6 bS 7 i.b3 d6 8 c3 0-0 9 h3 .tb7 10 d4 11 ~bd2 i.f8 12 a4 h6 13 i.el 13 ... ed Although, from the point of view of the opening, Black had nothing to complain about in the two Spanish Games of the 1986 return match, my losses in these games had a strong psychological effect on me, and after the end of the duel with Kasparov I temporarily switched to another well-known move, 13 ...:b8 (14).

:e8


14 W

I shall now digress a little, to recall two games I played on these lines. Timman-Karpov, Tilburg 1986, went 14 ab (14 i.d3 i.c8 15 lllil i.d7 16lOg3 Wc8 17 i.n1i'b7 18 Wc8 19 1Oh2 liJe7 20 f4 ef21 i.xf4lOg6led to equality in Beliavsky-Gligoric, Sochi 1986) 14...ab 15 i.d3 i.c8 16 M (16lOb3 and 16 dS have also been seen) 16...b4 (an innovation; after 16...ed 17 cd lOb4 18 i.bl c5 19 i.f4, the advantage is with White) 17lOg3 be 18 be ed 19 cd lOb4 20 i.bl c5 21 i.f4 l:.b5 22 Wd2 l:.aS 23 l:.xaS Wxa5 24 dS Wd8 25 l:.dl i.d7. The game is about equal. As often happens, White worsened his own position in seeking the initiative, but finally there was a peaceful result. Hjartarson-Karpov, Dubai OL 1986, varied with 16...i.d7 (another novelty, in place of 16...ed or 16... b4) 17lOg3 Wc8! 18 i.e3 (better 18 J.d21i'b7 19 b4! l:.aS 201i'b3 l:.xal 21 l:.xal :a8 22 l:.xaS "xaS 23 1i'bl 24 Wa2! with initiative to White, Geller-Gligoric, Sochi 1986) 18 ...1i'b7 19 de (after 19 dS, the game is level) 19...lOxe5 20:a7 Wc8 21lOxe5 de 22 l:.e6! 23 i.c2 c5. Black's chances are better, and White had to struggle for 50 moves to save himself. The pawn exchange in the centre combined with the knight sortie to b4 is currently the most popular system in the Zaitsev Variation. 14 cd 14 lOxd4 has also been played, but with that configuration in the centre White can scarcely count on a plus. 14 ••• lOb4 15 i.bl c5 (15) The capture on a4 also occurs frequently. In particular, it was played in game 2 of our 1990 match. I shall discuss this later. The move 15 ...c5 underwent thorough tests not only in the present match (1986) but also in our fifth one (1990). Another possibility is

as

"e8

"f3


15 W

15... g6, though Black's results with it have been decidedly modest. 16 cIS Considered virtually obligatory. After 16 b3 cd 17 tDxd4 ba 18 :Xa4 as 19 J.b2 g6 20 J.c3 :c8 21 J.xb4 ab 22:Xb4 J.aS 23 :a4 J.g7 24 J.d3 "'6 2S J.a6 :cd8, Black seized the initiative in TseshkovskyBalashov, Minsk 1982. The exchange on c5 similarly gives White nothing; after 16 dc dc 17 e5 clOd7 (17 ... ~7 and 17 ... are also playable) 18 ab ab 19 :xaS J.xa8 20 e6 :Xe6 21 :xe6 fe 22 ~ 'fIc7 the chances were equal in the game Kasparov-Balashov, Kislovodsk 1982. 16 ... clOd7 Grandmaster Dorfman, one of Kasparov's trainers in the 1986 match, regularly plays 16...g6. But in a game against Aseev (Lvov 1984), after 17 ~1 J.g7 18:a3 ba 19 :xa4 as 20:a3 J.a6 21lDg3 J.b5 22 J.f4 ~7 23 'fId2, he had White's dangerous activity to contend with. 17 l:a3 (16) An ingenious manoeuvre, typical of this opening set-up, which permits a quick transfer of the rook to the central files. Instead, 17 lDn allows Black to undermine the centre favourably with 17 .. .f5! 18 eS J.xdS 19J.xf5 J.xf3! 2O'fIxf3lDxe5 21 'fIdl c422 :e3 'fIf6 23lDg3 dS. The central pawn wedge is immensely strong, and Black won quickly in Dvoirys-Kruppa, 54th USSR Ch Semi-Final, 1986. After 18 ef lDf6! Black also has excellent chances, for example 19 J.d2lDbxdS 20 lDg3 'fId7 21 ~ b4! 22 lDh2lDxe4 23 J.xe4lDf6, de Firmian-Beliavsky, Tunis 1985. 17 ... c4 Subsequently the attention of theorists was wholly transferred to the counter-stroke ...f7-fS, which was tested with particular thoroughness

m


in our fifth match (New YorkJLyons, 1990). The reader will find the relevant games further on.

18

ab

Now after the reply 18...ab, White will post his knight on d4. In the 16th game of the match, he played 18 llXi4 at once. The difference might appear slight, yet we shall see that the play radically diverged in the two cases. I would point out, incidentally, that the natural exchange on b5 was played in the present game for the first time. The 14th and 16th match games gave rise to a lively theoretical debate. Large numbers of new games were played with this variation, and the assessments constantly changed. The most important material on this theme is included in the present book.

18 19

•.• lild4

ab

lba3

19... ~ is interesting. After 20 lOn l:xa3 21 ba lOd3 22 i.xd3 'ifxd4 23 i.e2 'ifxdl 24 l:xdl o!Oc5 the initiative is with Black, but a stronger line is 20 lOf5 (White can also play 20 102f3 o!Oc5 21 i.e3, and the pawn on e4 is immune: 2l...o!ilxe4 22 l:xa8 i.xa8 23 i.d2!) 20...g6 21 lOn <;Ph7 22 'ifd2! l:xa3 23 ba llXi3 24 i.xd3 cd 25 o!ilxh6 i.xd5 26 'iff4 (26 'ifxd3! is more precise), with some advantage to White; Efimov-Foigelson, USSR 1988. In analysing this game afterwards, the players decided that in reply to 20 lOf5 an immediate exchange on a3 is dangerous for Black, but on the following move it is playable, for instance: 20... g6 21 lOn l:xa3 22 ba lOxd5 23 ed l:xel 24 'ifxel gf 25 'ife8 lOf6 26 'ifb8 llXi7 27 'ife8 o!Of6, with a draw. 20 ba o!Od3 21 .i.xd3 cd (17) The diagram position should be considered as the point of departure for the following investigations.


If now 221Oxb5, then after 22 ....i.a6, 22 .....,,6 or 22 ...1faS 23lM4

1Dc5, Black has fully adequate compensation for the pawn. 22 .i.b2 Arnason-Shvidler, Beersheva 1987, went 22 l:e3, and after 22...1ff6 23 .i.b21Dc5 24 .i.al1fd8 251fcl1fc8 26 ""1! 1faS 27lDf51faS 28 1fb4 1fxb4 29 ab g6 30 lDxd6, White obtained a decisive plus. The second game of the Candidates Match between Sax and Short (Saint John, 1988) is interesting: 22 l:e3 .!i)eS! (instead of Shvidler's 22...1ff6; but 22...1Dc5! also deserves attention) 23lDxb51faS 24lM4 1fc3 25102b3 .i.a6 26 .i.d2 ""2 27 .i.b4 g6 28 f4 (too risky; better 28 1fd2 offering an immediate queen exchange, although after 28 ...lDc4 29 'iVxb2 .!i)xb2 30 lM2 Black would retain sufficient compensation for the pawn) 28 ... ~ 29 l:xd3 l:xe4 30 1ff3 l:e8 31 ~h2 .i.g7 32 lOc6 'iVe2 33 lObd4 1fxf3 34 J:xf3 l:e4 35 lOb3 l:e2 361Ocd4 l:e4 37 1Dc6 :e2 38 ~g3? (in extreme time-trouble Sax declines to repeat moves and commits a fatal error) 38...lOe3 39 h4 l:xg2+ 40 ~h3 .i.c8+ 41 f5 .i.xf5+ 42 l:xf5 gf 43 .i.xd6 l:b2 44lDe7+ ~h7 4s1Dc5 .i.f6 0-1. The following are two attempts to strengthen White's play at move 23. Hiibner-Short, Belfort 1988: 23 ""3 .i.a6 24lD4f31fc7 25 .i.b2 b4 26 ab lOxf3+ 27 .!i)xf3 1fc4 28 1fxc4 .i.xc4 29 .i.c3 g6, with full compensation for the pawn; the game ended in a draw. Ivanchuk-Kruppa, Frunze 1988: 23lD4f3! 1Dc4 (not 23 ...lOxf3+ 24 J:xf3!) 24 :xd3 1fd7 (24 ...g6 is more solid) 25lOxc4! be 26 J:e3 f5 27 00 :c8 28 .i.b2 ""5 20 1fal ! with a won position for White. A more accurate reply to 23lD4f3 is 23 .. .f5 24 .!i)xe5 l:xe5 25 .i.b2 J:e7 (25 ...:e8? 26 ""3 fe 27lDxe4 d2 28 1fd3! with the threat of 29 lOf6+) 26""3 fe! 27 l:xe4 l:xe4 281Oxe41fa8 291fxd3 .i.xdS 30 lOg3 .i.xg2! 311fg6.i.f3 32lDf5 'iVe4 331Oxh6+ ~h8 3400+ 112-112;

Closed Spanish: Zaitsev Variation 27 Glek-A.Kuzmin, Blagoveshchensk 1988. In Glek-Kharlamov, corr. 1989-90, White in tum found an improvement: 26 ef! :xe3 27 fe -*.xd5 28 "g4 "c8 29 "d4.tc4 30 lbxc4 be 31 f6 "c7 32 a4 gf 33 -*.c3! -*.e7 35 a5, with a powerful initiative. 22 ••• Numerous commentators consider this move best, but it is also worth considering 22 ...llX5, after which White cannot develop an attack with 23 ~ or 23 "g4. On 23lLlxb5, Black has the choice between 23 ...-*.a6 with initiative for the pawn, or a forcing line which virtually guarantees a draw: 23 ...'iVb6 24 a4 .txd5 25 ed llxel+ 26 "xellLlxa4 27 -*.xg7 "xb5 (27 .. .<~xg7 28 "al+) 28 -*.xfS xfS 29 "e3 "xd5 30 "xh6+ g8. It appears that White should play 23 :e3, with the threat of 24lLl4b3. 23 Surrounding the errant pawn at once does not work: 23 :e3lLle5 24 f4lLlc4; or 23 lLl2f3lLlc5 24lLlb3lLlxb3 25 "xb3 d2 26 :e2 "a4! 27 "xa4 ba 28lLlxd2 -*.xd5; or 23lLl2b3 "a4 24 "xd3 .txd5; in all variations Black obtains an advantage. Utilising the absence of the enemy queen, White endeavours to organise an attack on the king. The threat now is 24 -*.xg7 .txg7 25 "g4. 23 tDes (18)

"as

rn!

In making this move with the knight, I underestimated the following exchange on e5. Evidently 23 ... g6 was sounder; Kasparov considers 24 lLlb3 the main reply, and gives this sequence of moves as best for both sides: 24 .....a4 25 "xd3 lLle5! (25 ... gf 26 "g3+ h7 27 "f3) 26 -*.xe5 (26 "g3 .txd5!) 26...llxe5 27 f4 :e8 28lLlg3, with unclear play. Many annotators recommended 24 :e3 (when 24 ...gf fails to 25 :g3+ h7 26 'iVh5), but this time 24 ... lLle5 is completely sound. It may

28 Closed Spanish: Zaitsev Variation therefore be said that the outcome of the opening contest is quite acceptable to Black. But then, with the move actually played, Black has not gone all that seriously wrong as yet. 24 .*.xeS! At first sight, 24 f4 looks dangerous. If Black replies with the meek 24...~g6, he has a difficult game after 2S :n ~7 26 ~3, or 26 ~b3 'Ab6+ 27 .*.d4 .a6 2SlOxh6+ gh 29 fS. However, the active 24 ...~! would give rise to a very tense struggle, for example: 2S ~xc4 be 26 .*.xg7 .*.c8! 27 .*.xf8 .*.xfS 2S .*.xh6 :xe4 29 :xe4 .*.xe4 30 .g4+ .*.g6 31 fS (31 .cS+? C;;h7 32.£8 .cS+ 33 C;;hl .d4, with advantage to Black) 31....el + 32 C;;h2 .eS+ with complete equality; or 26....*.xdS! 27 .g4 .xel+ 2S C;;h2 .e2! 29 .g3 :e6!, and Black goes over to the counter-attack. 24 ... de After 24 ...:xeS 2S 1Of3 :eS 26 .xd3, Black has no compensation for the pawn. 2S lOb3 1fb6 Another possibility was 25 ...•xa3 26.xd3 g6 27 ~3 (27.xbS 1fb4!) 27 ...•a6 28 :al1fb6. 26 .xd3 Finally this pawn is devoured, but the pawn on a3 will not run away either.

26

...

:.8

It was worth considering 26....*.c8.

27:ct g6 Taking on a3 is premature: 27 ...:xa3 2S d6 g6 29 d7! :as (or 29 ...•dS 30 .xbS .ia6 31 .xeS :xb3 32 .eS :b8 33 :c8, etc.) 30 ~3 :dS 31 ~g4 .ig7 32 ~S hS 33 ~3.*.£8 34 ~xb7 .xb7 3S OO! with a considerable plus for White. 28 M .ixa3 After 2S ...La3 291Og4, Black has serious problems defending his pawn: 29 .....d6 (29 ....ig7 30 d6!) 30.xbS .ia6 31 .eS! :xb3 32 :c6!, or 29 ... 30 1fbl .a7 31 :c7 hS 32 :xb7 .xb7 331Of6+ and wins.

:a2

29

:.1

:'4

After 29 ...h5 30 .c3 f6, Black could resist stubbornly. 30 ~g4 .its 31 :ct (19) Not 31 ~xe5? .ig7! 32lOd7 :xal+ 33 ~xal .c7, and Black wins.


31 ••• "d6? It is only now, in time-trouble, that Black makes the decisive mistake. Another line that fails is 31...i.g7 32 d6 i.xe4 33 :c8+ ~h7 34 "dl, and there is no stopping the d-pawn. But a much more tenacious defence is 31...f6 32 :a6 (not 32...~g7 33 :c6!, or 32...i.g7 33 li:Ic5), and breaking into Black's position is not simple. 32 lLIcs 33 lhe4 be 34 li:Ixb7 cd 35 lLIxd6 i.xd6 36 c;M ~g7 37 f3 f5 38 dl 39 ~e2 i.b4 40 lLId3 i.c3 41 lLIcs 1-0

"f3

:c4

m

Black resigned since he loses the d-pawn and has no saving chances in the endgame. Game No. 5 Kasparov-Karpov World Championship Match game 3/16 LondonlLeningrad 1986 1 e4 e5 2 li:IOli:lc6 3 i.b5 86 4 .i.a4li:1f6 5 0-0 .i.e7 6 :el b5 7 .i.b3 d6 8 c3 0-0 9 h3 .i.b7 10 d4 :e8 llli:lbdl .i.f8 1284 h6 13 i.c2 ed 14 cd li:Ib4 15 .i.bl c5


As I have said, the exchange on a4 is also seen. We shall study that position in detail later (Games 11-12). 16 cIS lM7 17:&3 c4 18 lM4 (20) 20 B

The knight move to d4 was first played in A.Sokolov-Psakhis, Volgograd 1985. After 18 ... lDe5 19 ab ~ 20 lLl2f3lLlbd3, White sacrificed the exchange with 21 .te3lLlxel 22lLlxel "c7 23lDef3 ab 24 lLlxb5 "d7, but failed to obtain adequate compensation. For the present game, I had prepared a surprise - 18 .....f6! - and therefore a detailed examination of the variation 18 ... lDe5 19 ab"flb6 (21) was postponed to a later date.

Before proceeding with the primary game, let us look at several important encounters that continued from diagram 21. 011-Kruppa, Uzbgorod 1987, went 20 lLlf5lLlbd3 21 .txd3lLlxd3 22 :e3 ab 23lLlxh6+!? gh 24 :g3+ .tg7 25 :axd3 cd 26lLlb3, with sharp

Closed Spanish: Zaitsev Variation 31 play; White has a strong attack in return for the sacrificed rook. In Sax-Short, Subotica 1987, White introduced a truly incredible idea: 20 ~c4!? (a bolt from the blue! White gives up a knight to clear the way for an attack on the enemy king) 20...ltJxc4 21 :g3 i..c8 22 i..xh6 ab (not 22...ltJxdS 23 ed :xel+ 24 "xel "xd4 2S :xg7+; but 22 ... i..d7 was more tenacious) 23 W :al 24ltJg5? (after the better 24 26 i..e3ltJg6 27 i..d4, White's compensation for b3! lDes 2S "d2 the piece would be fully adequate. In a slightly different situation a year later, the quiet move b2-b3 was played by A.Sokolov - see the next game from diagram 21) 24 ...:xbl? (now it is Short who goes wrong, giving up his rook for the slumbering bishop on bi. The material balance is restored, but Sax's attack is not diminished. A line recommended by Geller was worth considering: 24 ...g6! 25 i..xf8 ~xf8 26 b3 ltJa3, or 26...ltJeS 27 "d2ltJa6. It looks as if 24... lDeS 25 1i'hs g6 26 1i'h4 ltJbd3! would also have beaten off the attack) 25 "xb 1 gh 26 ltJe6+ ~h8 27ltJxfS:xf8 (27 .....d8 28 "cl "f6 fails to 29ltJg6+!, but the right move was 27 .....d4, aiming to shut the enemy queen off the c I-h6 diagonal; after all, the knight on fS has nowhere to go. Now 28 "cl would be answered by 28 .....d2, and 28 b3 by 28 ... lDd2. It is true that in the second case, after 29 "cl l:xfS 30 l:e2, some preference should be given to White; if instead 29...ltJa2, then 30 "c7 l:xf8 31 "xd6 wins) 28 "cl! ~h7 29 "c3ltJeS 30 "xb4 i..d7 31 "d2llX4 (the threat was 32 ~h2 and t2-f4) 32 "dl b4 (32...ltJxb2 33 "al!) 33 b3 ltJe5 34 ~h2 (White intends to push his f-pawn, or to activate his second rook if the pawn is taken) 34.....xt2 3S :f11Wb2 361i'h5 f6 37 l:f4! "d2?? (a terrible mistake in time-trouble. After 37...i..e8 38 "f5+ i..g6 39 "e6 "cl 40 l:xf6 :xf6 41 "xf6 "c7, it would be extremely hard to breach Black's fortress) 33 "xh6+! (the game concludes with mate on the file) 1-0. A.Sokolov-Portisch, Brussels 1988, went 20 ltJxc4ltJxc4 21 :g3 i..c8 22 b3! ltJe5 23 i..e3 (but not 23 i..xh6 ltJbd3! with advantage to Black) 23 ...ltJg6 (after 23 ... ab 24ltJfS and 25ltJxh6+, White's attack is irresistible; also 23 .....c7 fails to 24 i..xh6!) 24 f4 "d8 (Black still has no time for 24 ... ab, in view of 25 fS ~ 26lDe6 and 27 ltJxg7) 25 fS ltJeS 26 "d2 as 27 i..xh61i'h4 (27 ... ~h7 28 i..f4 i..e7 29:f1 i..f6 30 ~hl i..d7 31 "e2 :h8 was more stubborn, though after 32 ilX:6 Black's defence would still have been difficult) 28 ~h2 i..d7 29 i..gS 1i'h5 30 l:f1 g6 31 ltJc6! i..xc6 (31...i..g7 would not have relieved the position) 32 dc l:ab8 33 fg fg 34 c7 l:bc8 3S b61i'h7 36 l:xf8+:xf8 37 "xd6ltJbc6 38 i..f6 :xf6 39 "xf6 "d7 40 b7 1-0.

"as

32 Closed Spanish: Zaitsev Variation Before we return to the primary game, it remains to be said that the immediate 18 ...1i'b6 (instead of 18 ... ~S 19 ab 1i'b6 or 18 .....f6) is dangerous for Black, for example: 19lDislDes (19 ...g6 20 l:g3 ~h7 is more accurate) 20 l:g3 ~h7 (Fmcnik advises 20... g6 21 M lDed3 22 .*.e3 "c7 23 .*.xh6 ~el 24 "xel) 21 illf3 .*.c8 22 illxg7! .*.xg7 23 "d2 illbd3 24 .*.xd3 illxd3 2S l:xg7+! ~xg7 26 "xh6+ ~g8 27.*.e3 "c7 28 .*.d4 f6 29 "xf6, with a large plus; Sax-Nikolit, Lugano 1987.

18 ... "16 19 lDlf3 lDc5 (22) At this point the queen move to f6 could immediately have been justifled by 19...illd3 20.*.xd3 (20 l:xd3 is inadequate: 20... cd 21 "xd3 illcs 22 "c2 g6) 20...b4! (an important intermediate move) 21 .*.xc4 (21 l:al cd 22 "xd3 illcs 231i'bl .*.xdS, or 23 "c4 as 24 illbS l:ac8, with advantage to Black) 2l...ba 22 b3 (22 b4 l:ac8) 22 ...illcs, and Black has a pleasant game. Incidentally, this variation makes it quite clear why the queen is better placed on f6 than on b6 (after 18 ...1i'b6); in the latter case 19...illd3 is not dangerous since White can interpolate 20 as.

20

ab

A game A.Sokolov-Karpov, Rotterdam 1989, went 20 l:ee3 ba 21 l:ac3 illbd3 22 l:xc4 illxc 1 23 "xc 1 l:ac8 24 .*.c2 g6 2S illc6 .*.xc6 26 dc hS 27 l:ec3 (better 27 "al I?) 27 ...l:xc6, and we agreed a draw ten moves later.

20 21

...

ab

~bS

21 l:xa8 l:xa8 22 illxbS l:al 23 illc3 illbd3 24 l:n illb3 is no good for White. However, 21 l:ee3!?, a move 'invented' quite recently. deserves attention. After 2l...l:xa3 22 ba illbd3 23 .*.xd3 cd 24 "el "f4

Closed Spanish: Zaitsev Variation 33 25 g3 .f6 26 .td2 (threatening .tc3) 26 ....tc8 27 ~g2 .g6 28 lbh4 .f6 29 .tc3, White had a slight edge in de Firmian-Timoshchenko, Moscow 1990.

21 22

•••

:xa3

lbxa3 On 22 ba, the black queen would penetrate to a1. 22 ••• .ta6 It is not yet time for 22 ...lLlbd3 on account of 23 .txd3 lLlxd3 24 ':e3!, for example 24 ... c!l)xb2 25 .txb2 .xb2 26 c!l)xc4; alternatively 24 ....ta6 25 .a4 ':a8 26 .td2 .xb2 (26 ...lLlxb2 27 .c2! lLld3 28 c!l)xc4 .a1+ 29 ~1, or 28 ... .txc4 29 .xc4 ':al+ 30 .tel) 27 c!l)xc4 1Ib1+ 28 .tel! lLlc5! 29 .c6 ':c8 30 1Ib6 .xb6 31 lLlxb6 ':b8 32 .taS, and Black has no compensation for the pawn. For this reason I fll'st defended the pawn on c4.

23':e3

':b8

This fits in with the plan of pressurising the queenside. However, at this point 23 ...lLlbd3 was perfectly playable. Kasparov gives the variation 24 .txd3 cd 25 b4lLlxe4 26 b5 .tb7 27 :xd3lLlc3 28 .tb2lLlxdl 29 .txf6lLlxt2 30 ~xt2 gf 31lLlc4, with advantage to White. But after 27 ....al (instead of 27 ... lLlc3), it is unlikely that White can keep the pawn. 24 eSt? 24 ':c3 is less active: 24 ...lLlbd3! (this would also be the reply to 24 c!l)h2 or 24 ~1) 25 .txd3 cd 26 .te3lLlxe4 27 ':c6':a8 28.a4 d2, Anand-Timoshchenko, Fronze 1987. White now starts an attack against the king, but at the same time Black's queen and his passive bishop on f8 acquire scope for action.

24 2S 23 B

de lbxeS (23)

34 Closed Spanish: Zaitsev Variation Another critical position. The time has come to dispatch a knight to d3, but unfortunately I chose the wrong one. The correct move was 2S ...ltXd3!, when Black can look to the future with confidence. This manoeuvre was, of course, examined immediately after the game; it was first seen in practice in Nunn-Psakhis, Hastings 1987/88. After 26 ~g4 (on 26 J.xd3 ~xd3 27 :xd3 cd 28 ~7 "d6 29 ~b8 "xb8 30 "a4, a draw could similarly be agreed) 26 ...'iVh4 27 :g3 ~h8 28 J.d2 J.d6 29:t'3 ~b2 30 "e2 "e7 31 "xc7, the players concluded peace. An alternative to 26 ...'iVh4 is 26..."d4 (26...'itb6 is dangerous in view of 27 :g3 g6 28 J.e3) 27 ~2 ~xc2 28 J.xc2 (24) 24

B

This position occurred in two games between Dvoirys and Timoshchenko (USSR 1988). In the Semi-Final of the national championship in Bamaul, Timoshchenko played 28 ... J.c5, and after 29 "f3! ~xcl (29 ...:t'8 30 :e4!) 30 ~6+! ~h8 (30...gh 31 "g3+ ~ 32 "xb8+ ~g7 33 :g3+ ~6 34 "d8+ ~e5 35 :e3+ wins) 31llli5 "xe3 (otherwise White mates with 32 'iVh5+ ~g8 33 ~7+ etc.) 32 fe :xb2 33 ~7! J.xe3+ 34 "xe3 :xc2 35 "e5, he acknowledged defeat. In the other game, Timoshchenko chose the stronger 28...J.d6!, and there followed: 29 b3 (or 29 :e2 ~xcl 30 "xcI "xd5 31 ~xh6+ ~f8! 32 :el, with unclear play; 32 :d2 is bad on account of32 ...J.f4! 33:Xd5 J.xcl) 29 .....aI30bcJ.xc4 31 J.xd3 J.xd3 32:el J.g6 33 J.d2 :b134 "e2 :xel+ 35 "xel "xel+ 36 J.xel J.e4 37 ~3 J.c5 lh-1h. In fact, we can now draw a line under the results of the Spanish theoretical duel in the 1986 return match. The variations I have quoted show that Black's plan has firmly stood up to examination. We shall go more quickly through the remaining - and bewildering - part of the present game.


25 26

..•

~bd3?

~g4?

My opponent returns the favour. Instead, 26 "c2! could have had grave consequences for me. The pawn on f2 is defended, the queen has occupied an aggressive diagonal, and as a result the knight on d3 is unpleasantly pinned. Now 26...~b3 fails to 27 ~7, and 26...:d8 to 27 lOaxc4 :xdS 28 b4. H 26...:b4, then 27 ~6 :b7 28 :e8 g6 29 .te3 .rI.xb2 30 .td4 with a clear plus for White, though the game is not yet over. After White's inaccuracy, a state of dynamic equilibrium is restored.

26

•••

""6

27 :g3 g6 A solid continuation, though 27 ... ~h8 and 27 ...~e4 are also playable. 28 .txb6 "xb2 29"f3 tM7 30.txfB ~ 31 ~h2 :b3! (25) Many annotators criticised this rook move. In actual fact it is stronger than 31.. ...xa3, 31.. ...cl or 3l...~g7.

32 .txd3 cd?? A ghastly mistake; let me try to explain it. In the first place, I have to admit that when I started analysing the position, I assumed that Black already stood better. Indeed, White's queenside has been decimated, and his kingside attack appeared to me to be ineffective. I naturally began by considering the move 32 ...cd. I studied it for a long time and eventually recognised that it was unsatisfactory. Thming to other variations, I kept discovering that Black has, alas, no advantage. Having

36 Closed Spanish: Zaitsev Variation established that depressing fact, I took a look at the clock and saw that the hand was getting close to the fatal hour-mark. Mechanically, I played the one move that I had thought about most. I had of course seen the following simple variation: 32...ltxd3 33 1i'f4 1i'xa3 34%6 1i'e7 35 lhg6 1i'e5 36 1i'xe5 (in this case, the combination played in the game does not work: 36 ltg8+ ~e7 37 d6+ ltxd6 38 c!Of5+ ~f6 39 "xe5+ ~xe5 40 ~xd6 ~xd6) 36...~xeS 37 ltxa6 ltxdS 38 lta8+ ~e7 39 ~f5+ ~e6 40 ~3, and White's nominal endgame advantage cannot be turned into a win. An even stronger move is 32...ltxa3; Kasparov gives the following long variation: 33 1i'f4 ltxd3 34 "d6+ ~g7 35 1i'xd7 ltxg3 36 fg .i.b7 37 h4 .i.a8 38 1i'd8 1i'd4 39 "xa8 "xg4 40 "a1+~! 41 d6 ~e8 42 "a4+ ~d8 43 "as+ ~e8 44 'ifb5+ 'iVd7 45 "e5+ 'iVe6, with a draw. 33 "r4 "xa3? This capture in severe time-trouble throws the game away irrevocably. Afterwards it was discovered that 33 ...d2! would have kept some saving chances. However that may be, I consider 32... cd in the present game to be the most dramatic mistake of the 1986 match. 34 ~6 "e7 3S ltxg6 "eS 36 ltg8+ ~e7 (26) 26

W

37 38 39 40 41

d6+! lte8+ lheS+ d7 fi)xf1 1-0

~e6 ~dS ~eS

ltb8

Closed Spanish: Zaitsev Variation 37 Game No. 6 Kasparov-Karpov

World Championship Match game 5/4 New YorkILyons 1990

m

m

1 e4 eS 2 ltX6 3 JobS 86 4 Jo84 5 0-0 Joe7 6 :leI b5 7 Job3 d6 8 c3 0-0 9 h3 Job7 10 d4 :le8 lllDbd2 Jof8 12 84 h6 13 J.c2 ed 14 cd lDb4 15 Job! c5 16 dSlDd7 17 :la3 17 5!? (27)

This counter-stroke in the centre has been known for some years, but it was evidently not until the present match that it underwent such serious investigation. 18 ef In the 20th game of the match, Kasparov preferred 18 :lae3 (see the notes to Game No.8 in this book); afterwards, the manoeuvre 18lDh2 was devised (see Game No. 10). But this is not all. The break 18 eS!? led to a quick win for White in Raaste-Rantanen, Helsinki 1990: 18...g6 19 e6lDb6 20 g4 .f6 21 gf gf22lDh2 :le7 23 :lg3+ :lg7 24lDe4! fe 2S lDg4 :lxg4 26 .xg4+ Jog7 27 .xe4 JoxdS 281i'h7+ 1-0. Of course, the e-pawn should have been eliminated: 18...de! (better than 18...lDxeS 19 lDxeS de 20 J.xfS with a very strong attack) 19 J.xfS JoxdS (Sznapik recommends 19 ... lDxdS!?, threatening ... cS-c4) 20 ab lDf6 (20...ab at once was sounder) 21 lDxeS ab 22 J.g6, and White has some initiative; Sznapik-A.Ivanov, Bie11990. 18 ••• 18...J.xdS 19lDe4lDf6 amounts to a transposition, but in the 22nd match game I decided not to wait for a surprise from my opponent, and introduced a novelty myself: 19...J.f7!1 (see Game No.7).

m

38 Closed Spanish: Zaitsev Variation 19

~

In Vasiukov-Razuvaev, Moscow 1987, the first game in which the counter-attack with ...f7-f5 was seen, Black achieved a fine game after 19lbh2 Jbel+ 20 "xel "e7 21 ..n :'e8!. Another possibility was tried out in A.Sokolov-Hjartarson, Manila 1990: 19 :'xe8 "xeS 20 lbh4 J.xdS (20... lDbxdS also deserves attention) 21lDg6 (in a later game Zsu.Polgar-Kamsky, New Delhi 1990, White played the weaker 21 :'g3?!, and after 2l...J.a2!? 22 J.xa2 lDxa2 23 lDdf3 lDxcl 24 "xcI ci?h7 25lDg6 d5! 26lDfe5 d4! 27lDg4 WI? 28 :'d3 "e4! 29lDxf8+ :'xf8 30 "xc5lDf4! 31:'f3 :'xfS 32 "d6 h5 Black obtained a large plus) 2l.....f7 22lDn ba 23 "xa4 J.e4 24 J.xe4lDxe4 25lDg3 :'e8 26 J.f4lDxg3 27 :xg3 lbel+ 28 ci?h2 "xf5 29 J.d2, and now after 29 ...:.n! (instead of 29 ...:'e2?) 30 "e8 .xf2 31 J.e3 "el 32 "e6+ ci?h7 33 lDxf8+ ci?h8, the game would have ended in perpetual check. 19 ••• J.xdS (28) I had prepared this move specially for the match. After 19...lDbxd5 20 ab ab 21 'ii'b3, or 20 lDxf6+ "xf6 21 :xe8 :'xe8 22 ab ab 23 :'a7, White has a substantial initiative. Less dangerous alternatives are 20 lbh2lDxe4 21 J.xe4 :xe4! 22 :'xe41Dc3 23 :'xc3 J.xe4 24 :'g3 ci?h8 with a good game for Black, de FirmianA.lvanov, Las Vegas 1989; and 20 lDh41Dc7 21lDxf6+ "xf6 22lDg6 :xel+ 23 .xel :'e8 24 :'e3 lbe3 25 "xe3 ba 26 J.a2+ J.d5 27 J.xdS+ lDxdS 28 "e4liJe7 with equal chances, de Firmian-A.lvanov, San Mateo 1989.

20 lOxf6+ Complex play would result from 20 :'ae3 J.xe4 21 J.xe4 dS 22 J.bl :'xe3 23 :'xe3 d4.

Closed Spanish: Zaitsev Variation 39 20 "xl6 21 .i.d2 Mter 21 ltae3 .i.f1, Black has everything in order. A game Rantanen-Ojanen, Finland 1990, continued differently with 21 ab ab 22 .i.d2 ltxa3 23 ba.i.xf3 24 "xf3 ltxel+ 25 .i.xel ~2 26"e4 d5 27 ~h8 28 ab "xbl 29 be .i.xc5 30 f6 gf 31 "xf6+ ~g8? (31...~h7! would have drawn; the wrong retreat with the king leads to loss) 32 "e6+ ~g7 33 ~h2 "cl 34 "d7+ ~f6 35 "c6+ ~e7 36 "c7+ ~ 37 .i.b4 1-0. 21 "xb2 22 .i.xb4 .i.f7! (29) This is the manoeuvre Black had in mind when he took the pawn on d5 with the bishop. An inferior line is 22....i.xf3 23 ltxf3 "xb4 24 lte6 "xa4 25 .i.c2 "d4 26 ltd3 1Wh4 27 ltde3. with a dangerous initiative for White. 000

"e6+

000

23 .Ile6! The only way to fight for the initiative; after 23 ab "xb4 24 ltxe8 ltxe8 25 ltxa6 'iVxb5 26 lta7, the chances are equal. 23 'iVxb4 It doesn't pay to accept the exchange sacrifice: 23 ....i.xe6 24 fe cb 25 ltb3 "f6 26 lte3 ba 27 'iVd5 .i.e7 28 'iVd3 g6 29lDh4. 24 ltb3! 24 ltae3 is no good in view of 24...ba! 25 .i.a2 c4. 24 'iVxa4 24 .....c4. threatening 25 ...ba, was probably even stronger; after 25 .i.d3 'iVd5. Black has the advantage. 25.i.c2 :adS 25 ...ltxe6 is risky: 26 fe .i.xe6 27 lte3 (after 27 ~g5? it is White 000

000

40 Closed Spanish: Zaitsev Variation who loses: 27 ...hg 28 'ii'hs 'iih4!) 27 .....a2 28 "d3 ':e8 29 "g6! (clearer than 29 'iih7+ ~f130 -*.g6+ ~e7 31-*.xe8 ~xe8 32 "g8 ~e7 33 ~4 .al + 34 ~h2 .f6).

26':be3 27 g3!

'ifb4

.f7!

White would have the advantage after 27 "e2!? "c4 28 ':xe8 ':xe8 29 ':xe8 -*.xe8 30.xe8 .xc2 31 .e6+ ~h7 32 .cl+ 33 ~h2 "f4+ 34 ~hl .cl + 35lOg11i'b2 36 "xf8 .xf2 37 lOn, but after the correct 28 ...•xe2 29 ':xf8+ ~xf8 30 ':xe2 Black has a good position.

27

•.•

as (30)

Black has an avalanche of pawns on the queenside, but he begins by advancing the wrong one; the immediate 27 ...d5! was much stronger. Let us look at these variations: (a) 28 ':xe8 ':xe8 29 ':xe8 -*.xe8 30 "xd5+ -*.f1 31 1i'b7 .c3 32 -*.e4 (or 32 "c4!) 32...•f6 33 g4 "e7 34.c6 "e8! 351i'b7 .e7!. (b) 28 "e2.c4 29 ':xe8 .xe2 30 ':xf8+ ~xf8 31 ':xe2 d4. (c) 28 lOe5 d4 29 ':b3 -*.xe6! 30 fe (or 30 ':xb4 -*.xf5! 31 -*.xf5 ':xe5!) 30... d3 31':xb4 dc 32 "xc2 cb. In all cases Black has absolutely no problems.

.e4

30 W

28 ~4 At this point, according to Azmaiparashvili (Kasparov's second), the game could have ended in repetition if the players had wanted: 28 ~g2 a4 29 .e2.c4 30 -*.d3 .d5 31-*.e4! (but not 31-*.xb5?! -*.xe6 32 fe ~e7) 3l....c4 32 -*.d3. Well, perhaps that result would have been objectively in accordance with the position. 28 d5

29"e2

"c4!

Closed Spanish: Zaitsev Variation 41 30 i.d3 'iVc1+ 31 ~g2 31 ~h2 is weaker; the pawn on f2 needs to be guarded. 31 ••• c:4 32 i.el The only move; after 32 llxe8 cd 33 llxf8+ llxf8 34 'iVxd3 'iVc4, Black has an obvious plus. 32 ••• i.xe6 White has distinct pressure on the light squares, but I do have two extra pawns. At this point it may have been worth pushing the d-pawn at once: 32 ...d4 33 llxe8 d3 34 llxf8+ ~xf8 35 i.xd3 cd 36 llxd3 'iVc6+, and White's position is not to be envied. 33 llxe6 llxe6 34 'iVxe6+ ~h8 35 lbg6+ ~h7 36 'iVe2?(31) White loses with 36lbxf8+? (36lbe5 lld6) 36...11xf8 37 'iVg6+ ~g8 38 f6 'iVg5! 39 fg 'iVxg6 40 gtW+ ~xf8 41 i.xg6 a4; or with 36lbe7? i.xe7 37 'iVg6+ ~g8 38 'iVe6+ ~h8! 39 'iVxe7 'iVg5. However, it seems that he could have saved himself with 36 'iVb6 llc8 37 'iVe6 lld8 38 'iVb6, drawing. Thus, taking the exchange has not proved justified.

36 ••• 'iVg5? In time-trouble I miss the win with 36 ...d4! 37 f6 (37 i.e4 d3 38 'iVf3 d2) 37 ...d3 38 'iVe4 'iVg5 39lbxf8+ ~g8 40 lbd7 llxd7 41 'iVe8+ ~h7 42 'iVxd7 dc 43 f7 'iVf6, and it is allover. 37 f6 'iVxl6

38

lbxf8+

~g8

39

lbg6

'iVf7?

42 Closed Spanish: Zaitsev Variation Here 39...d4!? was the final winning chance. After 40 .i.fS 'lVc6+ 41 .i.e4 'lVd6! (but not 41...lle8 42 .i.xc6! he243 .i.dS+
Game No. 7 Kasparov-Karpov World Championship Match game 5122 New YorkILyons 1990 1 e4 e5 2 m ~ 3 .i.bS a64 .i.a4lM6 5 0-0 .i.e7 6 Ilel bS 7 .i.b3 d6 8 c3 0-0 9 h3 .i.b710 d4 1le811 ~bd2 .i.1812 a4 h613.i.c2 ed 14 cd ~b41S .i.bl c:S 16 dS lild717 1183 IS 18 ef 18 .i.xdS 19 lLle4 .i.f7! (32) 32 W

An innovation, after which Black obtains a promising position. Black's counterplay in the fourth match game was similarly based on this retreat with the bishop, but in the present case it occupies the f7 square immediately. The idea was conceived by Portisch. 20 ab Black benefits from 20 .i.f4 dS 21lild6 .i.xd6 22 .i.xd6 'IVb6! 23 .i.f4 d4, while 20 ~d610ses outright to 2O...hel + 21 ~el ~b6 22 .i.f4 lLlc4 23 Ild3 .i.xd6 24 Ilxd6 lilxd6. However, 20 .i.d2 and 20 llae3 are worth trying. 20 ... dS!

Closed Spanish: Zaitsev Variation 43 This pawn sacrifice is the key to Black's idea.

21

M

lbe1+

22 ~el d4 23 lDa2 The knight can take a different route: 23lbe4 ab 24 f6 g6 (better than 24...lba3 25 ba 00 26 fg j,.xg7 27 llXi6; Black would lose with 24 ...1t1xf6 251t1xf6+ gf26 "g4+ cRh8 27"f5 j,.g8 28 j,.xh6! j,.xh629 'iVh5 etc.) 25 IlxaS (25 "g4 Ilxa3 26 ba 00 27 'iVh41t17xf6 28 j,.xh6 j,.xh6 29 "xh61t1xe4 30 j,.xe4 "f6 is in Black's favour) 25 .....xa8 26 "g4 "ai, with complex play. After 23 j,.e4 dc 24 j,.xa8 "xa8 25 "xd7"e4 26 "dl c2 27 "d2 ab, there are chances for both sides; Black may also play 23 ...lla7 24 ltle21t1f6 25 j,.f3 d3, with compensation for the pawn. 23 ••• ltlxa2 I also thought about 23 ... a5, but came to the conclusion that after 24 ltlxb4 ab 25 1lxa8 "xaS 26 b3, White retains a slight edge. 24 j,.:xa2 c4! Of course, 24 ... j,.xa2 25 Ilxa2 ab 26 'iVb3+ c4 27 "xb5 1lxa2 28 "xc4+ is no good for Black.

25

lba6

ltlcS!

26 Ilxa8 ..xaa 27 j,.bl d3 After 27 .....al 28 j,.f4 "xb2 29 j,.e5 ltlb3 30 b6, Black's position gives cause for concern. He can play more accurately with 28 ...llXi7 (29 f6? "xb2), or 28 ...d3 29 j,.e5lbe4 30 j,.d4 j,.c5 31 j,.xc5lLlxc5 32 f6 "xb2 33 "g4ltle6 with a tense situation. The position in the match (this was my last chance of winning it) forced me to play more sharply. 28 j,.e3 (33)

"as

33 W

44 Closed Spanish: Zaitsev Variation Black is two pawns down, yet all his pieces have taken up ideal posts, while for example the opponent's bishop on bl is completely shut out of play. However, White's pawn advantage permits him to carry out a freeing operation involving a piece sacrifice. 29 b3! More convincing than 29 f6, 29 00 'ilxb5, or 29 ~xd3 at once. 29 ~b3 30 ~d3! cd 31 .i.xd3 lilc5 Perhaps 31...'i1b4 promises more, though after 32 'ilg4 'ilxg4 33 hg a draw is in prospect. 32.i.n 'ilc7?! A last attempt at fighting for the initiative was 32...'iIb4!? 33 'ilg4! ~h7? After 33 ...h5 34 "d4, a peaceful outcome is inevitable. 34.i.c4 .i.xc4 There is no avoiding the bishop exchange; 34....i.e8? (34 ...h5 35 'ile2!) 35 .i.xh6 ~xh6 36 'ilh4+ .i.h5 37 g4 etc. 35 "xc4 'ileS 36 'ilf7 .i.d6 A safer line was 37.....f6 38 'ilxf6 gf, with a drawn ending. 37 g3 'ile7 And here, 37 ...~e4 was sounder. 38 'ilg6+ ~h8 (34)

39 .i.d4 A draw in this game enables Kasparov to retain the world title, so he brings it about by force. (Of course he could have done so much earlier, with 11 ~g5 etc., but that would hardly have looked respectable,

Closed Spanish: Zaitsev Variation 45 whereas now the contest has been fought out.) Objectively, 39 b6! was stronger. Now after 39 ...~g8 40 f6, or 39...tM7 40"e6!?, or 39...lDe440"e6-with thethreatof41"xe7, 42 b7 and 43 .*of4 - Black would be in quite a dangerous position. If instead 39...clL!b7 40 f6 gf (40.....xf6? 41 "e8+ ~h7 42 "e4+) 41 "xh6+ ~g8, a draw would probably have resulted just as in the actual game, but in this ending White's chances are of course better. 39 .*oeS 40 .*oxeS "xeS 41"e8+ ~b7 42"16+ ~g8 43 "e8+

1z

1/2· 1

Game No. 8

Kasparov.Karpov Amsterdam 1990 1 e4 eS 2 clL!f3 clL!c6 3 .*obS 86 4 .*084 clL!f6 5 0·0 .*oe7 6 :e1 bS 7 .*ob3 d6 8 c3 0-0 9 b3 .*ob710 d4:e8 11 clL!bd2 .*om 12 84 b6 13 .*oc2 ed 14 cd clL!b41S .*ob1 eS 16 dS lOci717 IS 18 he3 The consequences of 18 ef were examined in the two foregoing games, while 18 clL!h2 is the subject of Game No. 10. 18 ••• clL!f6 18... f4 will be studied in the next game. 19 clL!h2 (35)

:a3

35 B

This knight retreat had brought Kasparov success in the World

46 Closed Spanish: Zaitsev Variation Championship Match. Before turning to the improvements I had prepared for the Amsterdam encounter (which took place a few months later), let us recall the match game. Kasparov-Karpov (SI20): 19...~h8 (evidently not the most effective move, though the king does feel more secure in the comer) 20 b3 ba 21 ba c4 22 .i.b2 fe (22 ...':c8 23 .i.c3 is more solid) 23 lLlxe4lLlfxdS 24 ':g3! (on 24 'ii'hs, Black would lose with 24 ...lLlxe3 2S "xh6+ ~g8 26 lLlgS! !, but instead he has the good retort 24 ...c3!, for example 2S lLlxc3 lLlxe3 26 "g6lLlec2; or 2S "g6 c2 26 "xh6+ ~g8 27 ':g3 "c7; or 2S lLlgS "xgS 26 "xgS hg 27 :Xe8 cb 28 ':xa8 .i.xa8 29 ':e8 ~g8 30 l:txaSlLlc3, with winning chances for Black) 24...l:te6! 2S lLlg4 "e8? (36)

as

(in such a sharp position, one incautious move can be fatal. That is the case here. It was essential to play 2s ... lLld3 26 .i.xd3 cd 27 l:txd3 with roughly equal chances - but not 27 ...lLlf4 28 l:tde3lLldS, on account of 29 lLlgS!. White now succeeds with a quick attack) 26 lLlxh6! (although some quite complicated variations arise here, I am inclined to omit them; the opening contest can be said to have concluded in White's favour, and to recall this game in detail is none too pleasant an affair) 26 ...c3 (accepting the sacrifice would not have saved Black either: 26 ...l:txh6 27lLlxd6 "xel + 28 "xe! l:txd6 29"e4:h6 30 ':g6 l:th7 31 :Xg7! and mates) 27lLlf5! cb 28 "g4 .i.c8 29'iVh4+ l:th6 30 lLlxh6 gh 31 ~h2! 32lLlgS! "f6 33 l:te8 .i.fS 34 "xh6+ "xh6 3S lLlf7+ ~h7 36 .i.xfS+ "g6 37 .i.xg6+ ~g7 38 ':xa8 .i.e7 39 l:tb8 40 .i.e4+ ~xf7 41 .i.xdS+ 1-0. For my next game with Kasparov in this variation I prepared a new move, and on this occasion things turned out much more successfully for Black.

"as,

"eS

as


19

.d7!? (37)

37 W

20

ef

The innovation 20 :t3 was employed in OIl-Hjartarson, Philadelphia 1991. Play continued 20 ...':eS! (it is easy to see that an immediate capture on e4 favours White) 21 b3 (if 21 ':xfS ':xfS 22 ef, Black can take on dS with the bishop or the knight on b4) 21...lDxe4 22lDxe4 fe 23 lhe4 ':xdS (23 ....i.xdS is inadequate: 24 ':xeS de 25 ll)g4 with an attack) 24 .e2:eS 2SlheS de 26 ':g3 ':d8 27ll)f1 e4!? (Hjartarson gives the drawing line 27 ...•dI28 .xeS .xc129 .e6+ "'h8 30 .g6 "'g8 31 'it'h7+ "'f132 .f5+ "'g8 33 'it'h7+) 28 ab ab 29 .txh6 .d130 'it'b2 .d4 31 .e2 .dl 32 'it'b2 .d4. Now after 33 .e2 the game would have ended in repetition. Instead there followed 33 .cl? ll:ki3 and Black seized the initiative, though a draw was the eventual result. In Game No. 10, White will bring his rook to f3 in one go: 19 :a3f3. But again Black will obtain a good game.

20

...

]be3 .

This is more precise than 20...ll)bxdS 21 :e6ll)f4 22ll)g4. 21 fe After 21lhe3ll)bxdS 22 ':g3 :eS, Black has nothing to fear. 21 .txdS 22 ll)g4 .i.e7! 23 e4 .tn

24

m

24 e5 de 25ll)xeS .d4+ 26 "'hI ':d8 gives White nothing.

24

':d8

2S.e2 26 .e3(38)

.tc4!

At this point 26 .f2! is interesting. Then 26....td3 27 .g3! lDxg4

48 Closed Spanish: Zaitsev Variation (27 ... j,xb1? 28 lbxh6+ ~f8 29 lbg5 j,a2 30 b3, or 27 ...~h8? 28 j,xh6! W 29 "f2! j,xbl 30 j,g5 lb
"C

38 B

26 27

bg

lbxg4! j,f6

28 j,dl (39) A critical moment; the position after 28 e5 de 29 "xc5lbc6 30 g5 hg 31 j,xg5 j,xg5 32 lbxg5 "d4+ would have offered mutual chances. Now Black gradually begins to dictate matters. 39 B


28

•••

1te8!

It doesn't pay to be greedy with 2S ....i.xb2; after 29 gS! White has sufficient compensation for the pawn. 29 b3 29 .i.xb4 is no good in view of 29...cb 30 .d2 b3, when 31 eS? fails t03l....a7+. 29 ••• .i.rr 30.t2 .e7! A position with chances for both sides would result from 30....i.xb3 31 ab ab 32 gS, or 30...ba 31 ba .xa4 32 gS.

31

ab

32.n

ab 1tb8

33 .i.d3 Again, taking off the knight on b4 is unfavourable: 33 .i.xb4 cb 34 .i.c2 .c7, with a clear plus for Black. 33 ••• lffiId3 If 33 ....d7, then 34 gS is unpleasant. 34 .xd3 c4 35 be be 36.a3 Itb3 37 .a8+ .e8 Removing the king to h7 is not so good: 38 gS! hg 39 ~f2, threatening 40 l:thl+. 38 .xe8+ .i.xe8 39 gS At last White has carried out the thematic pawn move, but without queens it has somewhat diminished in strength. The initiative has entirely passed to Black, who in addition has the advantage of the bishop pair. Incidentally, the attempt to attack the d-pawn with 39 £f4 is easily parried by 39...c3! 40 £xd6 c2 41 .i.f4 Itbl. 39 ••• hg 40 £xg5 £b2 (40) 40 ...ltxf3 41 .i.xf6 would lead to a draw. 41 Itdl? Now White's position becomes extremely difficult, although subsequently I missed the win that was certainly there. In a detailed analysis, Kasparov has shown that after the correct move 41 eS!, White could have counted on equalising. Here are the main variations he gives: (a)41...de42~xeS c3 43 f6! gf44.i.xf6c24SlM3 .i.xf646ltxeS+ ~f147 ItcS.


40 W

(b) 4l...dS 42 e6 (42 f6? .i.g6! 43 fg d4!) 42 ...c3 43 f6 gf 44 .txf6, and now:

(bl) 44 ...c2 45 .txb2 Ilxb2 46 ~4 .ta4 47 e7 ~f7 48 g4! ~e8 49 Ilb6 50 ~t2 .td7 51 lleI .txfS 52 gf Ilc6 53 ~e3. (b2) 44....ta3 45 ~4 .tc5 46 ~t2! ~h7 47 ~e2! Ilb2+ 48 ~d3 c2 49lOxc2! (49Ilcl? .tg6+50~c3 Ilb151 Ilxc2.tb4mate)49 ....tg6+ 50 ~c3 Ilxc2+ 51 ~b3 .tf5 52 e7, with a draw. 41... .taJ 42 eS c3! But not 42...de 43 Ild8 ~ 44 lOxe5 c3 45 ~g6+ ~f7 46 ~5+, with perpetual check. ~

43

ed?

A more tenacious line was 43 ~4 Ilb2 44 ed .txd6. 43

44 45

c2

llelS ~

46:as

Ilbl+ Ildl

cHi'

47 .txc1 .txc1 48 1la6 48 :al loses to 48 ....te3+, while 48 llaS loses to 48 ...~f8! (48 ...~t7? 49 ~5+ draws) 49 ~ Ilxd6 50~g6+~f7 51 lla7+~. 48 ... llelS 49 f6 IT 49 g4, then 49....tf4! is very strong. 49 ... g6 The pawn must be preserved; 49 ... g5 would be met by 50 d7!. 50 g4 gS Black could also move his king towards the centre: 50...~ 51 gS! .tf4 (if 51.. ..txgS, then 52 lOxg5+ Ilxg5 53 :a7+ ~xf6 54 d7) 52


:a7+ ~f8, and Black rounds up the d-pawn. 51 :.7 £f4 (41) Taking the d-pawn at once was stronger. After 51...:xd6, the struggle might continue 52lDxg5 :xf6+ 53 lbf3 £c6; or 52 :g7+ ~f8 53 lbxg5 :xf6+; or 52 :c7 £f4 53 lbxgS :d2+ 54 ~e1 £xc7 55 ~xd2 £f4+. In all variations Black has an easy win. 41 W

52

£xr1

f7+

53 d7 ~ 54 lbgl :d2+? A dubious continuation, whereas after 54 ...£e6 55lbh3 :d2+ (or 55 ...£d2) 56 ~e1 (56 ~f3? £d5 mate) 56...:xd7 57lbxf4 gf58 l:a4 Black would have gained a decisive plus. 55 ~el :d5? Another inaccuracy. Black could have won with 55 ...£e6! 56lbh3 :xd7 57 lbxf4 gf (57 ...:xa7 58 lbxe6+ and 59 lbxg5) 58 :a4 :n. 56 lbh3 £d2+? Instead of this, Black should of course have chosen 56 ...£e3! (so as to answer 57 with 57 ...£d2+) 57 :b7 (57 :a3 is bad in view of 57 ...£b6 followed by 58 ...£d8) 57 ...£e6 58 :b5 ~e7. 57 ~e2 ~g7? Only now - at last - did White begin to breathe freely. A 'rook and bishop against rook' ending would have arisen after 57 ...£c1 58 l:c7 £e6 59 :xc1 £xg4+ 60 ~ £xh3 61 :xd7 62 :xg5. 58:al! £c1 59:c2 £a3 60 lbxg5! £g8 61 :c7 ~g6

:n

:as

:cS

lJ2-1h


Game No. 9 Timman-Karpov Candidates Final, 9th game Kuala Lumpur 1990

1 e4 eS 2 ~f3 ~c6 3 j,b5 a64 j,a4 ~f6 5 0-0 J.e7 6 :el b5 7 J.b3 d6 8 c3 0-0 9 h3 J.b710 d4 :e811 ~bd2 J.t8 12 a4 h613 J.c2 ed 14 eel ~b4 15 J.bl cS 16 dS ~7 17:.3 fS 18l:ae3 18 ••• f4!? 19 l:3e2 (42) In de Firmian-A.lvanov, Chicago 1988, the rook returned to a3. After 19 :a3 'it'f6 20 ~b3 ~b6 21 ~a5 :ab8 22 ab ab 23 ~xb7 :xb7 24 J.d2 ~4 25 J.c3 ltles 26 'it'e2 'it'f7 27 ~xe5 (better 27 :a5! ~xf3+ 28 'it'xf3 g5 29 J.xb4 cb 30 e5 :xe5 31 :Xe5 de 32 :a6, with some advantage to White) 27 ...de 28:a5 c4 29 'it'd2 'it'e7, Black obtained a good game. The Dutch grandmaster prefers to retreat his rook to e2.

19

•••

20

m!?

~

The exchange on e5 is not promising: 20 ~e5 de 21 ffi J.d6 (but not 21...g51! 22 J.d2 h5 23 J.xb4 cb 24 d6! with a dangerous initiative) 22 J.d2 c4 23 J.c3 ~3! 24 J.xd3 cd 25 'it'xd3 b4 26 J.d2 as gives Black the advantage, although unclear play results from 22 b3, and if 22 ...g5 then 23 :d2, followed by ~h2. 20

•••

lOxf3+

21 gf ....4 If 21 ...g5, White has the sharp reply 22 e5. 22 ~ :eS 23 'it'd2(43)


"'xh3

23 ... But here 23 ... g5 was seriously worth considering, for example 24 lLlg4 £g7 2S b3 26 £b2 £c8 with a wealth of possibilities for Black. If 23 ...:g5+, White's only reply is 24 lLlg4 (24 Whl loses to 24 ......xh3 2S :gl £e7 26 "'xf41 :xgl+ 27 Wxgl £gS). Events may then take the following course: 24 ...hS (or 24 ......xh3 25 "'xf4l:r.hS 26 'iVh2 ba 27 "'xh3 :xh3 28 ~g2) 2S "'xf4 £e7 26 "'g3 "'xg3 (26...:g6 27 e5) 27 fg hg 28 £xgS £xgS (2S ... gf is bad in view of 29 i.xe7 fe 30 £xd6) 29 f4 £e7 (29...£f6?! 30 eS £e7 doesn't work here: 31 ed £xd6 32 :e8+ :xe8 33 :xeS+ ~ 34 :e6 J.f8 35 £g6+ ~g8 36 d6 £c6 37 ab ab 38 hg) 30 hg ba, with chances for both sides. With this we may conclude our discussion of the opening, but there is a fascinating middlegame in store. 24 "'xf4 ba But not 24...:g5+ 2SlLlg4 £c8 26 "'g3!. 25 "'g4! "'xg4 26 lLlxg4 27 f4 The position arising from the queen exchange should be assessed as somewhat favourable to White. 27 as 28 f3 J.a6 29:g2 Wf1 (44) If 29 ...WhS, then 30 f5 is unpleasant, while 29 ...£d3 fails to 30 lLlf6+ Wf7 31lLlxe8 J.xbl 32lLlc7. 30 :dl An interesting alternative was 30 eS!? lLlxdS 31 £e4 (there is no danger to Black in 31lLlxh6+1! gh 32 £g6+ ~e7 33 ed+ Wd7, but a

:ae8

:ee8


44 W

possibility is 31 .th7 h5 32 t'bh6+ gh 33 .tg8+ ~e7 34 .txdS) 31...lDb4 32 .txaS l:txaS 33 e6+ ~e7 34 f5. White would now retain a noticeable plus after 34... clOd3?! 35 ':dl! ':b8 36lOe3lbf4 37 ':g4, or 34....td3 35 .txh6! gh 36 f6+; but after 34...lDd5 he would just have slightly the better chances. 30 .tc4

31

ll)e3

32

':el

33

e5! 34.tg6+

.tb3 c4

de ~g8

~4 The position is unclear after 35 .txe8 ':xe8 (35 ...~3? 36.tc6!) 36

35

lbg4lDd3. ~

35 36 37 38 45 W

lOxb6+

gh

.txd3+

~h8

.tg6

':ed8 (45)

Closed Spanish: Zaitsev Variation 55 At this point, in time-trouble, Timman played 39 .i.d2, and after 39....i.b4! Black had no further worries. The game would be equal after 39 .i.fT .i.cS+ 40 ~h2 :a7. Some fantastic variations couid arise from 39 fer :XdS 40 e6! (40 .i.e4 :XeS 41 .i.d2 .i.b4 42 .i.xaS :xe1+ 43 .i.xel .i.xel favours Black), and now: (a) 4O•••:dl is inferior: 41 :Xdl .i.xdl 42 .i.d2! :a7 (42....i.b4 43 e7! .i.xd2 44 :xd2) 43 .i.e3! :e7 (43 ...:88 44 .i.d4+ and .i.c2+) 44 .i.c5 :xe6 45 .i.xfS .i.xf3 46 :g3, and White wins. (b) Another line that fails is 40 ....i.c5+ 41 ~hl (41 ~f1 c3!) 4l...:dl (4l....i.b4 42 .i.e4 :e5 43 .i.xh6) 42 i.d2! :xd2 43 :xd2 .i.b4 44 :de2, and again White wins. (c) 40... i.b4 might seem more precise, but after 41 .i.e4 :e5 42 .i.xh6! Black still cannot find his way to safety, for example: (cl) 42....i.xeI43 .i.g7+ ~g8 44 .i.xeS+ ~ 45 .i.d6+ ~e8 46 :g8 mate. (c2) 42...:Xe6 43 .i.g7+ ~g8 44 .i.dS .i.xel 45 .i.xe6+ ~h7 46 .i.f5+ ~g8 47 .i.c3+. (c3) 42...i.cS+ 43 ~f1 :a7 44 .i.d2! .i.b4 45 :h2+ ~g8 46 .i.h7+! :xh7 47 :XeS :m2 48 .i.xb4 ab 49 e7, and it is allover. Attempts to guard the critical g7 square at once are no help either: (c4) 42 ...:g8 43 :Xg8+ ~xg8 44 .i.h7+ ~xh7 45 :xeS ~xh6 46 e7 .i.xe7 47 :Xe7, and wins. (cS) 42...:a7 43 .i.e3! .i.c5 (or 43 ...:g7 44 .i.d4 .i.c5 45 .i.xcS :xg2+46~xg2:xc547 .i.bl)44i.xc5 :Xc5 45 ~f2! :g7 (45 ...:bS 46 .i.g6!) 46 .i.dS! :xdS (or 46...c3 47 .i.xb3 cb 48 :h2+:h7 49 :xh7+ ~xh7 50 e7 :c8 51.i.xa4) 47 :Xg7 ~xg7 48 e7, again winning. The above variations were published in the Swiss magazine SchachWoche. They seem to indicate that by playing 39 fer :xdS 40 e6, Timman would have retained chances of victory. However, Black can defend more stubbornly: (d) 40...:a7! avoids losing control of the g7 and h6 squares. There can follow 41.i.fT (41 f4 .i.b4leads to sharp play; Typesetter's note: 41 .i.d2! looks very strong) 41.. ..i.c5+ 42 ~hl (42.te3 is worse: 42...c3! 43 bca344e7 .txe3+45 :xe3 :dl+46~h2.txf7 47 e81f+.i.xe848 :xe8+ ~h7, with the better chances for Black) 42...:xfT! 43 ef:bS+ 44 :b2 :xh2+ 45 ~xh2 ~g7, with a probable draw. 39 .i.d2? i.b4! 40.i.el .txc3


41

be

a3 (46)

46 W

In this endgame, Black even has the better chances. 42 fe LdS 43 e6 :dl! .More exact than 43 ...a2 44 f4. 44 Ldl i.xdl

45

e7

White would lose at once with 45 :d2? i.a4 46 e7 ~g7 47 :d8 a2.

45

...

i.a4!

45 ....tb3 46 :e2 leads to a draw. 46 i.rn Now Black acquires a decisive plus. The right move was 46 f4!?, with these possible variations: (a) 46 ...:g8 47 f5 i.e8 48 :a2 (48 .txe8? a2) 48 ...i.xg6 49 fg :xg6+ 50 ~f2:e6 51 :xa3 :xe7 52 :xaS, with a drawn rook ending. (b) 46...:a7 47 :e2! (White loses after 47 e8'ii'+ i.xe8 48 i.xe8 :g7 49 i.g6 :Xg6! 50 :Xg6 a2) 47 ...:b7 48 e8'if+ .txe8 49 i.xe8 (but not 49 :Xe8+ ~g7 50 i.f5 a2 51 :el ci16 52 i.c2 :b2 53 i.e4 hS, etc.) 49 ...~g7 50 i.a4 :b2 51 :xb2 ab 52 i.c2, with a draw.

46

...

:e2

:b8

47 47 e8'if+ doesn't help: 47 ....txe8 48 i.xe8 :xe8 49:a2 :e3 50 :xa3:Xf3 51 :xaS :Xc3. 47

48cMl 49

i.xc4

50 51

~el ~xe2


52 ~b3 .i.d7 53 ~e3 f6 54 ~4 ~c6! After 54...~xe7? White escapes with 55 ~g5 ~e6 56 c4! 55 c4 ~xe7 56 c5 ~e8 0-1 White resigned in view of the threatened ....i.f7. Game No. 10 Khalifman-Karpov Reggio Emilia 1991192 This encounter effectively sums up the results of contemporary theory on the Zaitsev Variation. Although it ends in a loss for me (as a result of a blunder at the end), this game more than any other illustrates the abundant resources latent in Black's position. The play proceeds on classical lines; White concentrates all his forces on the kingside, Black dominates on the queenside - and achieves his object there. The theoretical value of the game is unquestionable.

1 e4 e5 2 ~f3 ~c6 3 .i.b5 a6 4 .i.a4 ~r6 5 0-0 .i.e7 6 :el b5 7.i.b3 d6 8 c3 0-0 9 h3 ~b7 10 d4 :e8 11 ~bd2.i.f8 12 a4 h6 13 ~c2 ed 14 cd ~b4 15 ~bl c5 16 d5 ~d7 17 :a3 rs 18 ~2 A new move; previously, as we know, practice had only seen 18 ef, 18 e5 and 18 :ae3. 18 ••. ~6 18... fe is risky; after 19 ~xe4 ~xd5 22 :g3 White has a dangerous attack. But 18 ...c4 and 18 ...~h8 deserve to be tried. 19 :f3 (47) 19 :ae3 would lead to familiar positions from my games with Kasparov - after 19...~h8 or 19.....d7 20 ef, etc. In the latter case, 20:n gives a position in Oil-Hjartarson, mentioned in the notes to Game No. 8 above. The transfer of the rook to f3, without going to e3 ftrst, is the actual point of White's innovation. All the same, we shall soon see that this line too presents no great danger to Black. 19 ... l:e5!? If 19 .. .fe, then 20 ~xe4 ~bxd5 21 "d3!? 20

:xrs


47 B

The other capture on fS - 20 ef lZxel + 23 "xel .i.xdS 23 lZg3 "e8 24 lZe3 .. gives a game with mutual chances.

n-

20

:us

21 et .i.xd5 22 ~ Another possibility is 24 ltlg4ltlxg4 2S hg (but not 2S "xg4 "gS I). 22 ••• .i.xe4 Other ways of exchanging off this knight turn out less well for Black: 22...ltlxe4? 23 .i.xe4 .i.xe4 24 lZxe4 dS 25 lZe6, or 22...ba 23 ltlxf6+ "xf6 24ltlg4 "d4 25 "e2. 23 .i.xe4 d5 24 .i.f3 Retreating the bishop along its customary diagonal would play into Black's hands: 24 .i.bl d4 25ltlg4 d3 26ltlxf6+ "xf6 27 .i.xd3? lZd8! (better 27 .i.d2, but after 27 ... c4! Black has excellent chances). 24 ••• c4! 25 lZe6(48) This rook does Black no particular harm, hence White should have continued 2Sltlg4ltlxg4 26 hg ltld3 27 :est? ltlxe5 28 "xdS+ "xdS 29 .i.xdS+ ~h7 30 .i.xa8 ltld3, or 26 .i.xg4 1013 27 lZe6 .i.cs with a roughly equal game. 25 ... ltld3 26.i.e3 d4! The most accurate move; the b-pawn will not run away. If 26...ltlxb2?!, then 27 "d4! is very strong. 27 .i.xh6 The only way to fuel the fire. 27 ltlxb2 28 ..a


48 B

A real hand-to-hand fight is in progress. White is aiming at the enemy king, and abandons his queenside to its fate. Incidentally, the queen also had other moves; 28 .cl?! lbd3 (but not 28 ...~a4?) 29 .gS ~ is not much good for White, but it was worth considering 28 .e2 c3 29 i.gS d3 30.eS d2 31 i.xf6 gf (31...dl.+ 32 i.xdl .xdl + 33 ~f1 c2) 32 :xf6 c2 33 .e6+, with a probable draw.

lS

•••

lOxa4

So White's last queenside pawn is eliminated, but the knight now needs to be brought back into the action. 29 i.gS d3

3O.d2

tOes!

After 30...c3, the white queen is activated: 31 .81+ ~h8 32 :xf6! gf33.n. 31 i.xf6 gf 32 The rook has to move away. Instead, 32 i.xa8 ~e6 33 fe .xa8 34 ~g4 .d8 gives Black the advantage. 32 ••• :eS The white rook can give Black distinct trouble, so it is worth exchanging it. On the other hand, 32... ~b3 33 .f4lM4 34 :c7 i.g7! was also playable.

:c6

33

lbeS

33 i.dS+ .xdS 34 :xc8 .xfS is bad for White. 33 ••• .xeS 34 i.d5+ ~b7 More precise than 34...~h8 35 .f4.d7 36 'lrh4+ 'lrh7 37.xf6+ .g7 38 'lrh4+ 'lrh6 39 .g4 (39 .d4+!? is interesting) 39...i.g7 40 f6, with chances for both sides.

35

.f4 (49)


35

d2!

It is essential to divert the white queen from the kingside. The alternative was more dangerous: 3S ....*.h6 36 Wg3! WfB 37 Wg6+ (37 ~g4 is not good: 37 ...d2 38 ~xf6+ ~h8 39 Wg6 dl W+ 40 ~h2 .*.f4+ 41 g3 .*.xg3+ 42 ~g2 Wxd5+, or 42 fg Wd2+ 43 .*.g2 Wdh6; or 42 ~xg3 Wgl+ 43 .*.g2 Wg7) 37 ...~h8 38 ~g4 d2 39 .*.f3. 36 Wg4 Editor's note: Khalifman describes this as a time-trouble mistake, recommending instead 36 Wxd2! WxfS 37 ~g4 with compensation, e.g. 37 ...~7 38 Wd4 .*.cS? 39 .*.e4!. 36 ... dlW+ 37 Wxdl WxfS 38 ~g4?! 38 ~1 was more stubborn. 38 •.. .*.h6?? An inexcusable mistake. Instead of ending the game victoriously at one stroke, I suddenly throw it away - in a single move. The annoying thing is that Black had conducted the fight almost impeccably up to this point. 38...Wd3! was immediately decisive: 39 WeI (39 ~xf6+ ~g6) 39...~g7 40 ~3 c3, and it is allover. 39 WeI! The e7 square has been left undefended, and White takes the opportunity to break into the enemy's rear with his queen. .*.f8 39 ... 39....*.g7 40 We7 etc. does not help.

40 41

We8

""1+

42

g3

i.d6+ Wg6

~h2

Closed Spanish: Zaitsev Variation 61 43

.d8

1-0 As we have seen, the result of the game doesn't entirely correspond to the outcome of the theoretical debate on this topical variation. Game No. 11 lIjartarson-Karpov

Candidates Quarter-Final, 5th game Seattle 1989 In my Candidates match against Hjartarson I decided to stick to defending the Spanish. In the first game, my opponent shied away from a fullblooded contest; he chose the Exchange Variation, and peace was concluded in another fifteen moves. By contrast, the third and fifth games took a highly interesting course and made a definite contribution to theory. In the notes to this, the final game of the match, I shall incorporate one of the games of the match with Kasparov that took place a year later, in which White introduced an important theoretical novelty. 1 e4 e5 2 00 tLlc6 3 i.b5 a6 4 i.a4M 5 0-0 i.e7 6 :el b5 7 i.b3 d6 8 c3 0-0 9 h3 l:te8 10 d4 i.b7 11 a4 In the third game of the match, the Icelandic grandmaster played the more restrained 11 tLlbd2 i.f8 12 a3 etc. This will be found in the notes to Game No. 13, which focuses on the a2-a3 system.

11

h6

12 tLlbd2 i.f8 13 i.c2 We had reached this position once before (Hjartarson-Karpov, Dubai OL 1986); on that occasion I played the quiet 13 ...l:tb8, and solved all my opening problems after 14 ab ab IS i.d3 i.c8 16 tLlfl i.d7 17 tLlg3 .c8 18 i.e31fb7 19 de tLlxe5 20 l:ta7 .c8 21 tLlxeS de. But White could have played more strongly, for instance with 18 i.d2 or 19 dS. ed 13 14 cd tLlb4 15 i.bl ba An alternative to the currently popular ...c7-cS. After IS ...cS 16 d5 tLld7 17 l:ta3, Black either plays ...cS-c4 (Games 4 and S) or undermines the centre with ...f7-fS (Games 6-10).


16:04 17

as

:.3(50)

50 B

I now played ...:as-a6, but let us first recall two continuations seen in earlier games: 17 ... g6 and 17.....d7. As illustrations I shall use games by Beliavsky, who became a 'casualty' of these opening battles. Sax-Beliavsky, Moscow IZ 1982, went 17 ...g6 18 l%ae3 (in the wellknown game Kasparov-Beliavsky from the same interzonal, White immediately went into action in the centre with 18 eSt?, and after 18 ...de 19 de eohS! 20 W .d5 21 ~f3 ':xe5! 22 eog4 ':xel + 23 .xel ~h7 24 ':e3 :d8 a complex position arose with chances for both sides; Black is a pawn up, but White has compensation for it. Here is how the game concluded: 25 .id2.d6 26 eogeS .id5?! 27 eoh4 ~g8 28 ~xg6! fg 29 .ixg6.in 30.ixhS .ixhS 31 ':g3+ ~n 32.e4 "xd2 33 .f5+ ~e7 34 ':e3+ .xe3 35 fe ':dl + 36 ~h2':d5 37 .c8 ~n 38 g4.id6+ 39 ~g2 .ig6 40 eoxg6 ~xg6 41 "g8+ 112-112. At move 26, an alternative is 26 ....ixf3) 18....ig7 19 eon c5 (19 ...d5 can be met by 20 e5 ~ 21 eog3!; but 19...~7 !?, taking control of the central squares, is interesting) 20 eog3 cd (again 20... ~7 deserves attention) 21 eoxd4 d5 22 e5 ~ 23 eoxe4 de 24 .ixe4 .ixe4 25 1lxe4 .d5 (Black's pawn sacrifice fails to guarantee equality) 26 ~ .xdl 27 ':xdl ':ac828 .id2 :ed8 29 ':al ':d5 30 .ic3 ~3 31:dl! a4 32:al ':b5 33 Ilexa4 eoxb2 34 ':a8 ':xa8 35 ':xa8+.ifS 36 e6! fe 37 ~4 1-0. Ehlvest-Beliavsky; USSR Ch, Lvov 1984, went 17....d7 18 ~4! (Black's waiting move with his queen, vacating the d8-h4 diagonal, has made this powerful knight excursion possible. In the game BalashovBeliavsky from the same championship, White played the less energetic 18 ':ae3, and after 18 ...a4 19 ~ d5 20 eS ~ 21 eold2 'iVb5! 22lDxe4 de 23 .ixe4 .ixe4 24 ':xe4 ~3 25 ':le2 c5! Black had

Closed Spanish: Zaitsev Variation 63 enough compensation for the pawn. One further game on these lines, A.Sokolov-Beliavsky, was played in the tournament; after 21 tb3d2 c5 22 tbxe4 de 23 .txe4 cd 24 :g3 White's edge was minimal, and this game, like the foregoing one, was soon drawn) 18..."'5 (18 ... g6 can be met by 19 :g3! .tg7 20 ttlxg6!? with an attack) 19:n tbh7 20 :g3 tbg5 21 tbf3 (21 tbf5!?, with numerous kingside threats, was even stronger) 21...ttlxf3+ 22 ttlxf3 (White's activity on the kingside is very dangerous) 22 ...1i'h5 23 .td2 :e7 24 .c 1 ~h8 25 dS! c5 26 :g4! .tc8 27 :f4 ~g8 28 eS g5 29 :f6 de 30.txb4 ab 31 .c2 ~g7 32 d6! ~xf6 33 de .txe7 34.e4 1-0.

17

...

:a6!?

This move was first played by me in a game with Balashov nearly ten years ago (50th USSR Ch, Moscow 1983). Its idea consists in prophylaxis - the black rook prepares in advance for the defence of either wing. For example, in the case of e4-e5 and an exchange on eS, it will be covering the weak points e6 and g6.

18

tbb2

The dispute about this variation was to be continued in the Candidates Semi-Final. Timman preferred the h4 square for this knight. In the first game he moved it to the edge of the board at once, and in the third he did so after 18 :ae3 a4. For more about this, see game No. 12, which focuses on 18 :ae3. In a game Sznapik-Szymczak, Poland 1989, White preferred 18 :c3, and after 18 ... g6 19 b3 .tg7 20 .tb2 c5? (20...dS 21 e5 tbd7 was correct, preparing...c7-c5) 21 dS tbd7 22 .cl a4 23 ba tbb6 24 :ce3 .txb2 25 .xb2 the dark-square weakness on the kingside soon made itself felt. On 18 dS, Black equalises with 18...e6 19 de .txe6 20 tbd4 .tb7 21 :ae3 g6 22 tbfl.tg7. The immediate 18 eS? does not work: 18 ...de 19 de tbd7 20 tbc4 .tdS 21 tbxaS .txf3 22 gf ttlxeS, and Black is better.

18

...

g6

The crushing defeat suffered by Black in a game Sax-Banas. Hungary 1984, is instructive: 18 ....a8? 19 :ae3 .a7 (rather an artificial queen manoeuvre) 20 eS! tbfd5 21 :g3 de 22 de:ae6 23 tbe4 ~h8 24 tbf31fb6 25 tbfg5! hg 26 tbxg5 g6 271i'h5+!!, and Black resigned in view of 27 ...gh 28 tbxfl mate. 19 tbg4 19 e5, as played in Horvath-Razuvaev, Sochi 1987, holds no danger for Black: 19...de 20 de tbh7 21 tbc4 .dS 22 .xd5 .txd5 23 :c3 tbf6 24 tbd2 tbd7 25 .te4 liz-liz.

64 Closed Spanish: Zaitsev Variation A month after the primary game, Ivanchuk played the new move 19 f4 against me. I now append this sharp skirmish, which ended with the exhaustion of forces ten moves later. Ivanchuk-Karpov, Linares 1989: 19 f4 d5! (better than 19...c5?! 20 dS J..g7 211Ohf3) 20 e5lOe4 211Og4 (accepting the pawn sacrifice is not good: 21lOxe4 de 22 J..xe4 J..xe4 23 l:xe4 c5. White also gains nothing from 21lOhn c5 22 l:ae3 c4!?, or 22 ...cd 231Oxd4 J..c5 24 102f3 f6!?) 21...c5 221Oxe4 de 23 dc (White loses with 231Of6+? l:xf6 24 ef cd, while 23 J..xe4? fails to 23 ...J..xe4 24 l:xe4 f5!. Equality is preserved after 23 dS J..xd5 241Of6+ l:xf6 25 ef c4 26 J..xe4! lOd3 27 J..xd3 l:xel + 28 .xel J..xa3 29 ba cd; in this last line, it is worth considering 23 ...lOxd5 24 J..xe4 c4!?) 23 ...J..xc5+ 24 J..e3 J..f8 251Of6+ l:xf6 26 .xd8 (26 ef? .xd127 l:xdllOdS 28 l:b31Oxe3 29 l:xe3 J..c5 30 ~f2 l:e6 is bad for White) 26 ... l:xd8 27 ef lOd3 28 l:dl J..xa3 29 ba J..dS 1/2-1f2. (For instance, 30 J..xd3 J..b3 31 l:bl l:xd3 32 ~f2 draws.) Kasparov prepared an exceptionally cunning novelty for our 1990 match. In game 512, he played 19 f3! (51). 51 B

This move contains a great deal of poison. Play continued 19...•d7 (my first reaction didn't prove a complete success. A more logical plan was the simple 19...J..g7 20 lOc4 c6 21 J..d2lOh5, with a pleasant game for Black. Less clear lines are 20...•a8 21 dS! 1Oh5 22 lOn, and 19...c5, to which the reply is not 20 dc? dS!, but 20 dS J..g7 21lOc4 lOd7 22 J..e3! and White has the advantage) 20 lOc4 'iVb5 21 l:c3 J..c8 22 J..e3 ~h7 23 .cl c6 (the weakening of the d-pawn quickly decides the outcome; 23 ...'iVb8 was more stubborn) 241Og41Og8? (52) (now White carries out a winning combination, though the exchange 24 ...J..xg4 is also in his favour: 25 hg'iVb8 26 g5!, but not 26 ~f2 dS!) 25 J..xh6! J..xh6 26lOxh61Oxh6 271Oxd6 'iVb6 28lOxe8 .xd4+ 29


52 W

~hl'ifd830l:dl'ifxe831'ifg5l:a732l:d8'ife633f4i.a634f5'ife7

35 'ifd2 'ife5 36 'if£2 'fIe7 37 'ifd41llg8 38 e5 00 39 fg+ fg 40 l:xc6 'fIxd8 41 'fIxa7+ lllde7 42 l:xa6 'fIdl+ 43 'fIgl 'fId244 'fin 1-0. So that game did not end as successfully for Black as the primary game we are analysing. Perhaps Black will have a chance to get his own back in tournaments to come. But in the 1990 match I was never to revert to the move ...b5xa4. 19 ••• /llxg4 20 'ifxg4 20 hg i.g7 211lln i.c8 22 g5 h5leads to double-edged play. 20 ... cSt? An innovation, prepared for this match. A weaker line is 20 ....lg7?! 21 tDf3 c5 22 dS c4 23 l:dl l:a8 24 h4 .lc8?! (24 ...h5 is more precise) 25 'ifg3, and White has substantial attacking chances; A.lvanovKIovans, Kuldiga 1987. 21 de I once happened to watch a well-known player make the stock move d4-dS here, missing the counter-stroke 21...i.xdS. 21 ... de 22 eS!? (53) It is hard to say whether Hjartarson was expecting to reverse the course of the match, but in any case this is the only way to try for the initiative. Once White allows ....lfS-g7, the e4-e5 break will never happen. The position looks dangerous for Black, but I succeed in finding what is virtually a forced method of simplification. 22 ... 'fId4! 22...hS has been suggested, but it weakens all the kingside dark squares. After 23 'fIg3, the sortie 23 ...'fId4 no longer has any point; while if 23 ...h4, the h-pawn is straying a long way from its comrades. In


playing ...•d8-d4, it was of course essential to foresee all the consequences of the tactical skirmish that now ensues.

23 .g3 :ae6 24 :ae3 Black wins after 24 ~b3? 25 ~xaS :xe5 26 :XeS :xe5 21 ~xb1 :el+ 28 ci>h2 :XcI. 24 ••• c4! 25 j,r5 lOcI3 The rook cannot leave e6, on account of the thematic eS-e6. Playable moves were 2S ... j,g1 and 25 ...j,c5, giving up the exchange for a certain amount of compensation. But my intention was to give up not the exchange but my queen. 26 j,xd3 Of course I had calculated the variation 26 j,xe6 :xe6 21 ~b311fb6 28 :Xd3 cd 29 .xd3 a4! 30 ~4 j,c5!. A game with mutual chances also results from 26 ~b3 21 j,xe6 :xe6. 26 cd 27 hd3 (54)

.dS

.dS

54 B

Closed Spanish: Zaitsev Variation 67 27 ... :XeS! This positional queen sacrifice settles the outcome of the match. The queen has, incidentally, no good squares to which to move. On 27 .....c5, White has 28 tbb3, while 27 ...1i'b6 is very strongly met by 28 tbc4, and Black's compensation for the pawn is insufficient.

28 29

:Xd4 ~h2

:Xel+ :Xci

In this open position, Black's bishop pair gives us reason to assess his game as wholly comfortable. If White succeeded in exchanging his rook for a black one, the power of the bishops would be reduced, but this exchange is not possible. 30 00 White's knight covers all his weaknesses, and Black needs to play with some accuracy; for example after 30....ixf3 31 "xf3 ltcS, his fortress is not as secure as in the actual game. Despite this, White's best move was 30 tbb3. 30 ••• ltc5! Thanks to this elegant manoeuvre the black pieces co-operate properly, while the white queen cannot activate itself. Black incidentally threatens 31...l:tfS, and it is now White who may run into difficulties. 31 ltd7 .ixf3 This exchange is now fully appropriate. After 31....tc6 32 ltd2, the black pieces could lose their co-ordination (the threat is ti:ld4).

32"xf3

:rs (55)

55 W

At this point White offered a draw and I accepted with pleasure, seeing that this ended the Candidates Match in my favour. I may add that in other circumstances I would have asked my opponent to play a move, and would only then have made a decision. After 33 "g3 .ig7!

68 Closed Spanish: Zaitsev Variation (alternatively 33 ...i.c5 34 :d2 i.a7 35 'iVa3, or 33 ...:e2 34 f4 i.g7 35 b3 :xf4 36 'iVxf4 i.xe5 37 :xt7 with a draw) 34 f4 i.xb2, Black has the initiative. The correct move is 33 'iVd1 (the f-pawn is not to be saved anyway) 33 ...:Xf2 34 :d2, exchanging Black's active rook. As I found out after the game, Hjartarson was intending to leave that rook on the board and exchange the other one with 33 'iVd3 :xf2 34 :dS :xdS 35 'iVxdS. In that case Black could still have had a try at winning with 35...:xb2 36 'iVxaS i.d6+ 37 ~gl i.g3, although by playing 3S 'iVdS+ and 39 'iVd3 White could eliminate the dangers. Game No. 12

Timman-Karpov Candidates Final, 5th game Kuala Lumpur 1990

:e1

1 e4 eS 2lLlO lLlc6 3 i.bS a64 i.a4lLlf6 S 0-0 i.e7 6 bS 7 i.b3 d6 8 c3 0-0 9 h3 10 d4 i.b711 a4 h6 12lLlbd2 i.f8 13 i.c2 ed 14 cd lLlb4 15 i.bl ba 16 llxa4 as 17 h6 All this had occurred in my Candidates Quarter-Final match with Hjartarson. In the last game of that match, White tried lSlLlh2 g6 19 lLlg4. The debate initiated by White's lSth move was later continued in the World Championship match itself (19 f3!). Some lively contributions were also made in between those two contests - in the Candidates Final, where Timman was keen on the idea of ffi-h4.

:e8

:&3

18

he3

In the first match game, Timman played lSlLlh4?! at once, underestimating the counter-stroke Is ...lLlxe4!. After 19lLlxe4 (19 i.xe4?! d5 20 i.b1 :xe1+ 21 'iVxe1 'iVxh4) 19...i.xe4 20 i.xe4 d5! (56), Black had an obvious advantage. Let us follow the game to the end.

Closed Spanish: Zaitsev Variation 69 21 :ae3 l:lae6 (21...de!?22lbe4 :xe4 23 :xe4 "d5! was stronger) 22 .lg6!? 'ibh4! 23 lhe6 :xe6 24 :xe6 fe 25 .le3 "f6 26 "g4 (Black now obtains a decisive plus. White had to play 26 .lbl I?, with the aim of penetrating to e8 with his queen via a4 or h5. If then 26...e5?, White has 27 de "xe5 28 "g4!. If instead 26 ...c5 27 dce5 28 .td2~6 29 .txa5! lLlxa5 30 "xd5+ ~h8 31.la2 ~h7 32 .lbl+, White maintains equal chances; 32... g6 fails to 33 .te4! with the threat ofb2-b4. At move 27 of this line, Black fails to improve with 27 ...lLlc6 28 "c2!, or with 27 .....xb228.td4 "a3 29"el! ~f7 30.lg6+! ~xg6 31 "e6+, drawing) 26....td6 27 h4lLlc6! 28 .te8lLle7 29 .ld7lLli5 30 h5 ~ 31 .lc8 ~e7 32 b3 c5! 33 "e2 cd 34 .ld2 d3! 35 "dl'ifh4 36 g3lLlxg3! 0-1.

18

••.

a4

19 lLlh4 (57) In the present context, this move is to be taken quite seriously. In the original game with 17 ...:a6 (Balashov-Karpov, 50th USSR Ch, Moscow 1983), play went 19lLln d5 20e5lLle4 21lLlld2c5 22 lLlxe4 de 23 .txe4 .txe4 24 lhe4 c4!?, and Black had enough compensation for the pawn. As it happens, White returned it at once with 25 e6! :exe626 :xe6 fe 27 lLle5 "e7, retaining the initiative. Perhaps 27 .....d5 was more accurate; in A.Sokolov-Bronstein, Reykjavik 1990, the continuation was 28'ifh5 :a7 29 "g6lLld3 30 lLlxd3 cd 31 'it'xd3 :d7, with approximate equality. A correspondence game Maeder-Haag (1984) went 19lLlh2 g6 20 lLlg4 c5 (exchanging on g4 was worth considering) 21lLlxf6+ "xf6 22 dc dc 23 e5! "e6 24lLle4!, and White retained the better chances.

19 ••• c5?! Here the blow against e4 does not work: 19...lLlxe4? 20 lLlxe4 .txe4


(20 ...'it'xh4 21 ~f6+) 21 .i.xe4 'it'xh4 22 .i.h7+. In addition to this, 19...~d5? fails to 20 ed ':xe3 21 fe'it'xh4 22 ~f3 'it'g3 23 e4. 20 de 20 d5? is met by 20...~xdS 21 ed ':xe3 22 fe'it'xh4. 20 ••• de The counter-stroke 20...d5!? is also interesting. 21 ll)fs If21 eS ~d5 22':e4a3! 23 ba~324 'it'g4~xb12S ~xb1.i.xe4, the advantage is with Black. 21 ••• .ic8 A solid move, but a playable alternative was 21...g6!? 22 ':g3 ~h8 23 eS ~fd5 (23 ...~S) 24 ~3 ~f4, with chances for both sides. 22 e5 ll)fdS 23':g3 ~4 24 'it'a .iDS! The only move; 24...~d3 is no good in view of 2S 'it'e4!, when 2S ...~e1 fails to 26 ~7+. 2S .ixfS tBe6 26 .i.bl 26 .ixe6 ':axe6 27 ~ would give equality. 26 ••• ~4 27 'it'g4?! 27 'iVb7? loses to 27 ... ':e7, but White could improve with 27 'it'c3 'iVdS 28.ie4 :XeS! 29 .ixdS (29 ':ge3 'it'd6 30 .ih7+ ~xh7 31 ':xeS ~bc2leads to equality) 29...':xe1+ 30 ~h2 ~xd5, or 27 'it'd1!? with an unclear position. Now Black seizes the initiative. 27 ~bc2! 28 'it'e4 (58) 58 B


28 .*.xc2lDxc2 29 .c4lDxel 30.xa6 :xe5 31 .xa4li.)d3 32 M :e2 was no better for White. 28 5! 29.d3 ~el 30 .xa6 lbeS 31 :e3 31 .xh61! (threatening 32 .*.a2+) is dangerous for White on account of3l....d5! (31...lOe2+ 32 ~h2lDxg3 33 .*.a2+:d5 34 .e6+), and if 32 :g51 or 321i'h4? Black decides the game with the striking 32...•xg2+! 33 lbg2lOef3+ 34 tbxf3 tbxf3+ 35 ~hl :el+ 36 :gl :xgl mate. Another bad line is 31 .c4+ ~h7 32 .*.a2lOe2+ 33 ~h2 .*.d6, but after 31.*.a2+ ~h8 32.*.c4 f4 33 :g4 f3 there would still be a tough fight ahead. 31 •.• .15 32 ~! The only move; 32 .f1loses to 32... lOe2+ 33 lbe2 lbe2 etc. 32 lbe3 33 fe .xe3 34 .c4+ (59) White can't manage to win one of the pawns back: 34 .xa41 .f4+ 35~xelM+.

59 B

34 ••• ~h7?! A thorough analysis of the position reveals that 34...~h8! was much stronger. Let us look at the variations arising from it: 35 .*.a2 (35 M tbec2!) 35 ...•d3+ 36 ~t2 (36 ~xel? tbc2+! 37 ~ .e3+ 38 ~f1 .el mate) 36...•xc4 37 .*.xc4 tbec2 38 tbbl .*.d6 39 tbc3 tbb3! 40 ~e2 (the point is that with the king on h8, 40 .*.e3 loses to 4O...f4!, whereas with the king on h7 White could reply 41 .*.d3+) 40... a3!


(40...~xcl+ is inadequate; after 41 ~d2! White picks up the knight and then the a-pawn as well, with drawing chances) 41 .txb3 (41 ba ~xcl+ 42 ~d2 ~xa3) 41...~d4+ 42 ~dl ab! and wins. 35 .tal Weak alternatives are 35 ~ ~2! 36 ~f6+ ~g6, and 35 ~f3 ~2!.

35 ••• lOd3?! A time-trouble blunder. Nor are things entirely clear after 35 .....e8!? 36 .tbl (36 ~ fe 37 "g8+ ~g6 38 ~xel e3 is wholly bad for White) 36....td6 37 ~f2, preparing g2-g4. But Black could have retained winning chances with 35 .....d3+ 36 ~f2 "xc4 37 .txc4 ~2 38 ~bl ~b3 39 .te3 .td6 40 ~c3 ~bd4. (Alas, ...f5-f4 fails again to the bishop check.) ~g6 36 "g8+ 37 "t7+ ~h7 38 .g8+ 112-1h.

Game No. 13 lijartarson-Karpov Candidates Quarter-Final. 3rd game Seattle 1989 1 e4 e5 2·~f3 c!Dc6 3 .tb5 a6 4 .ta4 ~f6 5 0-0 .te7 6 :el b5 7.tb3 d6 8 c3 0-0 9 h3 .tb7 10 d4 11 ~bd2 .tt'S 12 a3 (60)

:e8

This restrained move of the a-pawn is a popular alternative to 12 a4. White protects the b4 square from a sortie by the black knight, and

Closed Spanish: Zaitsev Variation 73 subsequently aims to work up an initiative on the queenside. Another way to sidestep the most critical continuations is to retreat the bishop with 12 .tc2. A.Sokolov-Karpov, Bugojno 1986, continued 12... iDb8 12 a4, whereupon I played the innovation 12... cS which didn't turn out too well. White sealed the queenside with 14 dSiDbd7 IS b4 c4 16iDn, then conducted an energetic attack on the kingside: 16... lObs 17iD3h2 g618 .te3 .te7 19 Wd2:t'8 20 .th6iDg7 21iDg3 eRh8 22 iDg4llli6 23iDxf6+ .txf6 24:n Wd7 25 f4 as 26 fS ab 27 cb ba 28 eRg8 29 Wf2 .th4 30 .txg7 j,xg3 31 :xg3 eRxg7 32f6+ etc. Instead of 13 ...cS, the right move was 13 ...iDbd7, postponing the central break. This was played a few rounds later in Ljubojevic-Portisch, which quickly ended in a draw: 13...iDbd7 14 j,d3 dS!? (the customary 14...c6 is also good) IS ab de 16iDxe4iDxe4 17 j,xe4 .txe4 18 :xe4 ab 19 j,gS f6 20 .th4 :xal 21 Wxal j,d6 22 Wa2+ eRh8 23 de iDxe5 liz-liz. After 12 j,c2iDb8 (12...g6 is also frequently played), in addition to the sharp 13 a4 White has another quiet pawn move: 13 b3. Here is one example from my own play, which shows that in this case too Black has nothing to fear: Ljubojevic-Karpov, Tilburg 1988: 13 ...iDbd7 14 .tb2 g61S a4 :b8 16 j,d3 c6 17 Wc2lObs 18 lOb2 (after 18 c4 ed 19iDxd4 b4 Black seizes the initiative, but an improvement is 18 j,niDf4 19 g3iDe6 20 h4) 18 ...llli4 19 .tniDe6 20 iDg4 .tg7 21 de (White achieves nothing with 21 00 h5 22 iDe3 ed 23 cd cS 24 dS iOd4 25 iDxd4 cd 26iOdl :c8, and Black has the advantage. In this line, if 24 dc, then 24... j,xe4!) 21...de 22 00 h5 23iDe3 Wc7 24 :adl :bd8 2S h4 112_112. I would remind you that the Zaitsev Variation has the purpose of preventing the standard iOd2-n followed by the transfer of the knight to its customary square g3. After 12iDn ed 13 cd, as I have said before (see Game No.1), Black has the good move 13 ... i0a5, exchanging the lightsquared 'Spanish' bishop. On the other hand, taking the e-pawn is dangerous: 13 ...:xe4 14 j,xf7+, or 13 ...iDxe4 14 :xe4! :xe4 ISiDgS :e7 16Wh5. 12 ••• h6 (61) The immediate 12...iDb8 fails to 13 de de 14iDgS :e7 ISiDxf7!? :xf7 16iDf3 Wxd119 :xdl cS 20 j,e6!, as in Kuporosov-Zhukhovitsky, USSR 1986. In Hjartarson-Short, Belfort 1988, Black played 12...g6 and White achieved a beautiful win: 13 j,a2 j,g7 14 b4 h6 IS j,b2 lObs 16 dS iDe7 17iDb3:t'8 18 c4 be 19i0a5 j,c8 20 :cl fS (2o ...iDf4 21lOb2 fS 22 f3 gS was more to the point) 21 ef iDf4 22 :xc4 j,xfS 23iDxe5!

:f3

74 Closed Spanish: Zaitsev Variation i.xh3 241Ob7! 'ifcs 2S ~d6 cd 26 llxcS i.xcs 27"d2 g5 2s1Oc6 lOfS 29 i.bllOh4 30 lle7W 31 i.xg7 ~g7 32 "c3lOhfS 33 g4! lOxe7 34 ~e7+ wn 35 i.g6+ Wxe7 36 "xg7+ 1-0. 6/ W

13 14

i.c2

lObS

b4

A combination of a2-a3 and b2-b3 hardly holds any danger for Black, as we can see from Sax-Short, Saint]ohn 1988: 14 b31Obd7 15 dS (with the closing of the centre, we reach a familiar position from the Smyslov System, in which White has wasted a tempo on a2-a3) 15 ...c6 16 c4 'ifc7 17 a4 llec8 18:a2 ba 19 ba as 20 i.a3 i.a6 21lOh2 g6 22 lOhfl cd 23 cd hS, and the chances are equal. At move 15, a more interesting line is 15 i.b2 g6 16 a4 i.g7 17 i.d3 c6 18 'ifc2 "c7 (better 18...:C8) 19 b4 (19 c4! is stronger: 19...ed 20 cb ab 21 ab llxal 22 i.xal with advantage to White - Ivanchuk) 19...dS! with chances for both sides; Anand-Ivanchuk, Novi Sad 1990. 14 lObd7 15 i.b2(62) 62 B

Closed Spanish: Zaitsev Variation 75 For this match I had specially prepared the move IS ...aS!? But first I wish to give two of my games with the more common IS ... g6. In the process I shall quote various other games that have played an important role in the theory of the variation. Tirnrnan-Karpov. Bugojno 1986. went: 15... g6 After IS ...c5 16 be ed 17 cd de. the game Nuon-Shvidler, Thessaloniki OL 1988. saw the new move 18 l:cl (instead of the more popular 18 .i..bl or 18 dS). The continuation 18...l:c8 19 .i..bl c!tlhs 20 eSllli4 21 .c2? gave Black the initiative after 21 ...g6 22 ~ cd 23 .d2 l:xc1 24 :Xci ~6. In the later game Nuon-Greenfeld, Groningen 1988. Nuon improved by 21 ~ cd 22 l:xc8 .xc8 23 .xd4 with a minimal edge for White. 16 c4 ed 17 cb ab 18liJxd4 c619 a4 An improvement was introduced in a game between two Hungarian women players: 19 .i..d3! .i..g7 20 l:cl ~ 21 ~3 ~ 22.i..n l:ad8 23 104f3lOxf3+ (it was better to bolster the centre with 23 ...lOfd7) 24 'i'xf3 l:e6 25 lOb3 l:de8 26 .i..d4 'i'a6 27 .c3. and Black's position is very difficult; Madl-VerOci, Hungary 1987. 19... ba 20 ha4.b6 This move was first played by Kasparov, in the first game of his match with Timman (Hilversum 1985). After 21 bS cb 22 .i..xbS dS! 23 l:xa8 .i..xa8 24 .a4lOcS 25 .c2 l:b8, Black seized the initiative and went on to win. In the third game of the match. Tirnrnan varied with 21 lLIc2. He repeated this move against me. 211L1c2 .c722 .i..b3 l:xa1 In the third Timrnan-Kasparov game, Black chose 22....i..a6. which led to equality after 23 l:cl .i..g7 24lOe3 .i..bS - although in the subsequent play. which was not free from errors (White too came to the brink of defeat). Tirnrnan achieved the win. 23 hal After 23 .xal .i..g7 24lOc4. an equal game results from 24 ...dS, but an interesting line is 24 ...cS 2S (better 25 eS, with equality) 2S ...•xa5 26 ba lLIxe4 27 .i..xg7 ~xg7 28 f3 dS 29 fe dc 30.i..xc4 l:xe4, and the initiative is with Black. 23... .i..g7 24lOe3 c5 This defuses the situation. I was strongly tempted to take the pawn with 24...lOxe4. Indeed. after 2SlOxe4 l:xe4 26 .i..xf7+ ~xf127 .f3+ lLIf6 28 .i..xf6 l:xe3! 29 .xe3 .i..xf6 30 .xh6 cS. Black has the advantage. However, with 2S lLIg4! White could obtain a very dangerous

.as

76 Closed Spanish: Zaitsev Variation attack, for example: 25 ...lOef6 (25 ...dS 26 i.xg7 ~xg7 27lbxe4 de 28 'it'd4+ lbe5 29 f4) 26lhe8+ lbxe8 27 i.xg7 ~xg7 28 lOe4 h5 29 i.xfl! ~xn 30 lbg5+ ~e7 31 'it'f3! lOef6 32 lbxf6lbxf6 33 'it'e3+ and wins. 2S bc lhxcs 26 i.xf6 i.xf6 M-M And now to the second game I mentioned: Hjartarson-Karpov, Linares 1989. It is notable that this game was played a month after our Candidates Match, and this time I refrained from the experiment 15 ... a5. Play proceeded:

IS... g616 :hI We have already examined 16 c4 in detail. But apart from 16 :cl, White has also played 16 'it'bl in order to over-protect the pawn on e4. After 16...i.g7 17lbb3 c6 18lbaS 'it'c7 19 i.b3! :ad8 20 'it'a2 dS 21 ed lDxd5 22 de lbxe5 23 lbxe5 ':xe5 24 c4, White has a substantial plus. A more accurate line is 17 ...:c8 18lbaS i.a8 19 dSlbb6 20 a4 'it'd7 21 ab ab 22 i.d3lDb5! 23 c4 be 24 lbxc4 lDf4 25lDxb6 cb 26 i.cl, with approximate equality; Psakhis-Portisch, Sarajevo 1986. The manoeuvre ...lbf6-h5 was successful against 22 i.d3, but in reply to 22 i.cl it is less good: 22 ...lbh5 23 i.e3lbf4 24:a3 ~h8 25 'it'dl g5 26 i.xf4 gf 27 lDh4 :g8 28lDfS, and Black is in trouble; Short-Hjartarson, Tilburg 1988. The correct reply is 22 ...c6! 23 dc i.xc6 24 i.e3

lbc4. 16... c6 An innovation; Black solidly fortifies the centre. Timman-Portisch, Amsterdam 1989, went 16...:b817 :cl i.a818i.bl :c819c4ed20 cb ab 21lbxd4 c6 22lbnlbe5?! (in Timman's view, 24 ...'it'b6, preparing ...c7-c5, was better) 25 lOe3 lDb5 26 :n! 'it'g5 27 lOe2lbf4 28 lbxf4 'it'xf4 29 g3 'it'f3 30 lbg4!, and White obtained a very dangerous attack although in the end the game was drawn. It is interesting that this game was played in the other Candidates Quarter-Final, which took place at the same time as the Hjartarson-Karpov match, the source of the 'primary' game we are examining.

17li:Jb3 :C818 de de 19 c4 cS! The game is now equal. 20 lhxcslhxcs 21 'it'xd8 After 21 be 'it'xd122 :bxdl :xc5, Black has the advantage. 2l ... :exd822 bc :XcS 23 LeSlbd7 24 i.d4 Or 24 :bdl :xc4! 25 i.b3 :Cc8.

24... :Xc4 2S i.b3 :c726lDh4? He had to play 26 a4 lDc5 27 i.xc5 lhc5 28 ab ab 29 e5, or 26 e5


i.xf3 27 gf J.c5 28 J.b2 M 29 e6 fe 30 J.xe6+. In both cases Black's game would be a little more pleasant, but now he quickly achieves his aim. 26...t/)c5! 27 J.e5 :Cc8 28 J.f61:d3 29 J.c2 fld2 30 l:b2 &i)d7! 31 i.b11:d632 e5ltb6 33 a4 &fuj6 34 elb4 35 J.a2 '4xf6 361:d1 :C337 &i)j3 J.c6 0-1 15 ..• as 16 i.d3 Exchanging pawns on as first, and only then bringing the bishop to d3, was evidently more precise. It was also worth considering 16 de de 17 &i)b3 ab 18 cb c5 19 be (19 &i)aS? cb 20 &i)xb7 'iVb6) 19...&i)xc5 20 'ifxd8 l%exd8 21 &/)xc5 i.xc5 22:abl l%ac8 23 l%bel b4 24 a4 i.d4?25 &i)xd4 ed 26 e5 &i)b5 27 J.f5! with a plus for White; Shabalov-Klovans, USSR 1989. At move 24, Black should play 24 ...J.b6! 25 J.xe5 &/)xe4 with equality. 16 ••. c6 17 &i)b3 The only line in which White can count on a minimal plus is 17 'iVb 1 'iVb8 18 ba l%xaS 19 c4 be 20 &i)xc4, as in A.Rodriguez-Rubinetti, Toluca IZ 1982. 17 ••. ab 18 cb 18 ab &i)b6 19 &i)aS 'ifc7 gives equality. 18 ed 19 &i)fxd4 (63) 63 B

Now Black seizes the initiative. He also has a good game after 19 i.xd4 c5 20 J.xf6 &i)xf6 21 J.xb5 l%xe4. The correct course for White is 19 &i)bxd4 c5 20 be de 21 &/)xb5 &i)xe4 22 &i)e5 &i)xe5 23 J.xe4, with a roughly equal game.


19 ... cS! A highly promising pawn sacrifice. Black takes aim at the enemy centre. 20 be 20 c!Llxb5 at once was more precise: 20 ... cb 21 ab .zhal 22 i.xal d5 23 ed (23 e5? i.xb4) 23 ... .:.xel+ 24 'it'xel c!Llxd5 25 'it'e4 c!Ll7f6 26 i.xf6 c!Llxf6 27 'it'xb7 'it'xd3 with sharp play, in which Black has sufficient compensation for the pawn. 20 de 21 c!LlxbS c!Llxe4! 22 'it'c2?! A major inaccuracy. He had to play 22 'it'n, giving rise to the following variations: (a) 22 ... 'iVb6 23 i.xe4 i.xe4 24 Itxe4 'it'xb5 25 a4, with unclear play. (b) 22 ... c!Lld6? loses to 23 ':'xe8 'it'xe8 24 c!Llxd6 i.xd6 25 'it'xb7 ':'b8 26 'it'dS. (c) 22 ... .:.b8 23 i.xe4?! (better 23 i.c4, with a complicated position) 23 ... i.xe4 24 ':'xe4 ':'xe4 25 'it'xe4 ':'xb5 26 'it'dS 'iVb8 27 'it'xd7 ':'xb3 28 i.cl ':'bl 29 ':'xbl 'it'xbl 30 'it'd2 c4 31 ~h2 'iVb3, and Black is slightly better. (d) 22 ... c!Lldf6!? 23 i.xf6 c!Llxf6 24 ':'xe8 (24 'it'xb7 'it'xd3) 24 ... i.xf3 25 ':'xd8 ':'xd8 with equality. 22 ... c!Lldf6 23 c!Llc3 (64) Not 23 ':'adl 'iVb6 24 c!Llc3 c!Llxf2!. 64 B

23 24

c!LlgS! i.bS

Closed Spanish: Zaitsev Variation 79 24 to:I2 was more tenacious, although Black then has 24 ...'iFb6, threatening 25 .....c6. 24 lb::el+

2S

'fie?

lb::el

26 .i.n But now, 26lDd2 can be met by 26.....f4!. Still, 26 :e3 was worth considering. 26 ••• 'fIc6! Black threatens the deadly ...00+ or ...lOxh3+. 27 lte3 .i.d6

28 29

h4 iOdl?

lOe6

29 :xe61fe30 'fIg6"e8 was also bad; he had to play 291Ob5!1 .i.f4 30 :b3lOe4, when White can still offer resistance. I now conduct a decisive attack.

IOg4

29 30 31

lb::e6 ci?hl

32

f3

.i.h2+! 'fIxe6

'fie!!! 0-1

33 fg is decisively met by 33.....xh4, and 33 33...:a4.

"c4

by 33....i.f4 or

2 Closed Spanish: 9 ...ltJd7

Game No. 14 Kasparov-Karpov World Championship Match game 5/12 New YorkILyons 1990 I e4 eS 2/Cf3 c!lk6 3 .i.bS a6 4 .i.a41Ct'6 5 0-0 .i.e7 6 :el bS 7.i.b3 d68c30-09b3 1Cd7 (65) 9 65 W

This plan, involving the fortification of e5, was originally introduced into practice by Chigorin. I would remind you that in my last match with Kasparov, the Spanish Game occurred eight times. In half of these encounters, the dispute centred on 9 ... .i.b7 (see Games 1-12), and in the other half on 9...ted7. Two of the games with the latter move are accorded primary status in this book (Games 14 and 15); the up-todate material on this line is assembled in the notes. 10 d4 .i.f6 In this way Black prevents the white knight from developing on d2. Another quite popular system involves 1O... /Cb6, though of late it has been seen more rarely. After U/Cbd2 .i.f6 12 d5/Cas 13 .i.c2 c6 14 dc 'fIc7 15/Cn 'fIxc6 161Ce3, White has a small but secure advantage. In place of the old move 13 ... c6, Spassky chose 13 ... g6 in a game against

Closed Spanish: 9.JtJd7 81 BeJiavsky (Barcelona 1989). The reply was 14 ~f1, and a complex struggle ensued in which White had the initiative. But an even stronger line is 14 a4! ba 15 .i.xa4 ~a416 "ii'xa4 ~b7 17 "iVc6.:a7 18 b4 .i.d7 19"iVc4 .i.b5 20"iVb3 £g7 21 c4 £d7 22 "iVe3, intending c4-c5 (Beliavsky). A game Ljubojevil:-Spassky from the same tournament varied with 12 lllil 13 .i.c2 ed, and the players soon concluded peace: 14 cd ~b4 15 £bl c5 16 a3 (a novelty, but the familiar 16 ~3 looks stronger) 16...ttlc6 17 e5 de 18 de "iVxdl 19 :Xdl ttla4 20 £e4 £b7 21 .i.e3 .:ad8 22 b4 1/2- 112. In Fedorowicz-Razuvaev, New York 1989, White was better after 13 ...g614 ~3 £g7 15 d5 ~7 16 b3:f8 17 a4 ba 18 baf5 (18 ...aS was more precise) 19 as ~7 20efgf21 £a3~h8 22 ~g5! ~f6 23 c4 £h6 24 ~6 .txe6 25 de e4 26 .i.a4 l:b8 27 c5! d5 28 c6! lllig8 29 l:bl. White's pressure eventually led to a win. 11 84 11 d5 and 11 a3 are unpromising. White also gains nothing from 11 .i.d5; in Sax-Karpov, Rotterdam 1989, the game was completely equal after 11 ....tb7 12 de de 13 £e3 ~a5 14 £xb7 ~xb7 15 "iVc2 "iVe7. The traditional break on the flank looks more thematic than 11 £e3, though the two moves often transpose. 11 ... £b7 I have also sometimes played 11...:b8, a somewhat passive but solid move. The position after 12 ab ab 13 £e3 ~7 (66) has arisen in three of my recent games.

:e8

66 W

Nunn-Karpov, Rotterdam 1989 continued 14 d5 ttlc5! 15 £c2 c6 16 b4 ~6 17 dc ~xc6 18 ~ fiX7, with equal chances. Sax-Karpov, SkellefteA 1989 went 14 ~g5 h6 15 £xf7+:xn 16 ~e6 "iVe817 ~c7 "iVd818 ~611z-11z.

82 Closed Spanish: 9... CDd7 Kir.Georgiev-Karpov, Reggio Emilia 1990 went 14ltJbd2 .i.b7 15 .i.c2 :as 16 "e2 c6 17 de de IS b4 "c7 19ltJb3ltJg6, with equality. If Black plays ll ...ltJas at once, then instead of the standard 12 .tc2, a very strong reply seems to be 12 .i.a2!, and if 12....tb7 (12...ItJb6 13 b41tJac414 ltJd7 15 dS!), then 13 dS .te7141tJbd2 c61S b4lDc416 ItJxc4 be 17 .i.xc4 and Black stands badly. In Geller-Krogius, Bad WlSrishofen 1991, Black played instead 12...cS, and White gained a big advantage with 13 dc dc 14 .i.dS :bS 15 ab ab 16.te3.

as

12 ltJa3 In our last match Kasparov searched for various ways to develop an initiative in this system. 1\vice - in this game and the ISth - he developed his knight on a3, and twice he preferred other methods: a pawn exchange on bS (6th game), or the move 12 .i.e3 (Sth game). The last two continuations will be examined in the context of Game No. 15 in this book.

12

•••

eel

This exchange was recommended by de Firmian. In a game de Firmian-Benjamin, USA 19S5, White had a clear advantage after 12...'iVbS 13 .i.gS! ed 14 .i.xf6ltJxf6 15 cd. White would also benefit ltJxc4 16.i.xc4. from 12 ...b4 13lbc4 be 14 be ItJb6 15 The interesting novelty 12...ltJe7 was used by Korchnoi in his Candidates Match against Sax (Wijk aan Zee 1991). Black defends himself without surrendering the centre. Play continued: 13 :bl (13 dS is worth considering, since the knight on e7 is constricted and Black cannot conveniently carry out ...c7-(6) 13...c6 14 .i.e3 "c7 15 ItJgS ba?! (1S ...h6 is solid-the threat was 16'iVhS-andifthen 16 de ItJxeS 17 1tJf3 :adS, Black's position is perfectly defensible; but not 16...de 17 ltJe6!) 16 .i.xa4 dS 17 .i.c2 h6 IS 00 ed 19 cd de 20.i.xe4 00 21 .i.d2. In view of the passive placing of the bishop on b7, White has an obvious plus (21...cS is dangerous on account of 22 :Cl, when 22 ...'iVbS or 22 .....dS would be met by the unpleasant 23 'iVb3).

as

13

cd (67)

ltJas!?

This knight manoeuvre was suggested by Podgaets, one of my seconds. White is slightly better after 13...lteS 14 .i.f4ltJas 15 .i.c2 b4 16 ItJbl cS 17ltJbd2, although it is worth considering 16... b3!? 17 .i.d3 cS ISItJbd2 cd 19 .txd6 'iVb6 20 eS .i.e7 21ltJe4 :adS! with complex play; Ernst-Tisdall, Gausdall991. Gutman recommends 14 ab ab 15 .i.a2 (not 15 "d3?! ItJxd4! 16 ltJxd4 llXS 17 "xbS .i.xd4, and already the advantage is with Black; Todorovi~-Smagin, Vienna 1991) lS ... b4 16lDc4, or IS ...ItJb4 16 .i.bl, with a small plus for White.

Closed Spanish: 9.JDd7 83

The next time the diagram position arose, I decided to nonplus my opponent with the novelty ...fod7-b6. But 'leave well alone' would have been good advice; I came up against a powerful prepared line myself. We will now digress to examine the 18th game of the match. Kasparov-Karpov, game S118: 13... lDb6?! 14 J.f4! More precise than 14 ab ab IS J.f4, when Black has IS ... b4 16 ~2 lDaS 17 lDxb4lDxb3, or 16~lba117 .xallDxc418J.xc4lDxd4, with a good game.

14... ba Not 14 ...lDaS? IS ab ab 16 eS J.e7 17lDxbS, and Black has no compensation for the pawn.

15 b 4 !iJxa4 The immediate IS ... aS was more exact.

16.xa4a5 Not 16...:e8? 17 dSlDeS 18lDxeS J.xe5 19 J.xeSIbe5 20 ~ and lDaS. However, Zaitsev suggested the rather ugly-looking 16...•e8!?, threatening 17 ...lDxd4. 17 J.d2 :e8 Gutman recommends 17 ...dS! 18 eS J.e7, with a sturdy position for Black.

18dS! Fixing the weak pawn on c7 and taking control of the c6 point. 18... lDb4 19 Lb4 ab 20.xb4:b8 (68) We had had this position on the board in our preparation for the match, and considered that Black has enough for the pawn. The obvious-looking 21 .d2 is met by 21...c6; the position opens up, and the bishop pair promises ample chances. 2l.c4!

84 Closed Spanish: 9...lcJ7

An unexpected and very strong move, which essentially decides the game. White returns the pawn, but after 21.. ..*.xb2 22 l:a2.*.xa3 23 J:xa3, the c-pawn is tied to Black's second rank and it is hard for Black to free himself. There is no improvement in 22....*.f6 (22...Wf6 23 Wxc7 l:ecS 24 Wd7 Wc3 25 lbb5 Wb3 26 l:a7) 23lbb5 Wd7 24 l:a7, and Black cannot extricate himself: 24 ....*.xd5? 25 ed l:xb5 26 l:xeS+ WxeS 27 Wxb5! and wins. 21...c6 does not work either: 22 de J:cS 23lbd4 (or 23 cb l:xc4 24 lbxc4 Wc7 25 l:aS l:bS 26lbaS). HWhite's queen manoeuvre to c4 had been postponed by one move, with 21 lbb5 Wd7 22 Wc4, Black could have played 22....*.xb2, with the continuation 23 l:abl .i.a6 (but not 23 ....i.xd5? 24 ed l:xel+ 25 l:xel l:xb5 26 Wxc7! and wins) 24 l:xb2 .i.xb5 and 25 ...c5, equalising. 21 ...Wc8221:iJd4! But not 22 l:abl c6, and Black has everything in order. 22... .LW Or 22 ....*.xd4 23 Wxd4 c5 24 dc Wxc6 25 f3, and the b- pawn confidently advances. 2311c3c5 After 23 ...Wb7 24lbxc2 Wb6 25 b3, White has a sound extra pawn. 24 de .Ld4 25 Wxd4 Wxc6 26 W! - and despite stubborn resistance, Black eventually had to concede defeat. 14 .*.a2 After winning a pawn with 14 ab ab 15 ~xb5lbxb3 16 l:xaS WxaS 17 Wxb3 .*.xe4 lslbxc7 Wc6 19lbb5 l:bS, White would be facing disaster. 14 ... b4

Closed Spanish: 9.JDd7 85 15 ~4 lbxc4 16 j,xc4 l:te8 Podgaets' idea (and that of the whole variation!) consists in just this - the elimination of the white central pawn. After 16...dS 17j,xdS! j,xdS 18 ed 1Ob6, Black fails to equalise: 19 j,f41OxdS 20j,e5!.

"'3

17 Preventing ...d6-dS, whereas after 17 .d3 Black can play 17 ...dS 18 ed l:bel+ 19IOxellOb6. If instead 17 j,d3 cS 18 dSIOeS, Black has good prospects. l:txe4 17 ... 18 j,xt7+ ~h8 IS...~fB is no good on account of 19 j,e3 .e7 20j,dS j,xdS 21 .xdS l:teS 22 as!, besides which White has the very strong 19 j,gS!. 19 j,e3 . 19 j,d2 deserved consideration, but not 19 .xb4? l:txel+ 20 lO~el l:tbS 21 .d2 .e7 and Black already has the initiative. 19 ... l:te7 20 j,dS c6?! Quite frankly, I underestimated the position of the bishop on e6. Unclear play would result from 20...j,xdS 21 .xdS as 22 l:tacllOb6 23 1Vbs .gS 24 l:tc61OdS 25 l:la6 l:lxa6 26 .xa6 gS, with ...h7-hS to follow. 21 j,e6 IOf8 22 j,g4 as 23 l:lad IOg6 24 j,h5 l:lc8 l:lb8 25 j,g4 26.c2 l:lc7 27.rs 1Oe7

28

.d3? (69)

After this inaccuracy I gradually succeed in equalising. Of course, 2S .xas?? loses to 2s ...1OdS 29 j,d2:as 30 j,xb4l:baS 31 j,xaS .as 32 j,xc7 IOxc7. But White could have retained some advantage with 2S Wh5! g6 29 1Vh6 (Salov gives 29 j,gS gh 30 j,xf6+ ~gS 31 j,e6+ ~fB 32 lOgS, which may appear very strong, yet Black has 32...j,cS! 331Oxh7+ ~eS) 29...j,cS (29....fB is no better: 30.xf8+ l:lxfB 31 j,f4, with unpleasant pressure against the pawn on d6) 30 j,gS lOgS 31 1Vh4 j,xg4 32 hg.

28 29

IOdS j,dl

cS!

86 Closed Spanish: 9...riJd7 69 B

30

i.e6

31

de

tOb6

31 b3 can be met by 31...i.xf3!, and already White has to think about equalising.

31

...

de

"'xdS+ White could have tried to keep the initiative by refraining from the immediate queen exchange: 32 "'fS!? i.xf3 33 i.f4, and if 33 ... i.c6, then 34 i.xc7 "'xc7 3S "'xcS i.xb2 36 "'xc6 "'xc6 37 :x.c6 fua4 38 i.c8 g6 39 :le8+ Wg7 40 :lc7+Wf641 :lxh7 with a clear plus. However, Black has the more accurate 33 ...i.b7 34 l:tedl "'e7 3S i.xc7 "'xc7 36 :lxcS "'e7 37 l:txaS i.xb2, when his chances are no worse. 32 lhdS 33 i.r4 l:te7 34 tOgS i.dS! (70) Black would have the better ending after 34...l:tde8 3S 00+ Wg8 36 tOh6+ Wf8 37 i.d6 gh, but 36 tOgS+ leads to a draw. 32

70 W

Closed Spanish: 9... ~7 87

35 J.xdS There is no danger to Black in 35 ~+ 1lxf7 36 J.xf7 J.xf7 37 J.c7 lOxa4 3S J.xdS J.xdS 391lcdl J.f6 40 1ld7 J.g6. 35 IlxdS J.xe7 36 he7 37 Ilel 1h.-1h. The final position appears to be more pleasant for Black. Let me give a few variations: 37...J.fS! (37 ...lbcs 3S 1le6, with the threat of Ilc6) 3S lleS ~gS! 39lbe6 (39lOxh7? ~7 40 1lxf8 c4!, or 39 IlbSibd7 40 IldSlOb6 41 1lb8, and now Black has either 4l...ibd7 with a draw, or 41...lbxa4 42lbe6 ~ 43lOxfSlOxb2 with dangerous threats for the piece) 39...~f7 40 1lxf8+ ~xe6 41 J.c7, and now 41...lOxa4 42 b3! lbc3 (or 42...lOb2 43 J.xaS c4 44 J.xb4 cb) 43 J.xaS Ildl+ 44 ~h2 Ilbl 45 1lc8 ~dS 46 J.b6 Ilxb3 47 IlxcS+ ~d6 4S 1lc4 00 49 J.cS+ ~e6. The b-pawn can give White a certain amount of trouble. But I decided not to risk anything, and signed the peace agreement. Game No. 15 Kasparov-Karpov

World Championship Match game 5/6 New YorkILyons 1990 1 e4 e5 2 00 lbc:6 3 J.b5 86 4 J.84lOf6 5 0-0 J.e7 6 Ilel b5 7 J.b3 d6 8 c3 0-0 9 h3lOci7 10 d4 J.f6 11 a4 J.b7 I have said that this position arose four times in this match, and that Kasparov played 12iba3 twice (see Game No. 14), 12 ab once and 12 J.e3 once. Before giving detailed attention to the exchange on bS, let us examine the following encounter in which Kasparov tried to improve on White's play. Kasparov-Karpov, game 5/8: 12 J.e3lba513 J.c2lbc414 J.cl d5! (71) In this familiar position, where Black had usually chosen 14...c6 or 14...ed, I introduce an important novelty.

15 de White does no better with 15 ed J.xdS 16 b3 J.xf3 17 .xf3ibd6. or 15 b3ibd6. 15... lbdxe5 An even stronger reply is IS ...de! 16 ef (16 J.xe4 J.xe4 17 Ilxe4 lbdxeS. and already Black is better) 16...ef 17 fg :eS, when the kingside attack gives enough compensation for the pawn. True, White can

88 Closed Spanish: 9.J&J7

71 W

improve with 17 b3 ~e5, and now 18fg:e819t'bbd2. However that may be, the next time I employed the 9 ...t'bd7 system (Game No. 14 in this book), Kasparov refrained from both 12 ab and 12 .te3, and chose 12t'ba3. 16lhxe5lhxe517 ab ab At this point, 17 ...de is worse: 18 'iVxd8 :fxd8 19 ba t'bd3 20.txd3 ed 21 a7! followed by .te3, with advantage; though after 19...ha6 20 ha6 .txa6 21.txe4t'bd3 22 .txd3 .txd3 23lbd2, White's position is only slightly better. 18 l:xa8 'iVxa8? (72) The right recapture was 18....txa8, and if then 19 f4t'bd7! 20 e5 .th4, Black brings his knight via c5 to e4 with a fine game. 72 W

19/41iJg6 The point is that Black can no longer play 19...t'bd7 20 e5 .th4? (20....te7 21 f5), on account of 21 'iVh5, exploiting the absence of the queen from d8. 20 e5 .th4 21,:/1 .te722!ild2 .tc5+?

Closed Spanish: 9.JiJd7 89 Analysis has shown that after 22...d4 Black would obtain ample play. Gutman's recommendation is also interesting: 22 ... b4 23lLlf3 be 24 bef5.

23 t:J6h2 d4 Not 23 ....i.e3 24 lLlf3 .i.xf4+ 25 .i.xf4lLlxf4 26 .i.xh7+ t:J6xh7 27 lLlg5+ etc.

U.e2! Double-edged play would have resulted from 24 lLlb2 .i.xg2 25 lLlxc5 .i.xfl 26 .xfl, but now White is better. The further course of the game is extremely tense, and after some errors by Kasparov, Black actually obtains a won ending. Nonetheless the eventual outcome is peace on move 84. The game lasted over ten hours! For the sake of completeness I will give the rest of the game with brief notes. U ... dc 25 bc :d8 26lLle4 .i.a3 By exchanging one of the bishops, I manage to reduce White's attacking potential. 27.txa3 Le4 28 .xe4 .xa3 29 f5lLle7 30 .h4 f6! A good defensive move, shutting the white 'Spanish' bishop out of play. On 31 ef, Black has 31....d6+ and 32...•xf6.

31 .g3 t:J6f8! Again parrying the threat of 32 ef, in view of 32 ... gf 33 .xc7 ':c8 and 34 ....:xc3.

32 t:J6h1.c5! 33 e/gf34.i.b3lLlti5 35.h4t:J6g7 36.:ttl1 c637.:ttl4? A mistake; the right move was 37 ':d3, not only threatening 38 ':g3+ but also defending the pawn on c3.

37...•xc3 38 ':g4+ t:J6h8 39 i.xd5 .a1 + 40 t:J6h2 .e5+ The game was adjourned here. By now it is only Black who has winning chances. 41 ':g3 cd 42 .g4 .c7 43 .d4 .d644 t:J6h1 ':e845 .g4 .d746

.:ttl3 ':e1 + 47 t:J6h2 ':e4 48 .g3 ':e5 49.:ta3 ':e8 50 ./4 .b7 51 t:J6h1 .b852 .h4 .b6 53 .b4 (73) 53... d4? It looks as if Black had a win here with 53 ...•12!!. After 54 .xb5? ':el+ 55 t:J6h2 .f4+ 56 ':g3 ':e3 it is allover. Nor is 54 t:J6h2? any better in view of 54 ... ':g8. White's only try is 54 ':al ':g8 55 ':gl, but after 55 ...•e2 56 .d6 he will hardly be able to save himself.

.e5.cl

54 f1g3 .c7 55 .:ttl3 ./4+ 59 t:J6g1 ':c8

+ 56 t:J6h2 ./4+ 57 t:J6g1

.cl

+ 58 t:J6h2

de Firmian's recommendation would have preserved winning chances: 59...t:J6g7! 60 ':xd4 ':el+ 61 .xel .xd4+ 62 q.,hl b4 63

90 Closed Spanish: 9... &iJd7

We7+ Wh6, although even here the odds would be in favour of a drawn outcome. 60 '4d1 '4d861 Wxb5 We3+ 62 Wh1 d3 63 Wa5 Wd4 64 Wa1 Wb6 65Wa2 Wg7 66Wd2 Wc5 67'4/1 '4d4 68:p Wd6 69 '4e3 '4a4 70 '4e1 h5 71 '4b1Wd7 72 Wd1 Wh6 73 Wd2+ ~g7 74 We3 h4 75 W/3 ~h6 76 We3+ Wg7 77 W/3 d2 78 'ilh5 Wp 79 'ilxj7+ Wxj7 80 '4d1 '4d481 Wg1 '4d5 82 cttJ2llif5+ 83 ~e2 '4g5 84 Wp. ~-~. Now, back to the fifth match game:

12 13

ab lha8

ab

Wxa8

13...J.xa8 is worse, for example: 14 dSIOe7 15 ~a3lt)c5 16 J.c2 c6 17 b4 ~a618 dc .*.xc619 "d3!, and now White has a won position aftereither 19...'fIb8 20~4! ed21 e5 ~g622ef, as in Tal-Tone, Bugojno 1984, or 19...~720'4dlWa821'i1xd6~622'i1d3~g623:el! 'fIb7 24 g3 h5 25 h4 'ilc8 26 ~2:d8 27 W£3 J.e8 28 J.b3! as in Hubner-Short, Skelleftell989. 14 cIS In Hjartarson-Short, Tilburg 1988, White played 14 Wd3?, and after 14...edl 15 Wxb5 (15 cd ~xd4! 16 ~xd4 ~5 17 Wxb5 .*.xd4 is winning for Black) 15... ~5 16.*.dS :b8 Black emerged with advantage. There is likewise no danger for him in 14 ~a3 b4 15 ~ be 16 be ~ 17 ~xa5 "xaS, Tal-Karpov, SkellefteA 1989. 14 •.• lDaS 14...1Oe7 15 ~a3 J.a6 has also been seen, with somewhat the better game for White. I had prepared a different plan. IS.i.c2 li.)c4! (74) The knight excursion to the active post c4, prepared specially for this game, is more accurate than the moves seen earlier, 15 ....*.e7 and 15...:b8. For example, 15 ...'4b8 16 ~a3 i.a6 17 b4 ~ 18 ~c4 be

Closed Spanish: 9.JiJd7 91 19 -*.a4 t2)b6 20 -*.c6 -*.b7 21 -*.xb7 'it'xb7 22.i.e3 with a clear plus for White; Tal-Keres, Tallinn 1964. 74 W

16 b3 After 16 t2)bd2 and the exchange of knights, White cannot count on anything either. 16 ... t2)cb6 17 Q)a3 .i.a6! The position of the bishop on a6 looks rather ugly, but then the knight on a3 is also out of it for the present. Incidentally, 17 ... c6!? was also playable. 18 t2)h2 c6 19 de 'it'xe6 20.i.d2 .i.e7 The initiative is already with Black. It is true that I later played inaccurately and, as you will see, was forced to work for the draw in the ending. But the opening battle has gone Black's way, which is why Kasparov changed his plan for the eighth game. All the same, in that game too, as we already know, I coped successfully with the problems of the opening. 21 t2)g4 :as The stock move 21...dS is premature here, if only because of 22 b4. But deploying the rook in the centre was quite a good idea: 21...:e8 22 t2)e3 t2)f6 23 t2)fs -*.£8 24 .i.g5 :e6. Another interesting line is 21...fS 22 ef .i.b7, with ...d6-dS in view, but after 23 'it'f3 the situation is not so clear. t2)f6 22 t2)e3 23 t2)fs .i.f8 t2)bd7 (75) 24 .i.gS

92 Closed Spanish: 9... ~7 24..."xc3? would lose to 25 .txf6 gf 26 lle3 ""2 27 Ilg3+ ~h8 28 ~h2!, and 29 "g4 cannot be stopped. However, at this point 24 ...d5 was perfectly playable; after 25 ed lDbxdS 26 lDxb5 .txb5 27 c4 the game would have drawish features, for example: 27 ...lDb4 28 .txf6 lDxc2 29 cb "xf6 (29 .....c3? 30 lDh6+! ~h8 31lDxf7+ ~g8 32 ~6+ ~h8 33 "d5) 30 "xc2 Ilb8. 75 W

2S c4!? By sacrificing a pawn, White brings his light-squared bishop which has played a mediocre role so far - into the attack. 2S ••• bc 26 bc .txc4?! With 26.....c5! 27 lle3 (27 lDb5 dS! 28 ed "xc4, or 27 d5) 27 ... h6 28 .txf6lDxf6, Black could still have kept the initiative (Gutman). The modest 26... h6 is also good. Now the situation changes abruptly. 27 lDxc4 "xc4 28 .tb3 29 ~h2 h6 Another possibility is 29 ... g6 30 ':e3 "al 31 "d2 gf 32 .txf6 f4 33 llel"a5 with unclear play (dePirmian), or31 gf32"xf5.tg7 33 ':g3 ~f8 34 .txf6lDxf6 35 ':xg7 ~xg7 36 "g5+ ~f8 37 "xf6 "a7 38 "xd6+ "e7 39 'it'h6+ ~g8 with approximate equality (Wolff). However, with the correct order of moves - 30 .txf6! lDxf6 (30... gf? 31 "dS), and only now 31 ':e3 "al 32 "d2 - White would have a dangerous initiative (Gutman). lDxf6 30 .txf6 31 ':e3 "c7 ~h7 32 lW

"f3

"c3

"f3

Closed Spanish: 9... !iJd7 93 Now, after the knight exchange, the white pieces begin to dominate the board in earnest. Black should have preferred 32... 'iVdS 33 :g3 (33 1De3 :a7) 33.. .'~hS 34 j.xnlDxe4. 33 li)e3! Incidentally threatening 34 :xf6. 33 'iVe7 34

~

35 36

j.xdS ""3

lbxdS l%a7

In the endgame after 36 :xt7 'iVxt7 37 j.xn :xt7, Black's fortress is impregnable.

r6

36 37

""8

g6 (76)

76 W

At this point White could have hemmed in the black king with 3S g4!. The threat is 1lf3-c3-cS, and if ... j.f8-g7, then l%hS+! and mates. The continuation 3S ...:d7 39 :c3 :dS 40 'iVb6 j.g7 (40 ...hS 41 l%c7 :d7 42 :cS Jog7 43 'iVbs :a7 44 gh gh 4S 'iVb3!) 41 :c7 :d7 42 :cS etc. is scarcely attractive for Black. On the other hand if 38 ...hS (which works well in the actual game), then 39 gS! is decisive. 38 :c3? This allows Black to organise his defence. 38 ••• h5! 39 g4 39 :cS j.g7 40 :hS+ j.xhS 41 'iVgS+ ~h6 42 'iVxhS+ 'iVh7 is not dangerous either. 39 ~h6! 40 gh ~xh5

94 Closed Spanish: 9...&&17

:c8

41 The game was adjourned here.

J.g7 (77)

The game was not resumed. In his adjournment analysis Kasparov must have come to the conclusion that he had not sealed the best move and that White no longer had winning chances, so he proposed peace. We too had analysed the position thoroughly; my second Ron Henley put in a special effort. We reached a draw in all variations, though not without difficulty. Here are some interesting lines from our analysis which Henley afterwards published. After the sealed move 42 :e8, Black indeed has no problems: 42 ....d7 43 :d8 .c7 44 .xc7 :xc7 45 :xd6 with a draw, or 43 J.c6 .xc6 44 .xa7 .xe4 45 .xg7 .f4+ 46 ~g2 .g5+ 47 ~f1 .cl+ 48 ~e2 .c2+49~f3 .f5+! (but not 49 ....c6+? 50~g3 .xe851 'iVh7+ ~g5 52 'iVh4+ ~f5 53 .g4 mate) 50 ~g3 .g5+ 51 ~f3 .f5+ 52 ~e3 .f4+ 53 ~d3 .d4+ with perpetual check. We spent a good deal of time on the moves 42 :c3, 42 J.e6, 4211b6 and 42 :cl, any of which might have been sealed. Let us look at the main variations: (a) 42 :c3 f5 43 h4 ~h6 (but not 43 ... ~xh4? 44 :h3+ ~g5 45 :g3+ ~f6 46 .g8 fe 47 'iVh7 g5 48 .xe4 .e8 49 .e3 'iVh5+ 50 :b3 .dl 51 :13+) 44 :g3 (44 ef e4) 44...f4 45 :g4 :d7 and draws; (b) 42 J.e6 1Ib7! (the draw is much harder to achieve after 42 ....xe6 43 .xa7 .xc8 44 .xg7); (c) 4211b6 f5 43 .e3 f4 44 .f3+ ~h6 45 :cl :c7 46 :gl 'iVh4, with a draw; (d) 42 :cl f5 43 :gl :a3!, and now White has various possibilities:

Closed Spanish: 9... ~7 95

(dl) 44 ef :xh3+ 45 ~xh3 Wh4+ 46 ~g2 "g4+ 47 ~1 ~g2 "g4+ with perpetual check;

"dl+ 48

(d2) 44 ~1 Wh4 45 "dl+ ci>h6 46 "cl+ "f4+ and draws; (d3) 44 .tb3 d5!? 45 ed e4, with adequate counterplay; (d4) 44 ~:f345 ef:Xf2+46~hl :f447 .i.f3+ci>h648:Xg6+ ~h7 49 ~3 e4 50 .tg4 e3 51 "d3 ~7+ 52 ci>h2 ~2+ 53 .te2 :f2+ 54 :g2 .te5+ 55 ~gl (but not 55 ~hl? "c1+ 56 :gl :b2 mate) 55 .....cl+ 56.tfl e2 57 :xf2 el" 58 f6+ ~h6 59 fT "g5+ 60 :g2 "ee3+ 61 ~hl ! "xd3 62 f8"+ .tg7 63 :xg5 .txf8 64 .txd3 ~xg5, again drawing.

3 Open Spanish Game No. 16 Short-Beliavsky Barcelona 1989 The Open Variation was one of the most frequently occUlTing lines in the title matches in Baguio and Merano, and in the latter it can be said to have played a decisive rOle. In my duels with Kasparov, however, the variation was never played at all. It has arisen rarely in my games over the past ten years, and the general interest in it has somewhat declined. Nonetheless there are some grandmasters, especially Yusupov and Korchnoi, who never give up their pet system under any circumstances and frequently employ it as before.

1 e4 e5 2ibBibc6 3 .tb5 a64 .ta4ibf6 5 0-0 5 ibxe4 6 d4 b5 7

.tb3 d5 de .too ibbd2(78) 9 .te3 and 9 c3 will be examined in detail later. As for the old move 9 "e2, it has gone out of fashion. Thirty years ago, a long forced variation became known to theory: 9 'ii'e2 .te7 10 ':dllbcs 11 .txdS .txdS 12lbc3 .tc4 13 ':xd8+ ':xd8 14 'ii'e3 b4 IS b3 .te6 16ibe4 ':dl+ 17 iDel ibd4 18 .tb2ibxc2 19 'ii'e2 ':xal 20 .txalibxal 21 ibxcs .txcS 22 ibd3 .tb6 23 ibxb4 0-0 24 lbc6 f6 25 h4 fe 26 "xeS ':f6. The known continuations here were 27 g4 and 27 lild4; more recently, a third has been added. Timman-Yusupov, Montpellier 1985, went 27ibd8 .to 28ibxf7 Wx029 'ii'xal ':xf2 30 Wh2 as 31 'ii'e5 h6 32 a4 g6 33 'ii'dS+ Wg7 34 'ii'eS+ Wf1 3S hS ':fS! with equality. So the verdict on the forced line of play from opening to endgame is the same after all those years: Black's position is safe. Ljubojevic prefers 9 a4, but cannot boast of any special achievements with it, for example: 9 ...b4 10 aSlbcs 11 .tgS 'ii'd7 12ibbd2 h6

8 9

Open Spanish 97 13 .1h4 .1e7 14 J.xe711xe7 (or 14... ltlxe7 15ltld4 0-016 c3 be 17 bc l:lab8 18 J.c2 J.g4 1911el :b2 201le3 J.f5! and Black has everything in order; Ljubojevic-Yusupov, Linares 1991) 15 c3 be 16 be ltlxb3! 17 ltlxb3 0-0 18 :el :ab8 19 ltlfd4ltla7! 20 ltle2? (he should have exchanged on e6) 20...c5 21 ltlf4 :fd8 22 lIc2 ltlc6, with advantage to Black; Ljubojevic-Hjartarson, Amsterdam 1991. 78 B

9

..•

lLJcs

The discussion of 9 ....1g4 will be left until a little later. 10 c3 d4 (79) The most popular continuation. The alternatives 1O...J.e7 and 10....1g4 will also be discussed below. 79 W

Before we go further (in the game White exchanged on e6), something should be said about 11 ltlg5!? This striking manoeuvre was thought up by Igor Zaitsev, my second, in preparation for the Baguio match. The knight coolly puts itself en prise to the black queen - not the sort of thing that occurs to everyone! However, Korchnoi didn't

98 Open Spanish decide to take the piece. Let us recall that fascinating game - and also what happened afterwards. Karpov-Korchnoi, 10th game, World Ch Match, Baguio 1975: l1...dc 121t1xe6 fe 13 be "d3 141t1£3 (after 14 i.c2 "xc3 Is1i'hs+ g6 16 i.xg6+ hg 17 "xhS "xa1 White has no compensation for the sacrificed material, but 14 "g4 deserves attention) 14.....xd1 15 i.xd1 i.e7 16 i.e3lLld3 17 i.b3 ~ IS :adllLldxes 191t1xesltlxes 20 i.f4 ltlc4 (not 20...i.d6 21 i.xeS i.xes 22 :fe1 i.xc3 23 :xe6 Wf8 24 :d7! etc.). I now exchanged with 21 i.xc4 be, and the chances soon levelled out. Tal recommended 21 :d7 cs 22 :el :adS 23 :b7 :d3 24 i.gS :eS 25 :e4 :xc3 26 h3, and White still has dangerous threats. But instead of 23 ...:d3 Black has the stronger 23 ...:dS!, after which he can look to the future with confidence. But the question is, could Black have taken the knight after all? On l1.. ...xgS 12 "£3, White has a very powerful attack after either 12...i.d7 13 i.xf7+ We7 141t1b3 (another strong line is 14 i.dSltlxeS 15 "e2 d3 16 "el c6 17 f4'iht6 IS i.£3!, with a large plus; WolffFlear, London 1990) 14.....xeS 15 ltlxcs "xcs 16 :el+ WdS 17 cd "d6 IS i.gs+ WcS 19 :acl!, or 12... Wd7 13 i.dS! i.xdS 14 "xdS+ i.d61s cdltlxd416ltlc4! ltle2+ 17 ~hl "fslSltlxd6 "d319"xf7+ ~c6 20 i.e3 21 "e7! "dS 22 :adl lLld3 23 e6, BlllndumBrinck-Claussen, Denmark 1979. However, a year after the Baguio match, in a game Timman-Smyslov (West Germany 1979), Black answered 12 with 12...0-0-0. There followed 13 i.xe6+ fe 14 "xc6 "xeS 15 b4 "dS 16 "xdS ed 17 be dc IS1t1b3 d4 19 i.a3 i.e7 20 i.b4 i.f6 21 a4 ~d7 22 ab ab 23 :a6 c6 24 :dl ~e6 25 :Xc6+ WdS 26 :xf6! Wc4!, and in spite of White's two extra pieces the chances may be rated about equal. So acceptance of the sacrifice appeared to have stood up to the test. But later, Lilienthal introduced a significant refmement: 23 :fd 1 ! ~e6 24 :acl ~ 25 ~f1 :IleS 26 :d3:e4 27 g3. Black's centre pawns have been stopped, and his position is very difficult. Still, perhaps theory has yet to say its last word on this knight sacrifice. 11 i.xe6 ltlxe6

:af8

"f3

12 cd 12 a4 and 121t1b3 have also occurred, but without achieving anything for White. lDcxd4 (80) 12 ...

Open Spanish 99

"f3

This position is one of the standard points of departure for contemporary theory on the Open Variation. If now 13 ~xd4 "xd4 14 l:d8 15 a4, we reach a position from the famous game CapablancaLasker, St Petersburg 1914. In the Merano match, I employed two dangerous innovations in succession: 13 ~e4 (games 14 and 16) and 13 a4 (game 18). It was the knight move to e4 that later became more popular. But let us look first at the consequences of a2-a4.

13 a4J.e7 We had also examined other ways of bringing the bishop out - to c5 or b4. In Nunn-Timman, Amsterdam 1985, Black played 13 ...l:b8, and after 14abab 15~e4J.e716~6+cd 17~xd4~xd418"xd4de 19 "xeS 0-020 J.f4 l:b7 21 "e4 l:d7 22 'iVc6 a draw was agreed. All the same, 13 ...J.e7 seems to me the most logical. 14~4

It is also worth considering 14 ab, although in Sax-Yusupov, Sofia 1984, the chances were equal after 14...~xb5 15 "c2 0-0 16 ~b3 c5 17 J.e3 'iVb6 18 l:a4l:fb8 19 ~a5 l:c8 20 "e4 ~bd4 21 ~xd4 cd 22 J.xd4. In Hiibner-Yusupov, Tilburg 1987, White played instead 15 ~,but again failed to achieve anything: 15 ...0-016 J.e3 c5 17 l:a4 'iVb8 18 "al ~bd4 19 J.xd4 cd 20 lLlxd4 ~xd4 21 l:xd4 "xe5 22 l:fdl l:fd8 23 b3 l:xd4 24 "xd4 "xd4 112-112. 14.....xd4 The correct recapture. In the 18th game in Merano, Korchnoi took on d4 with the knight: 14...~xd4 15 ~ b616 J.e3 0-017 f4 "xdl 18 l:fxdl l:tb8 19 l:d7. White has a significant plus, which soon became a decisive one. This win was my sixth in the match, and therefore concluded the Merano contest. Later, several games were played in which Black castled a move earlier: 15 ...0-0. After 16 ab ~xb5 17 J.e3 'iVc8 18 "c2 "e6 19 f4 f6 20 ef J.xf6 21 f5"eS 22 J.c5, Black still

100 Open Spanish had some problems in Ivanchuk-Yusupov, Linares 1989. In his game against Adams at Hastings 1989190, Yusupov improved with 19...llad8 20 lla4 lld7 21 llfal "d5 22 h3 f6 23 ef 1.xf6 24lDxf6+ llxf6 25 llxa6 llxa6 26 llxa6l2Jd4. But in the same Hastings tournament, Yusupov in turn had to face an important innovation from Chandler: 18"d5! lld8? (the correct reply, in Yusupov's view, was 18.....f5! 19lDg3 "g6, and only then ...llad8, with chances for both sides) 19 "c6 "f5 (now this queen excursion is less successful; better 19...l2Jd4 20 1.xd4 llxd4 21 f4 as) 20 f4. Black has a difficult position, and this time Yusupov failed to hold it.

15 ab"xe516 ba 0-0 17&!Jf3"b51S"a4"xa419l:txa4lDc5 The chances are now equal. This was demonstrated in the game Psakhis-Dolmatov, played in the 49th USSR Championship (1981) shortly after the Merano match. Eight years later (!), in Ehlvest-Marin, Tallinn 1989, White innovated with the immediate 17 "a4! llfb818 a7 llb7 19lDo "d5 20 1.e31.c5 21 lladl "'322 "xb3 llxb3 231.xc5 lDxc5 24 l2Jd4, with advantage. 13 l2Je4 1.e7 Black plays this move almost automatically, but not everything is clear in the case of 13.....d5 14l2Jxd4lDxd4. The position after 15l2Jc3 "d7 161.e31.c5 (S1) occurred twice in the 1981 USSR Championship.

Romanishin-Yusupovcontinued 17'iVh5 (17"d2lld818lladl 0-0) 17.....e6 18 lladl lld8 19 lld2 0-020 llfdl g6 21 'iVh41.e7 22"e4 lDf5. Beliavsky-Dorfman went 17l2Je41.a7 18 llcl 0-0 19lDc5 .i.xc5 20 llxc5 llfd8 21 .i.xd4 "xd4 22 "xd4 llxd4 23 llxc7l:td2 24 llfcl lle8 25 ll7c2 llxc2 26 :Xc2 f6.

Open Spanish 101

In both cases, Black had safely overcome his opening difficulties. Instead of IS to:3, perhaps IS l%el is more precise, for example: IS ...J.b4 (1S ...J.e7 16 ~f6+!) 16 to:3 .d8 (after 16...•d7 17 J.e3 J.cS, White has an extra tempo compared with diagram 81, though this is hardly of vital importance) 17 J.gS! .xgS 18 .xd4 J.xc3 19 .xc3 0-020 :acl, Malchikov-Odaev, USSR 1983. The weakness of Black's queenside pawns clearly makes itself felt. 14 J.e3 lMs In the 14th game in Merano, where this position arose for the first time, Korchnoi wrongly exchanged knights; after 14...lDxf3+ IS 0-016 :fdl 17lDf6+! J.xf6 (17 ...gf 18 ef J.d6 19 :d4 ~h8 20 :h4:g8 21 :xh7+~xh7 221ih5 mate) 18ef.c819 fg:d8 20h4!, White had a clear plus. The retreat to fS occurred in the 16th match game.

.xf3

.e8

IS.c2 16 :adl

0·0

In the 16th game in Merano, I played 16 ~gS J.xgS 17 ~gS g618 lDxe6 fe 19 :ael .d5 20 b3 :ac8 21 J.cS :fd8 22 h3 .c6 23 b4 :d7 24 :dl :cd8, and this time Black surmounted his opening difficulties. There is little promise for White in 16lDf6+ J.xf6 17 .xfS J.e7 18 :adl.c8 19 ~2 :d8; Van der Wiel-Korchnoi, Sarajevo 1984. 16 ... c!bxe3 17 fe The doubled isolated pawns have their good points - they control the important squares d4, d6 and f6, and in addition the f-file is opened for White's manoeuvres.

17

•••

.cS

(82)

17 ...•e8 is less accurate on account of 18 ~4. All the same, in Van der Wiel-Korchnoi, Wijk aan Zee 1987, after 18...:d8 19 ~xe6 :xdl 20 :xdl fe 21 .xc7, Black proved to have enough counterplay for the pawn: 21...'it'h5 22:fl :xfl+ 23 ~xfl .fS+ 24 00 'it'bl+ 2S ~e2 .xb2+ 26 ~f3 J.f8 27 lDe4 .xa2 28 lDgS (White has closed on the black king, but his opponent finds an elegant way to force a draw) 28 ...•d5+ 29 ~g3 ~h8! (the white queen is tied to the pawn on eS, and there is now a threat of ...h7-h6. So White is forced to repeat moves) 30 ~f7+ ~g8 31lDgS ~h8 112-112. It is worth considering 18 h3. 19 lDh2 etc., just as with the queen on c8. In this case, after the exchange of one pair of rooks, the other black rook would be left on fS - where it would be guarding the f7 point but would feel rather tied down.

, !

!

102 Open Spanish

18 :'d3 The modest move with the h-pawn which first occurred in KarpovYusupov, Linares 1983, also enjoys great popularity. Let us examine this continuation by following Tal-Korchnoi, Reykjavik 1987: 18 h3 :d8 19li:Jh2

In the original game with this line, I made the knight move later: 19 :'cl c5 (better than 19...-.,,7 20 "c6!) 20 "f2 -"'7 21 li:Jh2! "xe4 (21...:t'8 22 "n! and li:Jg4) 22 "xf7+ ~h8 23 "xe6 .tg5 24 ~hl. White has retained the initiative, but Yusupov eventually managed to salvage a draw. The immediate li:Jf3-h2 looks more logical. 19...'4xdl 20 "xdl It makes sense to keep a rook on the f-file, so as to concentrate all White's forces for a kingside assault. 20.....e8

Blackjust needs one tempo for ...li:Jg5, and then the weakness of the e-pawns will tell. But at the moment, 20...li:Jg5 is bad in view of 21 "d5 "e8 22 li:Jg4, with a strong attack. In Geller-Tal, Sochi 1986, Black played 20... li:Jc5 21 "n, and only then 21 .....e8. There followed 22li:Jg3 :'d8 (Black can play more flexibly with either 22 ...:'b8 and 23 ...:'b6, or 22 ... a5 and 23 ...:'a6, switching his rook to the defence of the kingside - see the primary game) 23 li:Jg4 li:Je6 24 li:Jf5 ~h8 25 li:Jxe7 (25 h4 or 25 "g3, piling up the threats, would be even stronger) 25 .....xe7 26 "xf7 ..xn 27 :'xf7 h5 28li:Jf2 :'d5 29:'f5 g61 (after 29...:'d2! Black would hardly be risking defeat. Now the weakening of f6 has an immediate effect) 30 e4! :'c5 31 l%f6 :XeS 32 :xg6 li:Jcs 33 :c6 li:Jxe4 34 li:Jxe4 :'xe4 35 ~f2. White has a won rook ending and soon achieved victory. As a result, Tal took this variation into his arsenal with White. 21"h5

Open Spanish 103 Another square for the queen is c2. In this connection, two games by Ernst are of interest. Prasad-Ernst, Gausdal1991, went 21 ttlg4:d8 22 'ilc2 eS 23 ttlg3 c4 (after 23 ...:d7 24 ttlf5 ~h8, Black has a difficult position. MolayErnst, Gausdal1989, continued 25.e4 'ilc8 26 ttlci6!? .txd6 27 ed hd628 :xt7!?ttlg5 29'ifb7 :d1+30~h2 .d8 31 ttles! with a won position for White. On 24 ....tf8?, Molay gives the decisive 25 ttlth6+! ~h8 26 'ilf5 ttlci8 27 e6! .xe6 28 'ilxe6 ttlxe6 29 ttlxt7+ ~g8 30 ttlth6+! ~h8 31 ttles) 24 .e4 :c8 25 ttlf5 b4 26 ttlth6+!? (26 'ifb7 :c7 27 .xa6 c3 leads to double-edged play) 26 ...gh 27 ttlxh6+ ~h8 28 ttlxt7+ (28 :xt7? ttlg5) 28 ...~g8 29 .g4+ ttlg7 30 e6 .tc5 31 'ilg5 .e7 (but not 31....xe6? 32 ttlh6+ cRh8 33 .xeS! and wins) 32 ttlh6+ ~h8 33 .e5 .c7 lh-1h. 2l... lik522 ttlg3 a5 (83) 83 W

By this ingenious method, Black brings his rook across to the defence of the kingside. Afterwards, though, he tries to be too clever, and Tal brings off a striking finish. 23 ~ :a6 24 ttlg4 :g6

The knights are beautifully placed, but Black's rook is covering all the wlnerable points in the vicinity of his king. 25 b3 .td8 26 ttlj2 White has to sound the retreat. His e-pawns have turned into a real weakness, though for the moment they are immune.

26...'ilc6 27 e4 .e8 The attempt to gain a comfortable advantage lands Korchnoi in trouble.

28 'ild1 .te7 29 .d2 .tf8 30.xa5 .xe531 .xb5 L6 32 .b8 h5 33'i1d8g6?

104 Open Spanish

After 33 ...:e8, a draw would have been the most likely result. Now Tal plays an elegant combination which wins the queen. 34lDh6+ rilg7 (84) 84 W

351:i:fg4! Tal's cavalry is on the rampage... 35... hg 36 '4xj7+ rilxh6 37"xj8+ rilg5 38 h4+ rilxh4 39"h6+ "h5 40 g3+ Kxg341 "/4+! rilh442 "j2+ g3 43 :'/4+ rilg5 44 "xg3+ rilh645 :'h4 The culmination of White's combination, leaving him with a won position. I should add that at one time, the usual continuation on move 18 was 18 llXi4lDxd4 19 ed "e6. White's pawns have been straightened out, but are well blockaded. In Tseshkovsky-Yusupov, Erevan 1982, the unexpectedly quick conclusion was 20 lDg3 f6 21lDrs fe 22 'ilb3!, and Black resigned. The correct line was demonstrated by Yusupov three years later: 20 ... c6 21 lDf5 :'fe8 22 :'d3 .*.f8 23 :'h3 g6 24 lDh6+ .*.xh625 'lJ.xh6 c5, with equality; Short-Yusupov, Montpellier 1985. Nor is there any danger to Black in 22lDxe7+ :'xe7 23 'IJ.f3 :'d7 24 :'c3 "xa2 25 :'xc6 :'ad8 26 :'c8 "d5, Smirin-Mikhalchishin, Klaipeda 1988. Alternatives to 20 lDg3 similarly give White nothing: 20 :'d3 f6 21 "xc7 fe 22 "xe5 :'xfl + 23 rilxfl, Chandler-Yusupov, Minsk 1982; or 20 "xc7 :'ac8! 21 :'c2 22 :'f2 :'fc8 and Black has no difficulties, Hiibner-Ljubojevic, Tilburg 1982. In this last example, 22 .....g4!? is interesting; there can follow 23 "el.*.b4! 241Dc3 :'xf2 25 rilxf2 f6 26 ef :'xf6+ 27 rilg 1 :'e6 28 "fl .i.d6 29 "f3 "xf3 30 gf, and a draw was soon agreed in de Firmian-Hellers, Bie11989. All that remains is to mention 18 lDg3!? The game Sax-Hellers, Haninge 1989 continued 18 ...:'d8! 19lDd4lDxd4 20 ed c6 21lDf5 .*.f8

"as

Open Spanish 105

"e6

22 :d3 23 "d2 J:[d7 24 :g3 ~h8 2S "gs "g6! 26 'A'h4 J:[h3 "g6 28 :t4 :ad8 29 :g3 "e6 30 J:[h3 "g6 112-112. Now at last we return to the primary game. 18 ••. c5

19

lDd6

"e6

27

"c7

This move can be regarded as a novelty. A bad alternative is 19.....c6? 20 lilfs :a7 21 lil3d4! lilxd4 22 ed, with a clear plus for White. But a different queen move, 19...1Ib8, is sound enough, for example: 20 b3 :a7 21 :tdl :d8!? 22lLlxt7!? ~xil23 :xd8 .i.xd8, as in Stoica-Marin, Eforie Nord 1985. Stoica now gives the following drawing line: 24 "fs+ (24 "xh7 :e7! 2S :f1 ~e8 26 "gS+ ~d7 27 :dl+~c7!, and White is worse) 24 ...~e7 2S "xh7 :d7 26 'A'h4+~e8 27 'A'h8+ ~e7. 20 :rdl :fd8 21:d5 .i.f8 22 b3 White tries to obtain some initiative. After 22 h3 c4 23 b3 cb 24 "xb3 lLlcs 2S "c2lLle6, the game would be drawn. 22 ••• 22...1IbS? loses to 23 :a7 24lilxil!. 23 ..a He could have retained slightly the better chances with 23 : 1d2, vacating dl for the queen. 23 24 e4 :ad7

"f5

:.7

"c6

25

h3?

Now the advantage passes to Black. Instead, 25 "g3 c4 26 be bc would have maintained mutual chances. 25 f6 26 "g3 c4 27 be be

28

"g4

White would lose with 28 :c 1 .i.xd6 29 ed :xd6 30 :xd6 :xd6 31 :xc4? :dl+.

28 29

lilc7

lDd4

"as

30 lil4CS? Another substantial error. After 30 lile6!? h5 31 "fslLlxds 32lilxd8 :xd6 33 ed lile3 34 "e6+ ~h7 35 "eS lilxdl 36 "xh5+, a draw would be appropriate. BeJiavsky claimed that Black could retain a plus

106 Open Spanish

"e6+

with 32.....xd8!? 33 Wh7 34 IlxdS c3, but then 35 exf6! appears crushing, e.g. 35 ...gxf6 361&8! or 35...1fb6+ 36 Wfl -"'1+ 37 Wf2 "xa2+ 38 ~e3 (Nunn). 30 "87+ j,xd6 31 lZSd4

32 33

eel fiJe7+

~bS ~h8

34

lOc6

1fb6

3S

eS 35 IOxd8 loses to 35 ...llxdB 36 eS fe.

3S 36

... e6

"xc6 Ilxd6

37 e7 1le8 37...lOxd4 may appear decisive, but after 37 ...lOxd4 38 .xd4! 1le8 39 .xd6 "xd6 40 Ilxd6 ~g8 41 1lxa6 Ilxe7 42 1lc6, followed by a4, White should draw (Nunn).

38 Ilxd6 39

"e6

40

~hl

After this the game heads towards a draw, but instead 4O...c3 wins: 41 "xd6 c2 (and not 41.. ...xd6? 42 Ilxd6 ~g8 43 Ilxa6 Ilxe7 44 Ilc6 1le345 a4!) and White can resign, or 411ld5 1fc4 42 ':xd6 "f1+ 43 ~h2 c2 winning, as Black threatens to mate by checks with 44 .....f4+ 4S g3"f2+.

41

1feS

Ild7

Black could still have kept a minimal edge with 41 ...'IIc8 42 "dS c3 43 "xb7 "xb7 (not 43 ...c2? 44 .xc8, and White wins) 44 Ilxb7 ~g8 45 lla7 ~f7 46 llxa6 Ilxe7 (Beliavsky). 42.fT 1lg8 43 Ilxb7 c3 44 1lc7 .el+

lb..lb. Game No. 17 Van der Wiel·Hjartarson

Rotterdam 1989 1 e4 eS 2 00 lOc6 3 j,bS 86 4 j,a4lOi6 S 0-0 iDxe4 6 d4 bS 7 j,b3 dS 8 de j,e6 9 ~bd2lDcS 10 c3 10 ••• j,e7

Open Spanish 107 10...d4 was examined in detail in the context of the previous game. Another move Black often plays is 1O....i.g4; but usually this simply transposes into the 1O....i.e7 line, since one way or another Black has to get castled. 11 .i.c2 This move is played almost automatically. The novelty 11 ~4!? was used in Bryson-Flear, Dundee 1991. Black could now have secured equality with 1l...~xd4 12 cd ~xb3 13 ~xb3 0-014 f4 f5, but he took the pawn instead, and after 11...~xe5 12f4 ~c4 (12....i.g4? 13 .el ~ed3 14 .g3, with a dangerous attack) White could have increased his initiative with a further infantry advance in the f-file: 13 f5 (instead of 13 .e2 as played). 11 .i.g4 (85) 85 W

One of the critical positions in the Open Variation. The main line here, which actually occurred in the game, is 12l:[el. But first we will examine another interesting continuation. Akopian-Todorovic, N~ic 1991, went 12 .el!? (White frees himself from the pin and prepares to jump to d4 with his knight, but his queen is occupying rather an eccentric post) 12...0-0 (in Akopian's view, 12...~6!? was more solid) 13 ~4 ~xe5?! (13 ...•d7 is met by 14 h3. A roughly equal game would result from 13 ...~xd4 14 cd ~e6. Taking the pawn is quite risky. In his notes to the game Akopian indicates that it is this knight sacrifice that breaks new ground, but the whole line with 12 .el has not occurred in any other grandmaster games I have seen) 14 h3! (14 .xe5 is also good for White, for example 14....i.d6 15 .e3l:[e8 16 ~c6 .i.xh2+ 17 ~xh2 ~4+ 18 ~gl l:[xe3 19 fe ~6 20 ~7+ ~f8 21 ~xd5 c6 22 ~f4) 14...~d3 15 .i.xd3 ~xd3 16 .e3 ~c117 hg .i.g5 18 f4 c5 19 ~6 l:[e8 20 ~xd8l:[xe3 21 fg ~3 22 ~xn, and White won.

lOB Open Spanish

12

:el

Of course, this position can be reached by various routes; another possible move-order is 9 c3 ~S 10 J.c2 J.g4 11 :el J.e7 12 ~bd2. In the game, Black continued with 12...•d7. The alternative 12...d4, which used to be regarded with complete approval by theorists, has completely disappeared from practice. A further possibility is 12... 0-0 13 ~f1 (86). . 86 B

Let us look at some important examples. Kupreichik-Kaidanov, Kuibyshev 1986, went 13...J.hS 14 ~g3 J.g6 IS~S.d716g4:ad817h4! ~ 18~xe7+~e719~~S20

J.e3 ~6 21 J.b3 cS 22 hS J.e4 23 h6!, with an initiative for White. Ivanchuk-Thkmakov, New York 1988, went 13...J.hS 14 J.e3 (the manoeuvre ~1-g3 is postponed for the moment; instead, White sets up the threat of 15 J.xh7+ ~xh7 16 ~gS+ J.xgS 17 .xhS+ and 18 J.xcS) 14...J.g6 15 ~g3:e8 16 h4! J.xc2 17 .xc2 ~7 18 J.f4llli8 19 hS, and White had strong pressure. However, in Ivanchuk's view, Black could have solved his opening problems with 14...~! 15 ~g3 J.xf3 16 .xf3 ~xe5 17 'ii'hS, and now not 17 ... ~g6 18 f4, but 17 ...g6!, and if 18 .xe5 J.f6 it is Black who unexpectedly wins. White should instead play 17 J.xh7+ ~xh7 18 'ii'hS+ ~g8 19.xe5 J.d6 with an equal game. At move 14, another possibility is 14... ~xe5 15 J.xc5 ~xf3+ 16 .xf3! J.xf3 17 J.xe7 .d7 18 J.xf8 J.xg2! 19 J.cS J.xfl 20 ~xf11Vh3+ 21 ~gl .g4+ with perpetual check. Wang-Yusupov, Novi Sad 1990, went 13 ...:e8!? (a novelty; Black doesn't waste time retreating his bishop, but immediately pressurises the centre) 14 ~3 ~e5 14 J.xh7+ ~xh7 16 .c2+ ~g8 17 ~xeS J.e6 18 ~6 .d6 19 ~e7+ :Xe7. Black equalised, and in fact soon seized the initiative.

Open Spanish 109

Wahls-Hiibner, Munich 1991, went 13...lIe8 14 h3 (instead of 14 14....th5 15 g4.tg6 16 .txg6 hg 17lDe3 .d7 18.xdS lIad8 19 lIdl (19 .xd7 lIxd7 favours Black) 19...•e6 20.xe6 fe 21 ~4 ~xe5 22 f4 ~f123 lin .tf6 24 ~c2 e5, and again Black seized the initiative. Geller-Unzicker, Bad Worishofen 1991, went 12...0-0 13 ~b3 ~ 14 .d3 g6 15 .th6, and White obtained somewhat the better chances after 15 ...lIe8 16 lIadl .tf5 17 .d2 .txc2 18 .xc2 .d7 19 lId3 lIad8 20 h3 .tf8 21 .d2 ~7 22 .txf8 lIxf8 23 ~d4 .c8 24 'ife3. 12 •.• .d7 13 eon We may recall that I played 13 ~b3 as long ago as the 28th match game against Korchnoi at Baguio. After 13 ...~e6 14 h3.th5 15 .tf5 ~8 16.te3 as 17 .tc5 a4 18 .txe7 'ifxe7 19 ~bd2 c6 20 b4 ~g5, a double-edged game ensued. At move 17, White could have occupied c5 with the other piece. An interesting example is Ehlvest-Hjartarson, Belfort 1988: 17 ~5!? .c6! 13 ~3 .txf3 19 'ifxf3 g6 20 .tg4 h5 21 .txe6 ~e6 22 lIadl lId8 23lld2 0-0 24 lied 1 ~g5? (a serious mistake; 24...•d7, followed by 25 ...c6, was better) 25 .txg5 .txgS 26 lIe2 lIfe8 27 'ifg3 .th6 28 lIdel ~h7 29 e6! f6 30 ~g7 (87) ~3)

.f3

87 W

.n

31 ~! fe 32 .n+ ~h8 33 lIxeS lIg8 34 e7 lIde8 35 lIe6 .d7 36 lIxg6 lIxg6 37 .xg6 .tg7 38 d4 39 cd .txd4 40 1Ie6 .tg7 41 g3! and Black resigned, since he has no adequate defence against the threat of lle6-e4-h4xh5+. Another game of importance is Nuon-Tal, Nlestved 1985, in which Black varied with 15....tg6 (in place of 15...~8). There followed 16 ~fd4 0-017 .tg4! ~xd4 18 cd as 19 f4 h5 20.txh5 .txh5 21 .xh5

110 Open Spanish a4 22lbcs lOxcs 23 dc 1.xcS+ 24 1.e3 1.xe3+ 25 :xe3 fS 26 e6 'fIe7 27 :ael :ad8 28 g4! d4 29 gS! g6 30 'fIxg6+, and White won in another ten moves. A bad reply to 131Ob3 is 13...0-0; a game Ivanchuk-Haba, Hungary 1988, continued 1410xcs 1.xcs IS h3 1.e6 16 lbd4lOxd4 17 cd i.e7 18 1.e3, and White had full control over the dark squares and the halfopen c-file. By bringing his knight to e3, White will drive the bishop back, so there is no point in wasting a tempo on h2-h3. A suitable illustration is Hubner-Korchnoi, Tilburg 1986: 13 h3 1.h5 14 lOn :ad8 IS lOg3 1.g6 16lbd4 0-017 1.f5 (171OgfS is more exact) 17 ...lbe6 18 1.g4 lbcxd4 19 cd cS 20 lOfS 'fIa7 21 lOxe7+ 'fIxe7 22 1.e3 cd 23 1.xd4 :c8. By now the initiative is on Black's side, and it soon became decisive. 13 :d8 14 lbe3 i.h5 15 M IS b4 is interesting. Hjartarson-Korchnoi, St John Ct 1988, continued IS ...lbe6 (1S ... lbe4 loses at once to 161OxdS!) 161OfS, and now Black replied with the unsound 16 ... d4? After 17 i.e4! 1.g6 18 g4 hS 19 h3 ~f8 20 a4! hg 21 hg 'fIe8 22 ab ab 23 :a6!, he was crushed (23 ...lOb8 24 :xe6 fe 2S lOxe7 1.xe4 26 :xe4 etc.). Black also had a hard time in Rodriguez-Marin, Novi Sad OL 1990, in which he reacted with 16...0-0. Play went 17 a4 :fe8 18 ab ab 19 'fId3 1.g6 20 'fIxbS! (previously White had played 20 :dl and achieved nothing) 20 ...lOxeS 21 'fIxd71Oxd7 221Oxe7+ :xe7 23 i.xg6 hg 24lOd4!, and the ending is in White's favour. Evidently Black's safest course is to withdraw his bishop at once with 16...1.g6. After IS b4 lOe6, the immediate 16 g4 is worth considering. In Zagrebelny-Neverov, Barnau11988, there followed 16...1.g6 17lbfs 0-0 18 a4 :re8 19 ab ab 20 1.d3 :b8 21 'fIe2lbd8 22 :a7 with advantage to White. White gained nothing from IS 1.f5 in Chandler-Hjartarson, Novi Sad OL 1990; after IS ...lbe6 16lbc2 0-017 a4 :re8 18 ab ab 19 'fId3 i.g6 20 i.xg6 hg 21 1.e3 b4 22 :edl bc 23 be i.f8 24 'fIfl lOe7 2S lOcd41Oxd4 26 cd lOfS, the players agreed a draw. 15 ... 0-0 Black has castled at last, and has no problems at all. 16 lOxe7+ CiJxe7

Open Spanish 111 17 1.e3 Hiibner-Korchnoi, Tilburg 1987, concluded amusingly with 17 b4 f&41 18 1.xh7+! ~xh7 19 e6! 1-0. Of course, this is not a refutation of Black's opening variation. The correct defence was demonstrated by Korchnoi against A.Sokolov in the same T"llburg tournament: 17...lbe4! 18 1.xe4 de 19 "xd7 l:lxd7 20 ~gS 1.g6 21 e6 l:ld3 22 ef+ 1.xn 23 ~e4 00. Black has his full share of the chances, and was able to win the endgame. 17 ••• ~4 18"d3 ~g6 After 18...1.g6 19 "d2, White is slightly better. 19 b3 19 e61 is inferior: 19 ...fe 20 ~S ~xb2 21 ~xd7lDxd3 22lDxfS ~xel 23 1.xg6 1.xg6 24 ~g6 ~2 2S ~7+ ~f8, and Black has a very promising position; Marjanovic-Korchnoi, Belgrade 1987. 19 1.xf3 20 gf(88) 88 B

20

3 1.d2 Better than 21 ba ~eS 22 "xh7+ "xh7 23 1.xh7+ ~xh7 24 ~g2 l:lfe8, with the better chances for Black. 00'

....

21

21

~

22"rs

"xIS

24

lOe6

23 j,xIS lUeS 24 1.e3 24 j,gS1! favours Black after 24 ...~eS 2S j,xd8 ~xf3+ 26 ~ ~1 27 l:lxell:lxd8 28 l:le7 86 29 l:lxc7 ~xb3 30 ab gf. 00.

112 Open Spanish 25 .i.xg6 hg After 25 ... fg Black has slightly the better endgame. 26 a4 cS 27 ab ab 28 f4 d4 29 cd cd 30.i.d2 d3 31.i.e3 :cIS Now the position levels out for good. Black could still have tried for the initiative with 31...taJ4. 32 :ed1 33 ~g2 :c3

:cs

34:a8+

~h7

35 :a6! :d8 If 3S ...:xb3? 36 :d6 :Xd6 37 ed, it is White who obtains a plus. 36 b4 ~g8 37 :d6 ~

38

:Xd8+

lOxd8

39

~

lOe6

:c4

4O:a1 41

:c3

:d1

1/2_112 Game No. 18 Speelman-Thuman

Candidates Semi-Final, 4th game London 1989

1 e4 e5 2lill3li)c6 3 .i.b5 a6 4 .i.a4li)f6 5 0-0 ll)xe4 6 d4 b5 7 .i.b3 cIS 8 de.i.e6

9

c3

The two preceding games featured 9li)bd2. 9 ••• .i.cS The position after 9 ... .i.e7 10.i.e3 will be studied in the next game (where the move-order is 9 .i.e3 .i.e7 10 c3). The variation introduced by the bishop move to cS is sometimes called the Italian System. 10 li)bd2 Other possibilities are 10 'iVe2, 10 'iVd3 and 10 a4, but according to present-day theory Black obtains a fully viable game in all cases. 10 ... 0-0

Open Spanish 113

11

.i.c2(89)

89 B

In this position Black has four continuations. Two of them - 11 ...fS and 11 ...~xd2 - are not seen in contemporary practice. The other two - 11 ...~xfl and 11 ....i.fS - remain popular in our own day. Before turning to the bishop excursion to fS, let us examine some modem examples of the capture on fl. As a basis we will take the game Ivanchuk-Yusupov, Linares 1990: 1l... l&j2 The idea of this positional sacrifice, involving an attack down the ffile by Black, belongs to Dilworth. In the 1940s, a major contribution to its theory was made by Botvinnik. At the present time the variation has not lost its attraction. 12 :x.J'lf613 ef J..xj2+ 14 rJilxj2 "xf615 ~f1 More precise than withdrawing the king to g1. 15... ~e516 .i.e3 Again there is no hurry for 16 rJilg1. With 16...~f3+ 17 gf"xf3 18 "xf3lbi'3, Black restored material equality and kept some initiative in Morovic-Yusupov, 'funis IZ 1985. 16... flae817 .i.c5 &iJxj318 gf:p (90) 19l1lg3! A valuable innovation. In Short-Yusupov, Belgrade 1989, Black obtained a fine game after 19 .i.d3 .i.h3! 20 ~g3 h5! 21 .i.fl .i.g4. Nor does 19 rJilg2 promise much, for example: 19.....gS+ 20 rJilhl (20 ~g3 h5 21 rJilhl h4 22 "d3 'fIb6 23 .i.e3 'fIb8 24 ~fl .i.fS is in Black's favour - Soffer) 20...d4! 21 "xd4 (21 .i.xd4? also loses: 21.. ..i.dS 22 ~ :Xf3! 23 ~ :el+!) 2l...lbi'3 22.i.e4:t4 23 :el :xe4! 24 :Xe4 .i.dS, and it is time for White to resign; Griinfeld-Mikhalevsky, Israel 1991.

114 Open Spanish

19....t.g4 20Wg1"xj3 21"xj3 .t.xj3?! After 21...:xf3 White would have a slight advantage, but now, after his next move, he has a very substantial one. 22 ~l! '/Zf6 If 22 ....t.g4, then 23 .t.xh7+. 23 b4! c6 White now played 24 .t.f5?, and the position quicldy levelled out. However, after 24 .t.d4 l:r.f4 25 .t.f5, threatening 26 .t.d7, Black would be left with some very complicated problems. .t.rs 11 ••• This continuation was studied in detail at the World Championship Match in Baguio, but has remained in fashion to this day. Utilising d4 as a base, White aims to start active play in the centre and on the kingside. Black normally prepares counterplay on the queenside.

12

tilb3 (91)

A critical position. Black now has the choice between 12....t.g4, which was popular in the seventies, and 12....t.g6 - a move that came

Open Spanish 115 into widespread use in the following decade. In Morovic-Murey, Thessaloniki OL 1984, Black played instead 12....txf2+ (in the spirit of the forties!) 13 l%xfl ~xfl 14 ~xfl .txc2 15 .xc2 f6, but after 16 e6 .d6 17 .te3 .xe6 18 ~bd4 ~xd4 19 ~xd4 White obtained a plus. 12 .•. .tg6 In reply to 12....tg4, I used to exchange automatically on c5; in this way I managed to acquire an advantage against Beliavsky and Smyslov in Leningrad 1977. But at Baguio, Korchnoi twice succeeded in obtaining a good game, and I therefore switched to 13 h3. Seeing that 13 ....txf3 loses a piece to 14 gf, Black is compelled to retreat with 13 ....th5, and after 14 g4 .tg6 15 .txe4 de 16 ~xc5 ef 17 .tf4 White has a clear advantage. Here are a few more moves of Karpov-Korchnoi, 14th game, Baguio 1978: 17 ...•xdl (a forced exchange; 17...•e7 18 .d5 ~aS 19 b4 ~4 20 .xf3 ~xeS 21.txe5 .xeS 22 ~7 is no good for Black) 18 l%axdl ~8 19 l%d7 ~6 20 ~xe6 fe 21.te3 l%ac8 22 :rdl .te4 23 .tc5 :re8 24 l%7d4 .tdS 25 b3 as 26 ~h2 l%aS 27 ~g3, and White converted his positional advantage into a win. In this variation Black doesn't appear to have come back with any worthwhile improvements in the past decade.

13

~d4

13 a4 has frequently been played, for example 13 ....tb6 14 ~bd4 ~xd4 15 ~xd4, as in Van der Wiel-Korchnoi, Wijk aan Zee 1983; and now 15 ...•d7 (not 15 ...c5 16 ~6 ~xc3 17 ~xd8 ~xdl 18.txg6 l%axd819.tf5!) 16.te3 ~5 17 as .ta718 f4 .txc219~xc2 f6 20ef l%xf6 21 ~h 1 c6 with eqUality. Black has nothing to fear from 13 e6 or 13.e2. In Short-Timman, Tilburg 1988, White played the new move 13 .tf4, and after some sharp play his opponent emerged with some advantage: 13 ....tb6 14 a4.d7 15 ab ab 16 l%xa8 l%xa8 17 ~fd4 b4 18 .td3 be 19 .tb5 ~xf2! 20 l%xfl (20 .txc6? .xc6!) 20...~xd4! 21 .txd7 ~xb3 22 be (22 .xb3 l%a1+) 22 ...l%aI23 .xal ~xa124 .tc6 .te4 25 c4 ~2. White can also occupy d4 with the other knight, as in Zso. PolgarVan der Steeren, Wijk aan Zee 1990. After 13 ~bd4 ~xd4 14 ~d4 .tb6 15 .te3, Black played the inferior 15 ...l%e8?! (Polgar gives 15 ...•d7 or 15 ...•e816 f4 f6 as correct), and White obtained the better game with 16 a4.d7 17 ab ab 18 l%xaS l%xaS 19 .td3 c6 20 f4 l%e8 21 ~hl. At this stage Black could have maintained a defence with 21 ....txd4 22 .txd4 .tf5, but after 21 ...f6?? White played the winning

116 Open Spanish combination 22 e6! lIxe6 23 fS! .txfS 24 ~xfS J..xe3 25 ~xe3 ~xc3 26 "'g4lDe4 27 .txe4 de 28l1xi6! lId6 29 "'xd7 lIxd7 30 lIxc6 lId3 31 lIc3 lId2 32 lIc2l1d3 33 lIe2 1-0. 13 ••• J..xd4 14 cd In the sixth game of the same match, Speelman preferred 14 ~d4. There followed 14......d7 IS ~c6 (1S f4 gives White nothing: KlovanDorfman, USSR 1981, concluded IS ...~xd4 16 cd f617 J..e3 lIad8 18 "'e2 Wh8 19 :acl c6 20 .td3 lIc8 21 lIc2 J..fS 22 lIfcl 112-112) IS ......xc6 16 J..e3 lIfe8 (92). 92 W

The chances are roughly equal here. But the move 17 f3? (better 17 J..d4, although after 17 ...t2)cs Black has no problems) allowed Black to deliver the surprising blow 17 ...~xc3! 18 "'d2 J..xc2! 19 "'xc2 :xeS 20 h i ~2+!, not only saving his piece but emerging with an extra pawn. Still. in the double rook endgame after 21 lIxe2 "'xc2 22 lIxc2 lIxe3 23 :xc7, White managed to escape with a draw. Curiously enough, the invasion on c3 had been overlooked by White on a previous occasion, K.orchnoi-Karl, Swiss Ch 1982. From the diagram, play went 17 f4? ~c3! 18 be "'xc3 19 fS "'xe3+ 20 Wh 1 lIxeS 21 fg hg, and the armada of black pawns in the centre is clearly stronger than the bishop. In J .Polgar-Hellers, Wijk aan Zee 1990, instead of exchanging on c6, White sacrificed a pawn with IS a4. After IS ... ~xeS 16 f4 ~c6 17 ~xc6 "'xc6 18 fS, Black could have secured a comfortable game with 18 ......cS+ (instead of 18 ...~6+ as played) 19 "'d4 J..hS 20 J..f4 "'c6. 14 as 15 .tel

Open Spanish 117 White can also play IS.i.d3a416.i.xbS!?lDb4?! 17lDd2lDxd218 .i.xd21Dc2 19 IlcllDxd4 20.i.gS "'xgS 21 "'xd4, as in TseshkovskyTukmakov, Tashkent 1980. An improvement for Black is 16... ab 17 .i.xc6 lla6 18 f3 Ilxc6 19 fe ba 20 Ilxa2 .i.xe4, with equality; Ljubojevic-Tal, Nik§ic 1983. 15 ... a4 In an old game Karpov-Savon, Moscow 1971, White acquired a big advantage after Is ...lDb4? 16 .i.bl a4 17lDd2 a3 18 "'ct!

16

Ilkl

At this point it is worth recalling the interesting game Karpov-Yusupov, Moscow 1983, in which White played 16lDd2. The continuation was 16...a317lDxe4ab 18Ilbl.i.xe419Ilxb2"'d720.i.d3 (this move - in place of 20 .i.xe4, which leads to equality - was first played by Hubner against Korchnoi in 1982. The bishop exchange took place all the same - 20....i.xd3 21 "'xd3 - but left a pawn sbucture more to White's liking. The players reached this position twice. In the first game, in Chicago, a draw was agreed after 2l...llfb8 22 Ilfbl b4 23 a3 ba 24 Ilxb8+ Ilxb8 2S Ilxb8+ lDxb8 26 "'xa3 "'c6 27 'fIIe7 'fIId7 28 "'a3. In the second, in Lucerne, play went 2l...b4 22 .i.d2 Ilfb8 23 Ilfbl 'fIIg4 24 .i.e3 1lb6 26 h3 "'c8 26 Ilc2 b3! 27 Ilxb3lDb4, winning the exchange and eventually the game. But White didn't have to throw away material; with accurate play his pressure is very effective. My game with Yusupov is a case in point ... ) 20....i.xd3 21 "'xd3 Ilfb8 22 I:[fbl b4 23 h3 h624 Ilcl 1lb6 25 'fIIbl llab8 26 IlcslDd8 27 :Cc21Dc6 28"'cl 1l8b7 291lcSftJe7 30cRh2lDfS 31 Ilbc21lg6321lxc7 Ilxc7 33 Ilxc7 'fIIbs 34 g4! lDh4 3S Ilc8+ cRh7 36 "'dl "'a6 37 Ilc2 fS 38 cRg3! fg 39cRxM gh40f4 'fIIe641 'ifh5! "'e7+42 cRxh3 431lh2!"'d7+ 44 f5 1-0.

"'f7

16

...

a3

After 16...lDb417 .i.bl a3 18 b3, White keeps a slight edge; Tseshkovsky-Geller, Vilnius 1983. 17 b3 Also 17 ba Ilxa3 18 .i.b31Dc3 19 "'d2 b4 20 lDd3 .i.xd3 21 "'xd3 has been played quite often. Short-Yusupov, Belfort 1988, now continued 2l...lDas 22 J.c2"'M 23 .i.cl Ilxa2 24 Ilxa2lDxa2 2S .i.gS 'ifhs 26 .i.e7 Ilb8 27 "'a61Dc4 28 'fIIxa2 lDd2 29 .i.xh7+ and White had a won position. In Nunn-Marin, Thessaloniki OL 1988, Black played the important innovation 21.. ....a8!, and after 22 .i.d2 a draw was agreed, although Black could have obtained the better chances with 22 ...lDxa2 23 IlfellDas 24 llabllDc4.

118 Open Spanish 17 18 19

ef 1De2(93)

f6 "xl6

93 B

All this had already been seen in the sixth game of my match with Korchnoi (Merano, 1981). Mter 19...ll)b4 20 .tb1 'fIe7 21 'fIel :re8 22ll)f4.tf1 Black went on to achieve equality in the game, although by playing here 23lOd3! lDxd3 24 .txd3 I could have retained slightly the better chances. 19 ••• 'fIe7 Black withdraws his queen at once. It is hard to say whether this is best. 20:el ll)b4 20...:f6, to double rooks on the f-file, may be better. Then 21 f3 is bad on account of 2l...ll)c3, while 21 .txe4 .txe4 22 .tgS fails to 22...:g6. Still, by continuing 21 .tb1 and 22ll)c3, White could maintain the pressure. 21 .tbl l:tae8 On 21 ... lOd6, White plays 22 .td2! lOd3 23 .txd3 .txd3 24 :e1, obtaining an undoubted plus. 22 ll)c3 At this point 22 :e I? or 22 ll)f4? would fail to the powerful retort 22...ll)xf2.!. But 22 .tf4 cS 23 dc ll)xcs 24 .te3 looks good; White acquires an object of attack in the shape of the isolated pawn on dS, while Black is left with a weak dark-square complex. It was also worth considering 22 'fIel I?~ 22 lDxc3 23 ]bc3 .txbl d;(94) 24 'fIxbl

Open Spanish 119

94

W

25 .tel? After 25 :'c5, threatening 26 .td2, White would retain the better chances. Then 25 ......d6 (25 ... Af6? 26 .tg5) could be quite well answered by 26 "'dl ! - White consolidates his forces and doubles rooks on the c-file. The English grandmaster misses his opponent's striking rejoinder. 25 ... "el! Attacking the rook, and also threatening 26 .....xt2+ 27 :'xf2 :'el+.

16:'e3 27 .txe3 28 lb::bl

lb::e3 "'xbl

:'e8?! A drawish rook ending would arise from 28 ...lLlxa2 29 AallLlb4 30 :'xa3lLlc2 31 Aa7lLlxe3 32 fe :'f6. But prompted by his opponent's error, Timman decides to have a try at winning, and ends up making some bad mistakes himself. 29 .td2! Weak alternatives are 29 :'cllLlxa2 30 :'allLlb4 31 :'xa3lLlc2 32 :'a6lLlxd4!, and 29 ~fllLlc2 30 :'cllLlxe3+ 31 fe :'xe3 32 :'xc6 :'d3 33 :'c5 b4; in both cases Black has good winning chances. 29 1Oc2 30 :'el! lLlxd4

31

~

:'e4?

3l...:'e2 fails against 32 .te3 :'c2 33 :'xc2lLlxc2 34 .tc5!, and the black knight is trapped. It was essential to play 3l ...:'c8! 32 b4lLle6 33 :'c3 c5 34 be :'xc5 35 :'xa3, with a balanced position. 32 b4! rM1

33:'c3 34 f3 3S b3

~e6?

Ah4 AhS

120 Open Spanish

After 35 .. .c~d6 36.i.g5 J%hS 37 h4, the black rook would be trapped.

36 37

lha3 :d3(95)

:rs

37

••.

cS!

95 B

Black sacrifices two pawns, and in return obtains maximum activity for all his pieces - particularly his king. ~

38 be 39 :b3 40 :Xb5 41.4 42 as 43
d4 :17 :'7
46:a4

:'6

47 (4 48:c4 49

g5 With every pawn exchange, White's winning chances are reduced. He now gains nothing from 50 :xc6 :xc6 51 fg :c4! 52 h4 :a4 53 hS ~e6, and Black has everything in order. The game concluded:

5013 gt 51 gt:'7 52:'4:g? 53 .tel:n 54 .6 lbt4 55 .7 :r3+ 56
64 .in h5 65 :h7 :'4 66 :xh5 :'1+ 67 .tel:a2 68 :"3+
Open Spanish 121

69 .tn :el 70 ~g2 :c4 71 l:lh8 :c3 72 :e8+ ~4 73 :el :b3 74 .tell1d3 75 .td2+ ~ 76.te3 ~e4 77 ~ :b3 78:a2:b4 79 ~e2 :c4 80 :a8 :c2+ 81 .tdl :c4 82 :e8+ ~d5 83 .te3 :a4 84 :"8 :a2+ 85 ~ :a4 86 :h5+ ~e6 87 .tc5 :c4 88 ~e3 :g4 89 .td4 :g3+ 90 ~4:g2 91.te3:al 92 ~e4 :e2 93 l:lh6+ ~e7 94:a6 :el 95 ~d4 Wf'T 96 .tf4:e6 97 .td6 ~ 98 .te5+ Wf'T 99 :a7+ :e7100 :al :d7+ 101 ~e4 ~e6 102 :a6+ Wf'TI03 ~5 ~e8 104 :e6+ :e7 105 :b6 :d7 106 l:lh6 :t7 107 ~e6 :e7+ 108 ~d5 :d7+ 109 .td6 :g7 110 l:lh8+ Wf'T III :t'8+ ~g6 112 ~e6 ~g5 113 :n :g6+ 114 ~e5 ~g4 115 :gl+ ~h5 116 :dl :g5+ 1h.-lJz Game No. 19

Dolmatov-Yusupov Candidates Quarter-Final. 3rd game Wijk aan Zee 1991

1 e4 e5 2 ~f3 ~ 3 .tb5 a6 4.ta4 ~f6 5 0-0 ~"e4 6 d4 bS 7 .tb3 d5 8 de.te6 9 .te3 We have already analysed in detail the popular continuations 9 ~bd2 and 9 c3. It remains for us to examine a third possibility which

occurs more rarely.

9

...

10

c3

.te7 "d7(96)

According to current theory, the other well-known continuations 1O... ~5, 1O... lDa.5 and 10... 0-0 - allow White a distinct plus. In this Candidates Match, the rare queen move to d7 was subjected to thorough investigation. Yusupov resorted to it three times, and overall it

122 Open Spanish stood up to the test. I have chosen the third match game as our 'priDW)" game. I should add that in the seventh game Yusupov preferrCd 10...lDcs, and failed to equalise. Here is what happened:

10...&&5 11 .i.c2 f:iJd7 In the extra eleventh (speed chess!) game, Yusupov opted for 11....i.g4 12 ~bd2ltle6 13 iWbl .i.h5 14 J.fs J.g6 Isl:dl, and then at last played his 'patent' queen move Is ....d7. After 16 .c2 0-0 17 ~1 ~aS 18 ~g3 cs 19 h4 ~ 20.i.c1 :fe8 21 b3 ~b6 22 h5 J.xfs 23 ~xfs J.f8, the game was about level. 12 :el! Another novelty. 12... 0-013 J.j4lDb6 14 f:iJd4 ~4 15 cd c5 16 f:iJd2 ~c4 17lDb3 l'Dxb218.bl c419J.xh7+ WhB 20:e3.c8?! A more accurate choice was 20....d7 211Dcs J.xcs 22 dc d4. Now White obtains a substantial initiative. 21 J.c2! c3 22 ~c5? An error which brings White to the brink of defeat, whereas with 22 .e1! b4 23 a3! he could have maintained a sizable plus. 22... J.xc5 23 lIxc3 J.xd4! 24 lIxc8 Lxc8 Black stands better, but after some mistakes by both sides the game ended in a draw. 11 ~bd2 :d8

12

:el

A move which Dolmatov had prepared specially for the match. Previous games had gone 12 ~xe4 de 13 .xd7+ J.xd7 14 ~gs ~xes 15 .i.d4 J.xgs 16 J.xes 0-0 17 J.xc7 :c8 18 J.b6 :fe8 with approximate equality, Timman-Korchnoi, Reykjavik 1987; or 12 h3 ~d2 13 .xd2 ~ 14 J.gs c5 IslUel ~616:adl h617 ....xe7 .xe71S ....c2 0-0 with a minimal plus for White, Short-Ljubojevic, Linares 1989. Let us follow the latter game a little further: 19 'iVd3 g6 20 .e3 Wg7 21 a3 22 .f4 :d7 23 :d2 a4 24 h4 f5 (24 ...~aS, aiming for c4, is more precise) 25 ef+ :xf6 26 .g3 'iVd6 27 ~ ~xeS 28 :XeS J.fs?! (he should have retreated the bishop to f7; now White obtains a significant plus) 29 :dxdS .xdS 30 :xds :xds 31 c4! etc.

as

12

...

0·0

In the event of 12...~xd2 13 .xd2 ....g4, Dolmatov was intending a pawn sacrifice: 14 ~4! ~xes 15 J.h6! with a big advantage. 13.i.c2 ~d2 An innovation. In the first game of the match Yusupov chose 13.......fs, which finally resulted in an ending favourable to White: 14

Open Spanish 123 lLlxe4 .ixe4 (the more accurate recapture was 14... de 15 'i'xd7 :xd7 16 lLld4lLlxd4 17 cd) 15 .ixe4 de 16 'i'xd7 :xd7 17 e6! (breaking up the black pawns) 17...fe (17 ...:d6 is worse: 18 ef+ :xf7 19lLld2 :e6 20 :adl, followed by lLlb3) 18lLld2lLle5 19 lLlxe4 lLld3 20 :e2 c5 21 g3 c4?! (better 21 ...:dS) 22 b4! :f5 23 a4 <;Pf7 24 ab ab 25 <;Pg2:e5 26 f3 .if6 27 :a3 h6 28 h3 .ie7 29 :a8 (29 :ea2 was more precise, to stop Black from obtaining counterplay with the exchange sacrifice which occurs in the game) 29 ....if6 30 :b8 :xe4! 31 fe .ixc3 32 :xb5 .ixb4 33 <;Pfl .id6! 34 g4 .ie5. Bringing the bishop to e5 has enabled Black to hold the position. The game ended in a draw. In Fedorowicz-Kamsky, USA Ch 1991, White replied to 13 ....if5 not by exchanging on e4 but with 14lLlb3. There followed: 14....ig615 lLlbd4 lLlaS 16 e6!? (16 a4 lLlc4 17 .icl c5 is in Black's favour) 16.....d6! 17 ef+:Xf7 18 a4lLlc4! (in the complex struggle Kamsky gradually outplays his opponent) 19 ab lLlxe3 20 :xe3 ab 21 lLlxb5 'i'b6 22lLlbd4 c5! 23lLlb3 .if6 24 :bl d4! 25 cd cd 26 :d3 <;Ph8 27 lOcI "c7 28 :al 29 lLlb3lLlg5!, and Black won. 14 'i'xd2 .if5 15 :adl :reS 16 b3 The alternative 16 .if4 is inadequate because of 16....ixc2 17 'i'xc2 f6 18 e6 'i'c8. 16 ••• h6 In the fifth match game Black played 16.....e6, and White could have obtained the advantage with 17 .tf4 (17 ....ixc218 "xc2 'i'g619 "xg6 hg 20 e6). Instead, after 17 .ig5?! a lively struggle commenced, with chances for both sides. Play proceeded 17....ixc2 18 'i'xc2 "g6! 19 "xg6 hg 20 .if4 .ic5 21lLld4lLlxd4 22 cd .ib6 23 :e2 :e6 24 :c2 f6 25 h4 <;Pf7 26 <;Ph2 :d7 27 .ig3 fe 28 .ixe5 c5 29 dc .txc5 30 f4 .ie7 31 <;Pg3 .if6 32 .ixf6 gf. The position is drawish, but in the rook endgame Black went wrong in time-trouble and eventually lost. If 16....ixc2 (16 .. .f6?! 17 ef, and after .ib3, thedS point is shaky) 17 'i'xc2 f6 18 ef .ixf6 19 .ig5! :xel+ 20 :xel 'i'f7 21.ixf6 'i'xf6 22 "d3, White has the better game (Dvoretsky). 17 .if4 .iIB 17....ixc2 18 'i'xc2 f6? loses to 19 e6 "c8 20 "g6. 18 .tg3 .txc2 If 18...'i'e6 at once, then 19 .ixf5 'i'xf5 20 lLlh4, preparing f2-f4. 19 'i'xc2 'i'e6 20 lLld4?

:ff8

124 Open Spanish

Better is 20 ~h4! ':d7 21lbd4lDxd4 22 cd, threatening t2-f4-fS and

':cl. ltlxd4

20

21 22

cd

.:ca

f4

White should have removed his queen from the c-file with 22 "d3 (22 "d2 cS 23 dc ':xcS is not so clear) 22 ...~b4 (22...cS 23 dc ':xcS 24 f4) 23 ':f1! and t2-f4. 22 cS 23 fS (97) 97 B

23 ••• cd! White has exchanged nearly all the minor pieces on the assumption that he will acquire a dangerous kingside initiative. However, the intermediate capture on d4 leads to the exchange of queens and thereby reduces White's attacking potential.

24"dJ 25 ~tl 26 ':e2

"00

"c2

"f3

The game also remains level after 26 ~xd4 ~cS 27 ~xd4+ 28 l:xd4 "xb2 29 "xdS "c3 and .....CS. The game concluded: 26.....xd3 27.:xdJ + 28 ~b2 g6! (this break ensures a successful defence) 29 g4 (the position is also approximately balanced after 29 fg fg 30 ':xd4 ':d8) 29••• ~g7 30 ~xd4 gf31 gf.:n 32 ':g3:XC5 33 e6 f6 34 ~cS ':e5 35 ':eg2':g5 36 e7 ~ 37 b4 ':xgJ 38 ~g3 f5 (with further simplification, the draw comes closer) 39 ~4 ~f6 40 ~5

.:ct

~xb4 41':16 ~xe7 42 ~d4 ':g8 43 l:xb6 ':g5+ 44 ~4 ':1645 .:xg6 ~xg6 46 ~e5 ~ 47 ~xd5 ~g5 48 ~oo 112·112

4 Marshall Attack

Game No. 20

Short-Pinter

Rotterdam 1988 In the Marshall Attack (as it tends to be called, rather than Counter-Attack), Black sacrifices a pawn on move eight, attempting to gain the initiative in return. An experienced player with White is able to give back the pawn and preserve a minimal advantage. For that reason, Black generally prefers more solid systems. But such overall considerations are not of course taken into account by lovers of adventure. The Marshall Attack, which features in our next three games, remains highly topical and has a good many adherents even at grandmaster level.

1 e4 eS 2 00 iLlc6 3 j,bS 864 1.84 7

8

m 5 0-0 j,e7 6 :tel bS 7 j,b3 0-0

c3

Among White's many possibilities for avoiding the Marshall, the most significant are 8 d4, 8 d3 and 8 a4. But according to theory Black has no difficulties in any of these lines. 8 ... dS 9 eel iLlxdS The old continuation 9 ...e4 is hardly ever seen in present-day tournaments; Black fails to obtain adequate compensation for the pawn. 10 iLlxeS iLlxeS l1:XeS c6 (98) 11...iLlf6, as played by Marshall himself, has long since gone out of use. The fianchetto 11...j,b7 is less thoroughly investigated. In the last quarter of a century, the modest 11 ...c6 has had no rivals.

12

d4

The more restrained pawn advance 12 d3 will feature in Game No. 22. After a period of disfavour, the exchange on d5 is just coming back into fashion thanks to Hubner's efforts. The game Hubner-Nunn,· Skellefie! 1989, proceeded as follows:

126 Marshall Attack

98 W

12.Ld5 cd 13 d4 i.d614 :'e3 'fIh415 h3 (99)

15.. ./5 Several games have ended in a draw after 15 ...'fIf4 16 :'e5 'fIf6 17 :'e3 "'f4. But White can bring his rook back to the first rank: 17 :'el "'g6(or 17 ...i.d71Si.e3 "'g619 ...n :'aeS 20~2h5 21 ~hl J.bS 22 J.f4 "'c2 23 J.xbS 'fIxd2 24 i.e5? f6 25 :'adl "'g5 26 J.d6 :'xel + 27 :'xel :'eS 2S :'xeS+ J.xeS 29 "'e2 i.g6 30 ~h2 'fIf5 with a probable draw, A.Sokolov-Geller, New York 1990. At move 24, the immediate 24 :'edl! would have given White the advantage) IS ~hl i.f5 19 J.e3 J.c2! 20 "'g4 'fIxg4 21 hg f5 22 gf :'xf5 23 ~gl :h5 24lOd2 J.h2+ 25 ~hl i.d6+ 26 ~gl J.h2+ 27 ~hlllz-11z Van der Wiel-Nunn, Amsterdam 1990. Another quite good line for Black is 15 ... g5 16 'fin i.e6 17 "'f6 :'feS 18 ~ 'fIbS 19 J.d2 i.e7 20 'fin 'fIg6 21 :'ael g4 22 "'g3 gh 23 gh J.d6 24 'fIxg6+ hg 25 ~2 ~g7 26 .!l::!b4 J.xb4 27 cb :'h8, as in Hiibner-Nunn, Haifa 1989, which was quickly drawn. In Anand-Nunn, Wijk aan Zee 1990, White tried 16 b3 f5 17 'fin i.b7 18 :'e6? (the

Marshall Attack 127 correct line is 18 J.a3! g4 19 'iVe2 f4 20 J.xd6 fe 21 J.xfS 'iVxf2+ 22 'iVxf2 ef+ 23 ~xf2 l:txf8+ 24 ~gl gh 25 gh J.c8 26 ~2 J.xh3 with equality - Nunn), which led to advantage for Black after the reply 18 ... l:tae8!. Now 19 l:txd6 would be bad for White: 19 ...l:te1+ 20 ~h2 l:txcl 21 'iVe3 'iVf4+ 22 'iVxf4 gf23 l:te6 ~ 24 l:te5 ~f6, with ...l:tf8f7-e7 to follow. The game actually went 19l:txe8 l:txe8 20 ~n g4! 21 'iVxf5 gh 22 'iVxh3 'iVxh3 23 gh J.c8 24 J.e3 J.xh3+ 25 ~e2 J.g4+ 26 ~d3 J.f5+ 27 ~e2 l:c8!, and White had a difficult ending. 16'iV/3 de Firmian-I.Sokolov, Biel1989, proceeded entertainingly: 16 ~2 f4 17 l:tel J.xh3 18 'iVf3 J.xg2 19 'iVxg2 'iVh5 20 ltlf3 l:f6 21ltlg5 l:f5 22ltle6 ~6 23 J.d2 l:te8 24 'iVxg7+ 'iVxg7 25 ltlxg7 l:xel + 26l:txe 1 ~xg7 27 l:te6 l:f6 28 l:txf6 1/7.- 1/7.. 16... J.b7 17 CDd2 g5?! The right continuation was 17 ...f4 18 l:e6 l:tad8. After the move played, Black would gain advantage from either 18 l:te6l:ad8 19 'iVe2 f4 20 ltlf3 'iVh5 21ltlh2 'iVf7 22 'iVel J.c8 23 l:e2 J.f5, or 18ltln l:f6 19 'iVe2 ~ 20 J.d2 f4 21 l:d3 :eg. But Hubner doesn't move away with either his knight or his rook (they will do nicely where they are!); he withdraws his queen to e2 at once. 18'iVe2!f4 In this situation it would have been better to push the g-pawn. 19 ~ 'iVh5 (100) After 19...'iVh6 20 l:e6 l:f6 there would have been plenty to play for, but Nunn falls into a sly trap.

20lDxg5! An unexpected stroke. Since capturing on g5 fails to 21 l:g3, Black remains two pawns down.

128 Marshall Attack

:xes

20..."ikg6 21 :e6 "YixgS 22 fZ.xd6 :ae8 23 :e6 g;P 24 :eS 2S de ~e6 26 i.d2 "YixeS 27"Yid3 "Yig7 28 :e1 + g;d7 29 f3 ~ 30 "fIdJl "fIp 31 a4 ba 32 "fIxa4+ g;c733 fib4 g;d8 34 J.x.t41-0 It is also worth mentioning the manoeuvre 12 "Yin, which brings the queen closer to the king in good time. It looks rather artificial, though. Nor does an old move of Fischer's, 12 g3, present any danger to Black. Let me give one example. Braga-Geller, Amsterdam 1986, continued 12...i.f6 (since the d-pawn has kept back, Black organises pressure against d4) 13 :el c5 14 d4 i.b7 (this is a novelty; after 14 ...cd 15 cd i.b7 16 ~3 lLlxc3 17 be White is slightly better, but now the c3 square is occupied and White lags in development) 15 dc :e8 16lLld2 lLlxc3! 17 be .txc3 18 c6 i.xc619 :Xe8+ "ikxe8 20 :bl :d8 21 "fIc2 (if 21 :b2, then 21...a5!. On 21 "ikh5, Black has the pretty variation 21... g6 22 "figS "fIe1+ 23lLln "fIe4 24 "fIxd8+ g;g7 25 f3 "fIxf3 26lLle3 ilhl+ 27 g;f2 "fIel mate) 21...i.xd2 22 .txd2 i.e4 23 i.xt7+ g;xt7 24 ilb3+ i.dS. Black has an endgame advantage, which he easily exploited to win. 12 ••. i.d6 13 :el Recently the alternative retreat to e2 has often been seen. A survey of the material on this theme will be given in the notes to Game No. 21.

13 14

g3 IS i.e3 The well-known continuations 15 i.xdS and 15 "fId3 are not dangerous to Black, and hardly ever occur today. But White does have one other manoeuvre available - the diverting 15 :e4!? (101). 101 B

Marshall Attack 129 Now IS ....1f5? loses to 161lh4. In the event of 15....!ill616lZh4 .fS 17 .1f4!, or 15 ....1d7 16 c4! bc 17 .1xc4:ae8 18lDd2lCJf6 19l:h4!, White has a clear plus. From the diagram Black used to play IS ...•d7 16lCJd2 .!ill6, which gives White the better chances. But not long ago, ISokolov introduced two innovations in the same tournament - Wijk aan Zee 1991. Nunn-Sokolov went 16....1b7!? 171lel (17.1xdS cd 18 l:el l:ae8 19.!illl fS 20.1f4 is more accurate) 17 ...cS! 18lCJe4 c4! 19 .1c2 .1e7, with double-edged play. Adams-Sokolov went 16...f5 17 l:el f418lCJe4?! (it was imperative to play 181irh5 first, and only then lCJe4) 18...1Irh3 19 .e2? (19lCJg5 .fS 20 lCJe411rh3 21 ~gS would have drawn) 19....1g4! 20.n1irh5 21 .1dl .1xdl 22 l:xdl fg 23 hg (102) 102 B

23 ...l:xf2! 24 ~xf2 l:f8+ 25 ~g2 l:xn 26 l:xfl .e2+ 27 ~ .1xg3!, and Black won. Another fascinating possibility for Black after 15 l:e4 is IS ... g5!? The g-pawn is immune because of 16...•fS!. The game TimmanIvanchuk, Linares 1991, saw some sharp play: 16.f3 (16 . n has also been played, but the extravagant thrust with the g-pawn cannot be refuted like this; Black secures a comfortable game simply by exchanging queens) 16....1fS 17 .1xdS (17 .1c2 .1xe418 .1xe4 .e619 .1xgS fS! 20 .1d3 h6 21 .1h4 l:a7! 22lDd2 l:g7 23 ~n f4 is good for Black) 17 ... cd 18 l:e3 .ie419l:xe4de20.f6.g4 21lDd2 (21.xgS+ .xgS 22 .1xgS fS 23 lCJd2 l:ae8 24 l:el ~f7 25 ~1 ~g6 is in Black's favour; Blackstock-Radovici, Hastings 1971n2) 21...l:ae8 22lCJn.ie7 23 .xa6 f5 24 .xbS f4 1/2- 1/2. .ig4 15

16

.d3

he8

130 Marshall Attack

Black can also play 16.. .f5 first, and answer 17 f4 with 17...:aeS. Sometimes Black refrains from bringing his rook to e8 at all. The interesting game Sax-Ehlvest, SkellefteA 1989, went as follows: 16...f5 17 f4 Wh8 18 -*.xdS cd 19 tbd2 g5 20.f1 'it'h5 21 a4 (Ehlvestdid better in his game with A.Sokolov in Rotterdam 1989, in which White took on g5 at once: 21 fg f4 22 -*.xf4 :xi4 23 gf:t"8 24 :e5 -*.xe5 25 29 h3 'it'h4 30:f1 -*.f5 31 de h6 26 :el hg 27 f5:xi5 28 .e3 :xf1+ 32 ~xfl-*.xh3 33 e6 .g4+ lh-lh) 21...ba 22 fg (or 22 c4 :ab8 23 :abl -*.h3 24 .g4 25 cd gf 26 -*.xf4 -*.xf4 27 .xf4 .xf4 28 gf :b4, and Black had a good game in Short-Ehlvest, SkellefteA 1989) 22... f4 23 -*.xf4 :xf4 24 gf:f8 (103)

.f2

.d3:n

103 W

25 :e5 -*.xe5 26 de h6 27 .xa6 hg 28 .d6 :xf4 29 :f1 -*.f5 30 .xdS -*.h3 31 hf4 gf 32.0 (the storm has died down; White remains with a material plus, and gradually eliminates all dangers) 32...•g5+ 33 Whl 'it'h4 34 .a8+ Wg7 35 'it'b7+ Wh8 36 'it'b8+ ~h7 37 .c7+ Wh6 38 .c6+ Wh7 3900 .g4 40 .e4+ Wg7 41 ~gl .f5 42.xf5 -*.xf5 43 ~2 -*.e4+ 44 Wgl Wg645 e6 Wf6 46 ~xf4 We5 47 e7 -*.c6 48 ~2 Wd6 49 ~4 -*.d7 50 ~5+ 1-0. The exchange 16...~xe3 is not good for Black, for example 17 :xe3 c5 18.f1 'it'h619 ~d2 :ad8 2000 -*.xO 21 :xf3 cd 22 cd .d2 23 :d3!, and White's chances are better; Fischer-Donner, Santa Monica

1966. 17

1012

:e6 (104)

The immediate 17 .. .f5 leads to quite different variations. Judging from overall results, this pawn advance too gives Black sufficient counterplay. Let us look at the evidence. Dolmatov-Vladimirov, Moscow 1989, went 18 f4 Wh8 19 -*.xdS cd 20.n 'it'h5 21 a4 (A.Sokolov-Nunn, Rotterdam 1989, concluded

MarshLzll Attack 131

entertainingly: 21 'ii'g2 g5 22 'ii'xdS :d8 23 'ii'c6 gf 24 Axf4 Axf4 25 gf Ae2 26 Whl :de8 27 :gl 'ii'h4 28 'ii'g2 :g8 29 'ii'c6:gf8 30 'ii'g2 :g8 31 'ii'c6 112_112) 21. ..ba (ECO considers the main line to be 21...g5 22 ab ab 23 :a6 gf 24 Axf4 .txf4 25 :xe8 :xe8 26 'ii'xf4 :el+ 27 ~fl 'ii'e8 28 :b6 Ah3 29 l:b8 :xfl+ 30 'ii'xfl, with advantage to White; Ramirez-Velasquez, Nice 1984. White also does well out of 23 fg :xe3 24 l:xe3 f4 25 gf .txf4 26 :g3 'ii'xg5 27 Whl Ad6 28 'ii'g2 Axg3 29 'ii'xg3 h5 30 l:fl l:xfl+ 31 ~xfl 'ii'c1 32 Wgl 'ii'xb2 33 h3 Adl 34 ~e3, and the white knight is obviously stronger than the enemy bishop; Timman-Nunn, Brussels 1988. The capture on a4 alters matters by deflecting the rook from the first rank) 22 :xa4 g5 23 :aal (before capturing on g5, White has to bring his rook back. If 23 fg, here is what can happen: 23 ...:xe3! 24 :xe3 f425 gfAxf4 26 :g3 'ii'e8!, exploiting the fact that the rook is still on a4. Instead, winning the queen at once - 26 ... Ae3+ 27 :xe3 :xfl + 28 ~xfl - is pointless. After 26...'ii'e8, Ulmanis-Van der Heiden, corr. 1986, went 27 :xg4 Ae3+ 28 Wg2 l:xfl 29 ~xfl .tc1! 30:a5 'ii'e2+ 31 Wg3 'ii'xfl and wins) 23 ...:e6 (in Ph.Schlosser-Nunn, Krefeld 1986, the players agreed a draw after 23 ... a5 24 fg f4 25 .txf4) 24 fg :re8 25 'ii'f2 f4 (Black sacrifices a third pawn so as to get at the enemy king. The ensuing sharp play finally leads to a peaceful result) 26 gf h6 27 ~1 hg 28 ~g3 'ii'h3 29 fg:US 30 'ii'g2 'ii'h4 31 Af2 :xel+ 32 :xel.tO 33 'ii'f! Ae4 34 :xe4 de 35 'ii'g2:O 36 g6 Wg8 37 ~fl 'ii'f6 38liJd2 l:d3 39 ~fl :f3 40 ~ :d3 41 ~fl 'ii'f5 42 ~3 'ii'f3 43 'ii'g5 :d1+ 44 ~xdl 'ii'xdl + 45 Wg2 'ii'f3+ 46 Wfl 'ii'd3+ 112-112. Aside from the traditional continuations 17 ...:e6 and 17 ...f5, a novelty was introduced in lvanchuk-Adams, Terrassa 1991, when Black immediately withdrew his queen with 17 ...'ii'h5. Ivanchuk played two timid moves: 18 ~1 (better 18 a4) 18...:e6 19 .tdl (again 18 a4 was better), and after 19 ...f5 20 Axg4 'ii'xg4 21 Ad2 :g6 22 Wg2 f4 23 0 'ii'h5 24 g4 'ii'h4! 25 :e2 :xg4+! 26 fg f3+ 27 Whl fe 28 'ii'xe2 Wh8 White's position was very difficult in spite of his extra pawn. Even Ivanchuk, though such a connoisseur of the Marshall, shortly had to concede defeat.

18

84

The most thematic continuation. The alternatives 18 c4, 18 Adl and 18 'ii'fl are unpromising. Now Black can either take on a4 or postpone this capture for a couple of moves. 18 ... fS The traditional march of the f-pawn. But sometimes Black retreats

132 Marshall Attack

104 W

his queen at once with 18 ......h5, without waiting for the enemy queen to be re-positioned with "'d3-fl. Here are some examples of this. Chandler-Nunn, Hastings 1987/88, continued 19 ab ab 20 lDfl (in Hellers-I.Sokolov, Haninge 1989, White withdrew his bishop at once with 20 .td1, and there followed: 20....txd1 21 ':axd1 f5 22lDfl f423 .tc1 ':ef6 24"'e4 ~h8 25 'iVd3 h6 26 b3 b4 27 cb fg 28 fg .txb4 29 lDd2 ':12 30 h4 1Dc3, and Black won) 20....:fe8 21 .tdl .txd1 22 "'xd1 "'f5 23 .td2 ':xel 24 .txel h5 25 h4 c5 26 b3 cd 27 "'xd4 ':e4 28 "'d3 .tc5 29 ':a8+ ~h7 30 ':e8 lDxc3! 31 ~g2 ~g6 32 ':xe4 "'xe4+ 33 'iVxe4+ lDxe4 34 f3lDd6 35 .t12lDb7 36lDd2 ~f5 112_112. In Karpov-Short, Tilburg 1989, the Marshall Attack was played against me for the first time in my lengthy career. I introduced a novelty but failed to acquire an opening advantage. I shall now give the game in full: 19 ab ab 20 lDf1 .tf5 Not20... f5?21.i.f4! ':xe122':xel.i.xf4 23 gf.i.f3 24lLlg3 "'g4 25 ':e3 .te4 26 £3. Another inadequate line is 20....tf3 21 .td1 ':fe822 .td2!? ':xel 23 .txf3 "'xf3 24 ':xa1 25 b3. 21"'d1 In Ivanchuk-I.Sokolov, Bie11989, White preferred 21 "'d2, and the players' resources were quickly exhausted: 21.. ..:re8 22 .i.xd5 cd 23 .tf4 ':xe1 24 ':xe1':xe125 "'xe1.i.e4 26lDd2 .i.xf4! 27lDxe4 de 28 gf'iVg4+ 29 ~fllh-lh. A game Rubinchik-Vitomskis, corr. 1989-91, went 21 "'d2 .te4! (another novelty, adequate for obtaining counterplay) 22 .tc2 f5 23 .i.d1 'it'h3 24 f3 f4 25 fe fg! 26 "'g2 gh+ 27 ~h1 "'xg2+ 28 ~xg2 ':xe4 29 .tb3 h1"'+ 30 ~xhl':h4+ 31 ~g2 ':g4+ 32 ~hl ':h4+ 33 ~g2, and this game too ended in perpetual check.

"'xf3

Marshall Attack 133 But then, the new move for White which I chose against Short doesn't achieve much either. 21 ... i..g422 .d2 'fIh3 23 i..d1 i..xd1 More precise than 23 ...:fe8 24 f3 i..fS 25 i..d2, with the better chances for White. 24 Lxti1 f5 25 f4 Not allowing ...f5-f4, as White did in the previous example. 25...g5

An equal game similarly results from 25 ...:fe8 26 i..f2 .g4 27 :xe6 :xe6 28 :el i..xf4. 26.g2 If 26 fg, then the typical thrust 26... f4! is good. Ljubojevic-Nikolic, Belgrade 1991, concluded 27 i..xf4 i..xf4 28 gflLlxf4 29lLlg3 .g4 30 :xe6lLlh3+ 31 ~g2lLlf4+ 32 ~gllLlh3+ 33 ~g2 112-112. 26....xg2+ 27 ~xg2 :fe8 More accurate than 27 ... gf 28 i..xf4 :xel 29 ltxel i..xf4 30 gf lCixf4+ 31 ~O, with the better ending for White. 28 i..d2 LeI 29 LeI LeI 30 LeI gf31 ~f3 fg 32 hg ~p 33 b3 ~e6 34 e4 be 35 be lCif6 36 i..d2 h5 37 i..J4 i..b4 38 ltJd2 i..xd2 39 i..xd2lCie4 40 i..b4 ~f6 41 ~f4 ~e6 ~-~. After 18 ...'fIhs 19 ab ab, apart from 20 lCin and 20 i..dl, White has sometimes played 20 .n, as in the following two games which quickly reached a peaceful conclusion. Sax-I.Sokolov, Haninge 1989: 20 ... i..h3 21 i..dl (or 21 .e2i..g4 22 . n i..h3 23 .e2 i..g4 112-112 Sax-NuRD, Reykjavik 1988) 21.. .•f5 22 .e2 c5! 23lCin cd 24 cd lCib4 2S :a3!ltJc6! 26 :d3 i..b4! 27 d5 :d6 28 i..d2 :xdS 29 i..xb4lCixb4 30:0 .d7 31 i..b3 i..g4! 32 i..xdS .xdS 33ltJe3! .xO 112-11'1.. 19.0 ....S 20 f4 ba

21

:xa4 (105)

Such is the extent to which the Marshall has been investigated: you can reel off twenty-odd moves without breaking new ground. In place of the last move, White occasionally takes his time with the capture on a4, and plays 21 i..xd5 cd 22 .g2. Here is one illustration. Tseshkovsky-Agapov, Kiev 1984: 22 ...:fe8 23 .xdS ~h8 24 i..f2 i..e2 25ltJc4 (25 ~g2 has been played; so has 25 :xa4 i..xf4! 26 gf i..c4, and now if 27 .xc4 then 27 ....g4+ 28 ~hl :xel+ with equality, or if 27 :xc4 then 27 ...:xe1+ 28 i..xel :xel+ 29ltJil :xf1+! 30 ~xf1 .dl + with a draw. Instead, White plays a new move which does

134 Marshall Attack

not tum out well. Directing the knight to e5 looks tempting, but the weakening of f3 is more significant) 25 ...J.c7 26 lDes J.xe5 27 :xe2 (27 de allows the striking finish 27 ...J.f3 28"c4 "xh2+!) 27 .....xe2 28 de "xb2 29 :xa4 "xc3 30 :c4 "al + 31 ~g2 h6 32 "d3 1fa2 33 :c7 34 :c2 :b2 35 :xb2 "xb2 36 "xa6 "c2, and Black soon won. 21 ..• :b8 Lilienthal's move. Alternatives are 21...:fe8 22 ~h8 23 J.xdS cd 24 c4!, and 21...g5 22 :xa6 ~h8 23 :xc61Oxe3 24 with advantage to White. In the latter case Black can play more accurately with 22... gf 23 :xc6 (not 23 J.xdS? cd 24 J.xf4 :e2! 25 :xe2 J.xe2 26 "g2 J.xf4 27 "xdS+ ~h8 28 :a8 J.e3+ 29 ~g2 :xa8 30 "xa8+ ~g7, and Black won; Yagupov-Purgin, Moscow 1991) 23 ...:b6 24 J.xdS+ ~h8 2S "g2 fe 26 :xe3 f4 27 gf J.xf4 28 :g3 :xc6 29 J.xc6 J.xg3 30 hg "fS 31 J.e4 'ifbs 3210n J.e2 331Od2 J.g4, drawing. 22 J.xdS cd (106)

:b6

"f2"f2

Marshall Attack 135 At this point White innovated by capturing the pawn on a6. But before going further, let us consider the well-known continuation 23 'ifg2 "e8 24 'ifxdS ~h8 (Black also has 24 ...':b5 25 'ifa2 ':xb2 26 'ifxb2 ':xe3 27 ':eal ':e2 28 'iVb3+ ~f8 29 'ifxdS 'ife3+, with unclear play; Yagupov-Zhuravlyov, Moscow 1990) 25 ~f2. Some reference books now give 25 ...g5 26 ha6 ':xb2 27 ':a2 gf 28 gf ':xa2 29 'ifxa2 J.xf4 30 lDfl ':e4, with plenty of play for Black. ECO suggests 27 ':a8, but this is refuted by 27 ...J.b8!; Black avoids the rook exchange and is rid of the pin on the sixth rank, while maintaining all his threats. However, White has the more logical 27 'ifaS! ':b8 28 "c6!. Now 28 ...gf leads by force to a position with opposite bishops but two extra pawns for White: 29 "xe8+ ':bxe8 30 gf J.e7 31 ':xe6 J.h4+ 32 ~g2 ':xe6 33 dS J.xel 34 de .txd2 35 J.xd2. Nor is 28 ...'iVh5 any better: 29 h3! 'ifxh3 30 'ifg2 gf31 'ifxh3 J.xh3 32 gf':g8 33 ~e2 ':ge8 34 ~f2 ':g6 35l:hl, and White won; Popolitov-Trushchakov, corr. 1980-81. Does it follow, then, that by playing 23 'ifg2 from the last diagram, White gains the upper hand? No. The opening monographs have simply and unjustifiably ignored the obvious-looking 25 ...hb2 (instead of 25 ... g5). Play may then continue: 26':a2 ':xa2 27 'ifxa2 g5 28 dS! (in Chiburdanidze-Tseshkovsky, Tashkent 1980, a draw was agreed after 28 lBc4 gf 29 lDxd6 fe+ 30 he3 'iVh5 31 'ifxe6 'ifxh2+ 32 ~1 'iVhl+) 28 ...':xe3! 29 ':xe3.tc5 30 lDfl (107).

Until recently, this position was assessed as favouring White: 30.....e4 31 'ifd2 gf 32 gf'ifxf4+ 33 ~g2 J.xe3 34 'ifxe3 "xe3 35 lDxe3 etc. But Black was wrong to waste a tempo moving his queen, thereby giving White a breathing space. Instead, 30....th3! sets White some serious problems; his best reply seems to be 31 'ifxa6 J.xfl 32

136 Marshall Attack 1ff6+, and White forces perpetual check. So Lilienthal's 21...l:tb8, if met by 22 J.xdS cd 23 1fg2, leads to equality, which from the theoretical point of view is a success for Black.

23 lba6 :be8 In this case, capturing on b2 is too dangerous: 23 ...:xb2 24 "g2

1fe8 2S 1fxdS ~h8 26lbc4!. 24 ""5! An important strengthening of White's play. In the event of 24 1Ig2 l:txe3 2S :xe3 :xe3 26 1IxdS+ 1ff7 27 1fxf7+ ~xf7 28 l:txd6 :e1 + 29 ll)fl J.h3 30 Wf2 l:txf1+ 31 ~e3, Black's chances are better. After 24 1ff2 g5 25 :xd6! l:txd6 26 fg :de6, the most likely outcome is a draw with 27 1ff4 J.h3 28 ~f2 :e4 29ll)xe4 fe 30 ~gl l:tf8 31 g4 J.xg4 32 1feS 1ff7 33 1fg3 etc. On the other hand, in Prandstetter-Blatny, CSSR Ch 1986, White played the unsound 27 h4?, which was met by 27 ...f4! 28 gf J.h3! giving Black a decisive initiative: 29 ll)fl J.xfl 30 ~xfl 1Ig4 31.td2 :xe1+ 32 J.xe1 'ii'b3+ 33 ~gl 1Ie6 34 ~1 1fe4 35 1fd2 11f3+ 36.tf2 :e4! 37 g6 h5 38 c4 :xf4 39 1Ie1 'ii'b3+ 0-1. 24 ••• 11f7 (108) Black would lose with 24 ...:xe3 2S :xe3 l:txe3 26 :xd6 :el+ 27 ll)f1 h6 28 1fxdS+ ~h7 29 :d8 'ii'b3 30 1Ig8+ ~g6 31 :d6+ ~h5 32 1ft7+ (Short).

25 h3! Avoiding the trap 25 "xd5? l:txe3 26 1fxf7+ ~xf7 27 :xe3 l:txe3 28 :xd6 l:te1+ 29ll)f1 J.h3. 25 ... J.bS?! Black can play more tenaciously with 2S ...J.xh3 26ll)f3 h6 27ll)eS J.xeS 28 :xe6 :xe6 29 de, though White still has an undoubted plus.

Marshall Attack 137 26 27

1WxclS

.txf4

lhe6

lhe6

28

~

.txg3 Black's position is also hopeless after 28 ...h6 29 .tf2. 29 ~g3 f4 30 1Wxh5 :16 31 ~h2

1-0 An interesting game, which enabled us to present some up-ta-date material on the most popular branches of the Marshall Attack. Game No. 21 Beliavsky-Malaniuk Minsk 1987

1 e4 eS 2li:lDlLlc6 3 .tb5 a64 .ta4lLlf6 5 0-0 .te7 6 l:el b5 7 .tb3 0-08 c3 cIS 9 eel ~cIS 10 ~eS ~e5 11 lheS c6 12 d4 .td6 13 :e2(109) 109 B

Just recently the retreat of the rook to e2 has been rivalling the traditional :e5-el, yet only a few years ago it was dismissed in a couple of lines in openings books. The aim of this rook manoeuvre is easy to understand. White will save a tempo when bringing his queen to f1 (1Wdlf1 instead of 1Wdl-d3-fl); in addition, after g2-g3 and f2-f3, the rook will be switched to the defence of the king. In the notes to the present game, the current state of the variation will be elucidated. 13 ••• 1Wh4 13 ....tg4 has also been seen quite often, for example: 14 f3.thS 15

138 Marshall Attack .ixd5 (Ady-Littlewood, London 1983, concluded 1500 lOf4 16:£2 tOd3 17 :e2 ~f4 112-112) 15 ...cd 16 tOd2 "c7 (after 16... 'fIh4 17lOfl :ae8 18 .ie3, Black's initiative peters out) 17 ~n :fe8 18 .ie3 19 a4. This position arose in Van der Sterren-Pein, Brussels 1984. After 19....ig6 Black retains definite compensation for the pawn, but White too can be satisfied - his kingside is under no threat. In Kamsky-Ivanchuk, Linares 1991, Black played 16...f5 (instead of 16.....c7). After 17 'fIb3 .if7 18 ~ f4 19 .id2 "d7 20 :ael {better 20 a4) 20... a5! 21 a3 a4 22 "dl .ig6, strangely enough Black seized the initiative on the queenside, and eventually won. The correct course at move 18 is to return the pawn with 18 f4!? .ixf4 19 ffi .id6 20 ~e5, when White has the better game. In Kotronias-Nunn, Kavala 1991, Black innovated with 17...:e8!? (instead of 17 ....if7 as in Kamsky-Ivanchuk). After 18:xe8+"xe819 ~n (19 "xd5+ loses to 19....in 20 "xd6 "e3+ 21 ~n :e8 22 g3 .ic4+23~g2"e2+24~h3 g5!) 19....if720.id2f421 a4:b822ab :xb5 23 "c2 .ig6!? 24 "cl .id3 25 .ixf4 .ixf4 26 "xf4 :xb2 the chances were about equal, and the game ended in a draw. In Kindermann-Nunn, Dortmund 1991, White employed the interesting novelty 15 ..n I?~ There followed 15 ....ig6 16 "£2 "f6 (an unsuitable place for the queen; it would be better off on c7 or d7) 17 g3 .id3 18 :el "g6 191012 f5 20 f4! ~h8 21 ~f3.ie4 22 ~ .ixe5 23 de. White has an unquestionable plus, and went on to win. At move 14, retreating the bishop to f5 is worth considering: 14....if5 15 .ixd5 cd 16 ~2 .id3 17 :£2 "c7 18 g3 :ae8 19lOfi .ig6 20 ~3 "d7 21 ~d5 .ixg3 22 ~f6+ gf 23 hg. If now 23 ...:d8 24 .ih6 :fe8 25 ~g2 :e6 26 "d2, the advantage is with White; Kir. Georgiev-P.Nikolic, Wijk aan Zee 1988. The right move is 23 ... 'fIb3, when the game may conclude as follows: 24l:r.h2 "xg3+ 25 :g2 'fIb3 26l:r.h2, with a draw. The queen sortie to h4 is a standard precept in the Marshall Attack, but in Oll-Tseshkovsky, Sverdlovsk 1987, Black introduced an unexpected idea: 13... .ic7! 14lDd2 The prophylactic g2-g3 is impossible in view of....ic8-g4; here we see a drawback to having the rook on e2. After the move played, the black knight will come to f4, with tempo too - another defect of the rook's position. In Lensky-Al. Karpov, corr. 1989, White played 14 .ic2 (14 f3 and 14 ..n also deserve to be tried). There followed 14....ig4 15 f3 "d6

"c4

Marshall Attack 139 16 g3 .thS (16 ....txf3? 17 "d3) 17 :eS? (17 llXI2 was correct) 17....tg61S~2cS! 19.tb3c420.tc2(1l0) 110

B

20.....xe5!! 21 de ~e3 22 "e2 (22 .txg6 ~xd1 23 .te4 :adS, with threats of ....txe5 and ... ~xb2) 22 ... lDxc2, and White was in a bad way; if 23 ltle4, then 23 ... ~xa1 is decisive, while 23 :b1 is met by 23 ....tb6+ 24 ~h1.td3 2S "g2ltle3 26'ifh3 .txb127 ~xb1 :adS.

14.Jj)f415 :e3 Evidently the rook should return to the first rank.

15... c5! 16 ~f3 An improvement is 16ltle4. After 16...cd 17 "xd4'ifh4! IS ~g3 .te6! 19 :e1 :adS, Black had sufficient compensation for the pawn in Klovans-Shulman, Riga 19S5; but Klovans recommends a preliminary exchange of bishops -19 .txe6 fe - and then 20 :e1 :adS 21 "e4, with advantage to White. However that may be, the manoeuvre ....td6c7 could be acknowledged as a valuable piece of pre-game analysis if Tseshkovsky had not thought it up over-the-board, as he himself admitted.

16... .tb717dc ..f6! The white king is feeling more and more uncomfortable.

18"d7l:[ac819"e7 This loses by force; the queen should have taken a different route to g4. 19.....c6! 20 .tc2f5 21 .tb3+ ~h8 22 :e6 The rest is agony, but 22 .tf1 :fdS 23 .td2 ~xg2 24 ~xg2 f4 was no better.

22... lIDe6 23 "xe6"xe6 24 Le6 :ce8 25.td7:d8 0-1. 14 g3 ....5 On the subject of 14...'ihl3, let us start with the game Wahls-Khalifman,

140 Marshall Attack

Hamburg 1991: 1500 .tfS 16 a4 :aeS 17 :xe8 :xeS 18 ~fl b4 (a new move) 19 c4lDf6 20 cS .tc7 21.tc4 as 22 .td2 h5 23 ""3.te4 24 f3, with a slight edge for White. Instead of 16 a4, earlier practice had seen 16 "fl, 16 f3, 16lDe4 and (particularly often) 16 .tc2. In this last case, after 16....txc2 17 "xc2 fS 18f4 "g419~fl.txf420:f2.td6(better20....txc121 :xc1 f4) 21 c4 be 22 "xc4 f4 23 "xc6 "e6 24 .txf4 ~xf4 25 :xf4, the advantage is with White; Sokolov-Khalifman, Sochi 1982. At move 18, the immediate 18 c4 is inferior; in Ljubojevi~-Nunn, Szirak 1987, White was crushed as follows: 18 .....g4! 19 :e6 ~f4!? (19 .. .f4!? is also 21 cb :e2! 22 ~h8 23 "xe2 (the threat good) 20 :xd6? was 23 ...lDh3+; on 23 "f7, Black has the decisive 23 ...:e1+ 24lDfl :xfl+) 23 ... ~xe2+ 24 ~g2 f4! 25 be fg 26 hg ~f4+ 0-1. The game Hiibner-Timman, Tilburg 1987, diverged from Ljubojevic-Nunn with the more accurate 20 f3, and after 20...lDh3+ 21 ~g2 ~f4+ 22 ~gl lDh3+ 23 ~g2 ~f4+ a draw was agreed. It would be very dangerous to try 22 ~hl, in view of22 ...~3 23 gf .txf4 24 :e2 :ae8!. Exchanging the light-squared bishop not for the black bishop (16 .tc2) but for the knight - 16 .txdS - is scarcely more successful: 16...cd 17f3:ae81800 hS 19.te3 (19lDe3ismoreaccurate) 19...h4 20.tf2 .td7 21 gh (21 "d3 is more tenacious) 21...:xe2 22 "xe2 :e8 23 "d3 :e6 24 .tg3 :g6 25 ~f2 (two pawns to the good, White appears to have constructed a solid fortress, but it is quickly demolished) 2S ... b4! 26 a4 ba 27 b3 .te7 28 c4 dc 29 be :b6 30:a2 "f5 31 "xfS .txfS 32 c5 :b2+ 33 :xb2 ab 34lDd2 as 35 c6.tb4 0-1 Griinfeld-Pinter, Zagreb 1987. In the event of 14 ...~3 15 :e4, the distinction between 13 :el and 13 :e2 disappears. For some recent examples of this line, see the notes to Game No. 20. 15 liJd2 White gains nothing from 15 "g6 16 .tc2 fS (or 16....tf5 17 :e2 :aeS 18 .te3 .txc2 19 "xc2 "xc2 20 :xc2 ~xe3) 17 :e2 ~ 18lDd2 f4, with complex play; Elmes-Romanenko, corr. 1981. 15 ••• .th3 White benefits from 15 ....tfS 16 :el "g6 17 ~f3 .tg4 IslDh4 ~5 19 f3, Sax-Pinter, Hungary 1981. The alternative IS ....tg4 16 f3.txf3 17 ~xf3 "xf3 18"fl "g4 19 .td2 :reS led to equality in Balashov-Tseshkovsky, Vilnius 1981. But White has the stronger 18 :f2 "e4 19 "f3 (19 .tc2 "e6 20 "d3 g6 is

"c4+

:ae8

:e4

Marshall Attack 141 not so clear) 19...llae8 20 .i.d2 c!tlf6 21 llel, with the better chances; Sax-P.Nikolic, Plovdiv 1983. It is obviously in White's interest to simplify at the cost of a pawn while keeping the advantage of the bishop

pair. 16 f3 A game played some time ago - Psakhis-Geller, Sochi 1982 went 16 llel? llae8!, and Black's initiative increased conspicuously: 17 f3 f5! 18 c4lbe3! (not shrinking from sacrifices) 19 c5+ ~h8 20 llxe3 llxe3 21 cd f4! 22lbe4 (22 ~n .i.g4! 23 gf llfe8) 22...llxf3 23 .i.xf4 ll8xf4 24 d7 (24lOd2 "e8! 25li.)xf3 "e3+ 26 ~hl .. 27ll)h4 llxh4!) 24 ....i.xd7 25 00 .i.h3! 26ll)xf3 llxf3 27"d2 llf8 28 "e3 "f5 29 "e2 c5 30 .i.c2 "dS, and Black won. (A quicker method was 30.....f6! 31 dc "d4+ 32 ~hl.i.g4!.) If White has decided to withdraw his rook to the back rank, it is better to insert 16 lle4 "g6 and only then play 17 llel. On the other hand, 17 "el is weak; Mokry-Franzen, Supork 1984, concluded 17 ...f5! 18 lle2 (or 18 lle6 llae8! 19 llxe8 :Xe8 20 "dl f4!) 18.. .f4 19lbe4 Wh5 20 f3 fg 21li.)xd6 llxf3 22ll)e4 gh+ 23 ~xh2 .i.f1+ O-l. 16 ••• .i.c7 (111) Black has also played 16...f5, 16...llad8 and 16...llae8. The lastmentioned move looks quite good, for instance: 17 llxe8 llxe8 18lbe4 "g6 19 .i.xdS cd 20 li.)xd6 "xd6 with sharp play; Mithrakanth-Geller; Delhi 1989.

n

16....i.c7 was first played in a game Kuporosov-Malaniuk, USSR 1985, which continued as follows: 17 a4 b4 18 c4li.)f6 19 llel (better than 19 "el llae8 20"n llxe2 21 "xe2 lle8 22"n "fS, with initiative to Black) 19...llad8 20 ll)e4 li.)xe4 21 llxe4 .ifS 22 llel Wh3 (22 ...llfe8 23 .i.e3 Wh3 24 .i.n h5 25 llxe8+ llxe8 26 "d2 h4 27 %leI

142 MarsluJll Attack is not good for Black, but a playable line is 22...J.b6 23 J.e3 J.xd4 24 J.xd4 cS 25 J.xcS :xdl 26 J.xdl :d8 27 J.xb4) 23 :e2 :fe8 24 J.gS f6 25 J.e3 h5 (25 ...J.xg3! 26 hg :xe3 27 l:txe3 "xg3+ 28 ~fl 'ilh3+ would also lead to a draw) 26 c5+ ~f8 27 J.c4 J.xg3 (27 ...h4 28 "n!) 28 hg (at this point, 28 ..n J.xh2+ 29 :xh2 "xfl + 30 ~xn :xe3 31 ~! :ee8 32 :xh5 J.c2 33 :hS+ ~e7 34 :el + ~d7 35 :exeS :xe8 36 :xe8 ~xe8 37 as would have preserved the better chances for White) 28 ...:xe3 29 :xe3 "xg3+ 30 ~f1 'ilh3+ 31 ~gl "g3+ 32 ~n 'ilh3+ 113- 113. In Short-Nunn, Brussels 1986, White improved with 19 llk4! (in place of 19 :el) 19.....g6 (Nunn has shown that 19...~xe4 20 :xe4 and 19.....xf3 20 ~g5 are in White's favour) 20 ~f2 J.fS 21 J.c2 lUe8 22 J.xfS "xfS 23 :xe8+ :xe8 24 ~g2; Black has no compensation for the pawn. But then, the knight excursion to e4 is perfectly playable two moves earlier.

17

~!

The immediate centralisation of the knight proves so unpleasant for Black that Malaniuk, for all his knowledge of the Marshall, lasts a mere ten moves longer.

17

•••

:ae8

In Ehlvest-P.Nikolic, Zagreb 1987, Black decided to take the pawn on 0, but after 17 .....xO 18 ~g5 'iIh5 19 ~xh3 "xh3 20 J.d2:ae8 21 "n! "d7 22 :ael :xe2 23 "xe2 White had a clear plus (two bishops and pressure in the e-file), which he exploited in the ending. That game lasted longer than our primary one, but the result was the same.

18

"d3

:e6(112)

This rook manoeuvre must be considered a failure. The theoretical debate was continued a few months later in the game Ehlvest-Geller,

Marshall Attack 143 Vdac 1987 - a short skirmish containing many adventures: 18 ...f5!? 19 ~g5 (19 00 fails to 19...:xe2 20 "xe2 :e8) 19.. .f4 (19 ...:xe2 20 "xe2 :e8 21 f4 is playable, with compensation for the pawn) 20 ~xh3 fg (the right reply was 20.....xh3 21 g4 hS, with a complex struggle) 21 :xe8! (not 21 ~g2 gb 22 .td2 "g6+) 2l...he8 (Black would lose outright with 21...gh+ 22 ~hl :xe8 23 .tf4) 22 ~g2 gh 23 i..d2 ~h8! 24 ~g5? (this move concedes the initiative to Black; it was essential to play 24 i..xdS cd 25 ~g5, with advantage) 24 ...i..f4! 25 ~ l%e6 26 ~hl.txd2?? (in time trouble, Black forgets to play the move he intended - 26...~3!, which would have won immediately; if 27 "e2 then 27 ...l%xe4, or if 27 .tdl ~ 28 .txe3 :g6 29 "e2 .txe3 there is no defence against mate) 27 .txdS cd 28 ~xd2 "g5 29 f4! ~4 30 "f5 l%e8 31 :f1 h6, and Black resigned. 19.td2 :g6 White has comfortably completed his development, and his opponent undertakes a desperate attempt to stir up complications on the kingside.

"f2

20 But not 20 l%ael f5 21

g4!

00 .txg3! .

:zo ...

.txg4

20...~4 is more stubborn. 21 fg

22 ~g3 23 :g2 24m White's forces have all been brought in to help, and Black has no compensation for the piece minus.

:t6 h5

24 25

26

27

:rJ!

"e2

:tg6

.tel 1·0

So in the main line of the Marshall, there is scope for further work by both White and Black.

Game No. 22 Andrijevi~.M.Pavlovic

Yugoslavia 1988

144 Marshall Attack 1 e4 eS 2 ~ ~ 3 .tbS a6 4 .ta4lDf6 5 0-0 .te7 6 l:el bS 7 .tb3 0-0 8 c3 cIS 9 eel lOxclS 10 lOxeS lOxeS 11 :XeS c6 12 d3 This modest move of the d-pawn (White fortifies the e4 point and in some cases leaves d4 free for a rook) can hardly be better than the standard d2-d4. Nonetheless this ancient move does sometimes emerge as a rival to the double advance of the pawn. 12 .td6 13 l:el (113) JJ3 B

"f3

13 . ... 'iVh4 Another possibility is 13 ....tfS. Sharp play then results from 14 l:e8 15 l:xe8+ "xe8 16lClci2 (after 16 .td21C1f4 17 .tc2 l:d8, Black is better; Whittaker-Harding, corr 1977. The correct line is 17lCla3 "e2 18 .txf4 19 gf .txf4 20 d4 l:e8, with a minimal edge for White) 16.....el+ 171C1fl .tg6 18 h3 l:e8 19 .td1 .th2+ 20 ~xh2 "xfl 21 .td2 "xd3 22 "xd3 .txd3 23 a4 f624 ab ab, which was played in 001matov-Khalifman, Moscow 1990. In Smagin-Geller, Moscow 1989, Black answered 14 with 14 .....d7. There followed 15 .txd5 cd 16.tf4.txf417"xf4.txd318 lDd2 l:ae8 19 l:e3 l:xe3 20 "xe3 .tg6 21 l:e1 h6? (a bad innovation. In an old game Konstantinopolsky-Abramov, corr. 1949, the right method was demonstrated: 21 ... b4! 22 M be 23 "xc3 - White has a slight edge here - 23 .....a4 24 a3 h6 25 h3 .tfS! 26 "c5 27 l:c1 "f4 28 "e3 "d6 Ih-lh) 22lDb3 l:c8 23 a3 "c7 24 h4 a5 25lDd4 26 h5! .txh5 27lDf5:rs 28 "e5 "g4 291C1xg7, with a won position for White. At move 14, the immediate 14lClci2 is interesting; after 14 ...lDf4! 15 lCle4 (15 d4lDxg2!) ls ...1C1xd3 16 .tgS "d7 171C1xd6 "xd6 18 .tc2,

"xf3

"f3

"c4 "c4

MarshaliAttack 145 the game is level. Kir.Georgiev-Nunn, Dubai OL 1986, ended quickly with 17 :e3~xe418:Xe4:ae819"g4 "xg420:xg4.i.eS21 :bl hS 22l:r.b41Oxb2 23 .i.e3lOd3 24 :dllOb2 2S :bllOd3 lb.-lb.. Recently Black has started to refrain from 13 ...~fS in an attempt to exploit his opponent's timid play and make a serious bid for the initiative:

14

g3

....3

This position can arise from various move-orders, for example 12 g3 ~d6 13 :el "d7 14 d3 1Vh3.

15:e4

..rs

The queen retreat to d7 deserves attention, although the game Smagin-Nunn, Dortmund 1991, turned out badly for Black: IS .....d7 16lOd2.i.b7 17 :el (171Of3 cS 18 a4 b4 19lOe5 "c7 20 lOc4 be 21 .i.e7 22 be .i.f6leads to equality; de Firmian-Wahls, BieI1990) 17...cS 18liJe4 .i.e7 19 .i.gS! (more accurate than 19 a4, which Smagin had played against Wessman, Stockholm 1989) 19...:ad81 (he should have played 19... f6 20 ~e3 "c6, or 20 c4 fg 21 cd ~xdS 22 a4 ~h8 23 .i.xdS "xdS 24 ab ab 25 :xa8 :xa8 261Oc3 "d7 27 lOxbS, with somewhat the better ending for White) 20 ~xe7 "xe7 211Vhs g622 "gS f6 231Vh4.i.a8 24 :e2 ~h8 25 :aellOb6 261Vh61Vb7 27 f3! c4 28 lOgS! 1-0. An important improvement was introduced at move 16 in the game Benjamin-Kamsky, USA Ch 1991: 16...fS! (postponing for a few moves the transfer of the bishop to b7) 17 :el ~h8 (and not the premature 17...f4 181Vh5 ~h8 19liJe4 :fS1 20 "xfS! "xfS 211Oxd6, when Black is in trouble-Kamsky) 181VhslOf6191Vh4cS 201Of3 ~b7 21 lOgS h6 22 .i.d2 "c6 23 f3 c4! 24 ~dl cd 25lOe6:n 26 ~g2 ~cS 27 lOgs:ff8 28lOe6 (114)

"ft

114 B

146 Marshall Attack 28 ...lllg4! 29illxfB :xfS 30 b4 1.d6 31 Wh5 Wb6 32 :nille3+ 33 1.xe3 "it'xe3 34 1.b3 "'d2+ 3S:n "'xc3 36 :d1 1.xb4 37 "'g6 d2!. Black has a very strong initiative. and eventually won. In any case. Black need not huny to withdraw his queen. It is worth considering lS ....t.b7 (practice has shown that lS ....t.d7 and lS ...lllf6 fail to equalise) 16illd2 171llf1 cS 18 f3lllf6 19 :b4 "'fS 20 ille3 "'c8. with unclear play; Howell-Hebden. England 1990.

:ae8

16

Illd2

White scarcely succeeds with 16:b4 1.b7 17 .t.c2 "'e6 18 c41llb4 19 cSIllxc2 20 "'xc2 fS!. or 16 .t.c2 "'g6 17"'n fS 18 :e1 f4 when Black has the advantage.

16

...

"'16(115)

Black comes off badly from 16...1llf617 :e1 "'xd3 18ille4! "'xd1 1911lxf6+ gf20 :Xdl. 115 W

Diagram lIS is a kind of standard position in the 12 d3 variation. White has a wide choice of moves. Tunman-Hubner, TJlburg 1985, went: 17~1 This was the first time this move had been played.

17.../5 17 ....t.fS is inferior: 18 .t.xdS cd 19:d4 (this is where the vacantd4 square comes in useful) 19 ....t.cS (19 ....t.e6 20 ille3 .i.cS 21 IllxdS :ad8 221llf4!) 20:XdS .t.g4 21 "'d2 "'c6 22 "'gS .t.h3 231lle3 .t.xe3 24 .t.xe3 h6 25 Wh5 g6 26 'iVe5. and White has an obvious plus. In Hubner's view. Black can improve with 17... lllf6 18 :el 1.g4 19 f3 .t.fS 20 d4 cS 21 de .t.xcS+ 22 .t.e3 :ad8 23 "'e2. but here again White's chances are better. It is worth considering 17...h5!718 a4 .t.g4 19 'iVellllf6 20 :e3 :ae8 21 ab ab 22 d4 h4. with sharp play; KuzminShulman. USSR 1986.

Marshall Attack 147 Let us look at some further moves from Timman-Hubner: 18:ti4 Again White utilises the fact that d4 is unoccupied, but then this is the only move; after 18 :el f4 19 c4 fg 20 fg .tg4, Black has an attack. 18...f4 18 ....tb7 or IS ....te6 would be met by 19 c4. 19:xd5 cd 20 .txd5+ .te621 .txa8 lba8 22 'ilf3 In the event of 22 a4 .tg4 23 'ilb3+ .te6, it seems that White must agree to repeat moves, since 24 'ilc2 gives Black an attack: 24 ....td5 25 ab 'ilg4 26 c4 .to 27 :xa6 :xa6 2S ba 'ilh3 29 ~3 fe 30 fe .txg3, with a strong attack, e.g. 31 c5 .th4 32 c6 'ilg4 33 ~fl .txg2+ winning (Hubner). 22... :j8 23 'fIe4 .tf5 24 'ild5+ ~h8 25 a4 b4? Hubner gives 25 ....txd3 26 ab.te4 27 'fId4 fg 2SIOxg3 .txg3 29 fg ab 30 .tf4 h5 31 'fId6, which leads to equality. 26 cb fg 27 hg .txb4 28 .td2 .txd2 291Oxd2 .txd3 30 :a3 An ending has arisen in which Black has failed to obtain compensation for his pawn. However, thanks to some inaccuracies on his opponent's part, Hubner managed to save himself. Garcia-Bryson, Thessaloniki OL 19S4, proceeded differently (from diagram 115): 17100 f5 lSIOh4 'fIf6 19:d4.te6 20 00.tc5 21 l:[h4 h6 22 d4 .td6 23 .tf4lOxf4 24 gf :aeS. The eccentric position of the rook on h4 gives Black the better prospects. Feher-Hazai, Budapest 19S9, went 17 a4 fS 18:d4 ~hS 19 :xdS (19 ab is more thematic; in reply, Black should examine 19 ...lOf4 20 :xd61Oh3+ 21 ~g2 'fIxd6 22 ~xh3 f4+ etc.) 19... cd 20 .txdS :bS 21 ab ab 22 00 'fIh5 23 'ilb3 f4 241Oh4 fg 25 fg 'fIe2 0-1. We now return to the primary game, which illustrates White's most popular option. 17:el f5 In Smagin-Malaniuk, Kiev 19S6, Black preferred 17 ....tc7. After IS 100.tg4 191Oh4 'ilhs 200 .th3 21 'fIe2! :adS 22 d4 fS 23 f4, he went in for an unfavourable queen exchange: 23 ...'fIxe2 (a sharper line was 23 ....tg4 24 'fIg2 gS 2S fg f4) 24 :xe2 :reS 2S :xeS+ :xeS 26 ~f2 .tg4 27 .te3!, with a difficult ending for Black. The alternative 17....tb7 failed to justify itself in HjartarsonHebden, London 19S6: IS 100 ! :feS 19 :xe8+ :xeS 20 IOh4 'fIf6 21 .td2 b4 22 'flo 'fIxo 23/OxO, with the better chances for White in the endgame. 18 c4

148 Marshall Attack After 18lDe4fe 19de.tg420'ird4 ~5! 21 edc5! White's position is not to be envied, but 18 f4! is quite strong: 18....txf4 (18 ...~h8 19 .txdS cd 20 lOS) 19 'irs .txd2 20 .txdS+ cd 21 'iWxd5+ ~h8 22 .txd2 :a7 23 'irc5 with advantage; Koch-Blatny, Haifa 1989. 18 f4 19 ltle4 fg 20 fg .tg4 21 'iWc2 be 22 de 22 'iWxc4 has also been played, for example: 22 ....tc7 23 'ifc2 (a possibility is 23 .te3 ~h8 24 .tc5 ltlf4 25 .txf8 ltlh3+ 26 ~g2ltlf4+ 27 ~glltlh3+ with perpetual check) 23 ...~h8 24 .txd5 cd 25 ltlf2 :ac8 26 'ifa4 .td7 27 ~4 .td8 28 'iWb4 .tc7 29 ~4 .td8 30 'iWb4 .tc7 1/2- 1/2 HeUers-Wahls, Adelaide 1988. 22 ••• (116) In Smagin-Hebden, Moscow 1986, Black didn't risk giving up a piece and played 22 ....tb4. The 'game concluded as follows: 23 .td2 lOf4 24 .txf4 .txel 25 :xel .tf5 26 'ifg2 :fe8 (26 ...':ae8 is more tenacious) 27 .tc2 ':e6 28 g4! .txg4 29':fl .te2 (29 ....th3 30 'iWxg6 ':xg6+ 31lOg3 .txfl 32 .txg6 .txc4 33 .tb 1) 30 lOgS! .txfl 31 .txg6 .txg2 32 .tf7+ ~f8 33 .txe6 ~e7 34 ~xg2 h6 35 .tg4 hg 36 .txgS+ ~d6 37 b4 1-0.

:ae8

116 W

23 cd .to! 24 dc+ Stronger than 24 .tf4 .txf4 25 'ifxc6 ~5 26 d6+ ~h8 27 d7 ~3! (27 ...':xe4 28 'iWc8!) 28 'iWc2 'iWxd7 29 gf':xe4! 24 ~h8 2S .tdS

Marshall Attack 149 A game A.Ivanov-Agapov, Kiev 1984, ended abruptly with 25 j,d21 :'xe4 26 :'xe4 j,xe4 27 'ilc3 "f5, and White lost on time, though he could just as well have resigned. It might seem that White can extricate himself with 25 j,f4 j,xf4 26ilbf2 (26 c7 j,xc7), but after 26.....xc2 27 j,xc2 .i.e3! his position is scarcely defensible. 25 ••• .hg3! 26 hg lIxe4? Black is over-keen on sacrificing. He had a drawing line in 26... j,xe4 27 :'xe4 :'xe4 28 j,xe4 "xg3+ 29 "g2 "el+ 30 Wh2:'f2 31 j,f4 :'xg2+ 32 j,xg2 'ilh4+ 33 Wgl 'ilxf4 34:'f1 "e3+ 35 l:.f2 'ilcl+ 36 :'f1. I dare say that would have been the rightful result of this game. The game concluded: 27 j,r4 :rxr4 28 'ilf'l 'ii'h5 29 'ii'h2 "xdS 30 If The turmoil has abated, and White is left the exchange up. After some minor adventures he converts his material plus into a win. 30.....xc6 31 lIxe4 'ilxe4 32 "dl h5 33:'c1 "g6+ 34 Wf'l j,e4 35 "d8+ Wh7 36 :'gl At last the white king is in complete safety. 36.....f737 ..g5 37 ~e3 and 38 "d4 would have won at once. White's inaccuracy costs him an extra thirty moves. 37.....87+ 38 ~e2 "f7 39 b3 1.(5 40 :'g3 j,g4+ 41 Wdl "87 42 ~el "d7 43 h444 :'e3 "dl+ 45 Wf'l"d2+ 46 Wgl h3? Black should have picked up the pawn on a2, after which no forcing line is to be found for White. 47 'ile4+ ~h6 48:'d3 "c1+ 49 ~h2"cS 50"e3"c6 51:'dl 1.(5 52 'ile5 53 b4 Wh754 a4 1.16 55 b5 "c156 "e3 .i.h5 57 Wxh3 White's wish to eliminate his opponent's passed pawn is understandable, but again he misses a chance to speed up the win: 57 ba 1.£3 58Wxh3 etc. 57•••8b 58a5! Of course not 58 ab "f1 + and 59.....xb5, when Black's position is impregnable. 58...b4 59 Wh2 b3 60 "d4 .i.g6 61 :'g2 "c7 62 86 .i.c2 63 Wg3 "e7 64 87 b2 65 lIxc2! "el+ 65 ... bl" 66 "d3+ fails to save Black. 66"f'l bl" 67 as....e6 68 'ii'h2+ Wg6 69 (5+ ~5 70 'ilf3+ 1-0

"e5

"cS

5 Bird, Schliemann and Berlin Defences Game No. 23

Kamsky-Ivancbuk 1ilburg 1990 So far we have been examining those branches of the Spanish Game which have proved most popular in the major competitions of the last few years, including World Championship and Candidates Matches. There now remain three vacancies to be filled before we reach the target figure of twenty-five games. There were many candidates for these vacancies. and selection was not easy. I have already mentioned in the preface that I am passing over the Chigorin System, once the most popular of all. The Classical System (3 ...j,cS) is only played regularly by Beliavsky (occasionally by lvanchuk). The Steinitz Defence (3 ...d6) and Steinitz Deferred (3 ...a6 4 j,a4 d6) have today almost disappeared from practice. Black has gained no particular success with the Fianchetto Defence (3 ... g6); at present, it seems to have just one ardent supporter left - Vassily Smyslov. The Exchange Variation (3 ... a6 4 j,xc6 dc) shows little ambition on White's part, and from the theoretical point of view it is harmless for Black. So by process of elimination we arrive at three systems which repeatedly crop up at grandmaster level and are introduced by Black's third move: Bird's Defence (3 ... ~4), the Schliemann Defence (3 ...fS), and the Berlin Defence (3 ...llli6). Our three remaining games will illustrate the most up-todate material in each of these old variations.

1 e4 eS 2 00 ~c6 3 j,b5 3 ...

lbd4

This somewhat extravagant thrust with the knight was introduced into practice as long ago as the middle of the last century, by the English master H.Bird, and was soon named after him.

4

~d4

Retreating the bishop to a4 or c4 is less thematic, though it enables White to avoid sharp variations. A game Romanishin-Malaniuk, Tbilisi

Bird, Schliemann and Berlin Defences 151

1986, proceeded as follows: 4 .ic4 .ic5 5 lOxd4 .ixd4 6 c3 .ib6 7 d4 'fIe7 80-0 lOf6 9 a4 a6 10 .ie3lOxe4 (after 10...d6 11 de de 12 .ixb6 Black's pawn structure loses its flexibility) 11 :el 0-0 12.idS (12 de? lOxf2l13 'fIe2 j xe3 14 'fIxe3lOg4, and White has problems) 12...lOf6 13 de lOxdS 14 'fIxdS .ixe3 15 ':xe3':b8 16lOd2 b51, and Black had no difficulties. Kr.Georgiev-Tseshkovsky, Dubai 1986, went 4 .ia4 .ic5, and now White innovated with 5 b4 (in place of the usual 5 0-0); after 5 ....ib6 6 0-0 lOxf3+ 7 'fIxf3 'fIf6 8 'fIg3lOe7 9 .ib2lOg6 10 c4lOf4 11 :el c6 12 d4, he acquired a plus. Yet Black could have taken the pawn with 5....ixb4; after 6lOxd4 ed 7 0-0, all White can expect is enough initiative to compensate for it. The new move 4 ...b5 was chosen in Kr.Georgiev-Hector, Haifa 1989. After 5 .ib3lOxb3 6 ab lOf6 7 0-0 d6 8 d4 .ib7 9 ':el a6 10 1Oc3 .ie7 11 de de 12 'fIxd8+ hd8 13lOxe5, White had an endgame advantage. 4 ed (117)

So Black has brought about a position in which the bishop on b5 is doing nothing for the moment, and the pawn on d4 is hindering the development of White's queenside. On the other hand, White has gained some time and intends to create a dynamic pawn centre. On both sides, then, there are positive and negative points. 5 0-0 Recently 5.ic4 has begun to be played more often. Then 5....ic5 (the standard reply to 5 0-0) is unplayable, and 5 ...lOf6 6 0-0 lOxe4 7 .ixf7+ ~xf7 8 'fIh5+ g6 9 'fId5+ ~g7 10 'fIxe4 11'f6 11 d3 .ic5 gives White a slight edge; A.Sokolov-Tukmakov, Leningrad 1987. The game Short-lvanchuk, Linares 1989, is interesting: 5.ic4lOf6 6 'fIe2.ic5 7 e5 0-0180-0d51gefdc 10'flh5 (after lOfg ':e811 'fIxc4

152 Bird, Schliemann and Berlin Defences i.d6!, the black bishops are eyeing the white kingside menacingly: 12 d31rh4 13 g31rh3 etc.) lO... b6 11 fg :e8 12 d3 cd 13 cd i.a614 'iVf3 'iVe7 15 i.f4 'iVe2! (after the queen exchange the pawn on d3 is lost, and Black's endgame advantage becomes obvious) 16 ~2 'iVxf3 17lDxn i.xd3 18 :reI i.e2 19 i.xc7 d3! 20 a3 as 21 i.f4:e4 22 i.d2 :ae8 23 i.c3 a4! (Black's positional advantage assumes drastic proportions; White is practically stalemated, and each of his moves leads to new concessions) 24lDg5:c4 25 :adl :c8! 26 :a1 :d8 27lDf3 ':xc3! 28 bcd229:xe2d1'iV+30:Xdl ':xd1+31lDel ':c132':e4fS! 33':e8+ ~xg7 34 ~f1 :Xc3 35 ~e2 ':xa3, and White soon resigned. In place of 5 ... lDf6, Black preferred 5 ... d6 in a game Aseev-Guseinov, USSR 1989. Then 6 c3lDf6 70-0 i.e7 8 ':e1 c6 9 cd d5 lO ed lDxd5111Dc3lDb612d30-0 13 'iVf3! i.f614i.e3i.e615i.xe6fe 16 'iVg4 'iVd7 17lDe4 ':ae8 181Dc5 gave White some advantage. S ... i.c5 Today this move is more usual than 5 ...c6, although in many variations the two simply transpose. After 5 ...c6 6 i.c4 d5 7 ed cd 8 i.b5+ i.d7 9 i.xd7+ 'iVxd7 10 d3 i.c5, we reach a position from LjubojevicSalov, Rotterdam 1989, which continued: 11 ~2lDe7 12lDb3 i.b6 13 i.g5 f6 14 i.d2 as 151rh5+ g6 16 'iVf3 0-0 17 ':fe1lDf5 18 a4lDh4 19 'iVg3lLlf5 20 'iVg4 (it is only now that White plays a new move! The line known to theory was 20 'iVf3lDh4 21 'iVg3 g5) 20... ~g7 211Dc1 ':ac8 22 c3 ':f7 231rh3 h5 24lDe2 g5! 25lDg3! lDxg3 26 'ii'xd7 ':xd7 27 hg ~. The position is level, and the game ended in a draw. Black has a good position after 9 ':e1+ lDe7 lO i.xd7+ 'ii'xd7 11 1rh5 O-O-O!? 12 'iVxn1Dc613 'iVxd7+ ~xd7 14 d3lDb4 15lDa3 ':c8! 16 i.f4lLlxc2 17lDxc2 ':xc2. A game Oll-Guseinov, USSR 1989, saw 9 'ii'e2+ lDe7 lO b3 i.xb5 11 'ii'xb5+ 'ii'd7 12 'ii'd3 0-0-0, which is not so clear.

6

d3 (118)

White may also withdraw his bishop at once. On this subject, it will be interesting to examine Romanishin-Balashov, Erevan 1986 (we shall mention several other important examples in the process):

6 i.c4 d6 7 c3 Fashion changes, and it is hard to decide whether the c-pawn or the d-pawn deserves priority. In Gelfand-Kupreichik, Sverdlovsk 1988, White acquired an advantage with 7 d3 lDf6 8 i.g5 h6 9 i.h4 g5?! lO i.g3 lDg4 11 h3 lLle5 12 i.b3 i.e6 13 1rh5 i.xb3 14 abo Nor is lO...i.g4 11 f3 i.e6 12lLld2 'ii'd7 any better for Black; this occurred in Ghinda-Kotronias, Athens 1986, which continued 13 i.xe6 (in the


winner's opinion, 13 ~b3! .tb6 14 .tf2 c5 15 c3 dS 16 ed .txdS 17 :el+ is even stronger) 13 ...fe 14f4h515 fg~g416 "f30-0-0 17~b3 ~3 18 ~c5 dc 19 .te5!, and White won. We must look for the mistake before move 10. Black's 9 ...g5 seems risky; instead, after 9 ....te6 10~, White has only slightly the better chances. But at move 7, it is worth considering 7 ...'ilh4!. TseshkovskyKlaric, Moscow 1989, continued 8lO
"as

154 Bird, Schliemann and Berlin Defences

8cd j.m4 91ra4+ wj8I01Oc3 White would have had the better chances after 10;'xn c5 (Black does badly with 10...;'e5 11 ;'b3 ;'xh2+ 12 Wxh2lDg4+ 13 Wg3 1rg5 14 n!, or 10...;'xf2+ 111lxf2 Wxf112 d41le8 131Dc3 Wg8 14 ;'g5) 11 ;'b3 lDxe4 121Dc3. IO... c6 By maintaining his bishop in the centre, Black endeavours to hold back the white d-pawn. But the bishop will soon be forced to retreat, and White's pawn mass will become mobile. It was therefore worth considering 10...;'xc3 11 dc lOxe4 12 ;'xf7lDxc3 13irb3, with complexplay. 11 ;'e2 1rb6 12 d3 ;'e6 13 ;'f31:i)g4 14 ,Lg4 ,Lg4 15 h3 ;'e6 16 lDe2 ;'c5 17 d4 1rb4 I81rdI ;'b6I9 b3 f6 20 i.d2 1rb521 ;'e31le8 221lcJ 1rh5 23 1rd2 d5 24 e5 Wg8 25 efgf26 a4 ;'c7 27 J,f4 j.xJ4 28 lDxj4 1rg5 291lc3 Wp 30 Ilg3 1rh6 31 lleI lle7? 32 1rcJ! The 'x-ray' on the cl-a6 diagonal and White's control of the g-file permit him to finish the struggle with a problem-like motif. Black could have staved off the immediate crisis with 31...llhg8. 32... a5 33lDxe6 1rxcJ 341lg7+ 1-0. Another interesting novelty, 6 e5, occurred in Yurtaev-KIari6, Moscow 1989. After 6 ...c6 7 ;'c4 dS 8 ed ;'xd6 9 d3lDe7 10000-011 lDe4lDd5121lel ;'f513 1rn ;'g614lOxd6 1rxd615 ;'d2, White had some initiative. Plaskett-Kupreichik, Hastings 1984185, went 6 b4 ;'xb4 7 ;'b2 lOe7 8 ;'xd4 0-0 9 1rcl c5 10 irb2 cd 11 1rxb4, and Black had an excellent game. At this point he should have played 11...1Dc6! 121ra4 a6 13 d31lb8 14 ;'xc6 dc etc. White also has 6 c3, when 6 ...c6 7 ;'d3!? dS 8 ed 1rxdS 9 b4 ;'e7 10 IlellDf6 11 b5 ;'e6 12 be be 13 1ra40-0 14 1rxd4 1rxd4 15 cd Ilfd8

Bird, Schliemann and Berlin Defences 155 led to equality in Timman-Beliavsky, Tilburg 1986. Before going further with the primary game, let me mention two possible queen excursions on move six. A game Kuzmin-Malaniuk, Moscow 1986, went 6 'ihts 'fIe7 (6....i.b67 'fIeS+) 7 d3llli6 (7 ...c6 8 'figS with advantage) 8 'fIh4 c6 9 .i.c4 (9 .i.a4 deserves to be tried) 9... dS 10 ed (10 .i.b3 de ll.i.gS, with unclear chances) 1O...li)xdS 11 .i.gS f6 12 .i.d2 .i.e6, with equality. In Hjartarson-Malaniuk, Moscow 1987, White played 6 "O!?, and after 6 .. liJe7 7 d3 c6 8 .i.c4 dS 9 .i.b3 0-0 10 ed cd 11 .i.gS f6 12.i.f4 as? (12 ....i.e6 13 h3 would have led to a tense struggle) 13 a4.i.e6 14 lba3 gS IS .i.d2 'fId7 16li)bS, he obtained a considerable plus. 6 •.. c6 After 6...li)e7 7 'ihts! .i.b6 8 .i.gS 0-0 9li)d2, White's position deserves preference. 7 .i.a4 The retreat on the other diagonal is also popular: 7 .i.c4 dS 8 ed cd 9 .i.bS+ .i.d7 10 .i.xd7+ 'fIxd7 llli)d2ll)e7 12li)b3 .i.b6 13 .i.gS, and now Black has either 13...0-0 14 .i.xe7 'fIxe7 IS :el 'fIf6 16 'ihts :ac8, as in Chandler-Wolf, West Germany 1985, or 13 ... f6 14 .i.d2 as IS a4 0-016 :elli)g6 17 h3 :fc8, as in Lanc-Tseshkovsky, Trnava 1986. In both cases the game is level. In place of 9 .i.bS+, White also has 9 .i.b3ll)e7 (120)

Here are some of the most characteristic examples to demonstrate that Black's doubled pawns are more of a strength than a weakness; his opponent's position can be seriously cramped by them. Dvoirys-Balashov, S3rd USSR Ch 1986, went 10 c4 0-0 (1O...dc3 11 li)xc3 0-0 12 :el h6 13 .i.f4 .i.e6 14 d4 .i.b4 IS :e3 :C8 16 :Cllllis 17 :d3 :e8 18 h3 gave White the better game in Nikolenko-Gorelov,

156 Bird, Schliemann and Berlin Defences Moscow 1986) 11 cd ~xd5 12lDd2 ~3 13 fe de 14 'iVh5 ed+ 15 .xc5 dc. 16 :axc1 J.e6 17 J.xe6 fe 18 :xf8+.xf8 19.e3 "f6, with equality. Dvoirys-Kupreicbik, Volgograd 1985, went 1O:e1 0-0 11 'iVh5 (11 lDd2 12 a4 J.b4 is no improvement for White; Anand-Tseshkovsky, Calcutta 1986) 11...J.e6 12 ~2 13 a4 J.b4 14 :e2.d7 1500 J.g4 16.g5 J.xf3 17 gf :a6, with advantage to Black. Markland-Davies, COlT. 1985, went 10 'iVh5 0-0 nlDd2 12 a3 a4 13 J.a2 :a6 14 ~f3 :g6, with a double-edged game. BeJiavsky-Tsesbkovsky, 53rd USSR Ch 1986, went 10 J.g5 f6 11 J.f4 (11 J.h4 0-0 12 ~2 ~h8 13 :el gives equal chances) 11...0-0 (11...J.e6 12 :el J.f113 c3 0-0 14 cd J.xd4 15 ~3 ~g6led to complex play in Thipsay-Kupreichik, Frunze 1985) 12 ~2 ~h8 13:el 14 a4 J.b4 15 h3 g5 16 J.h2 :a6, with good chances for Black.

as

as

as

as

as

7 8

••• f4! (121)

liJe7

Other moves known to theory are 8 'iVh5, 8 J.b3 and 8 c3. Here are two relatively recent games in which 8 'iVh5 d5 was played. Lau-Anand, Moscow 1989: 9 lDd2 0-0 10 ~f3 f6 11 ed ~xd5 12 :el J.d7 13 J.b3 J.e8 14 'iVh4 J.f7 15 J.d2 ~7 1611e4 J.xb3 17 ab ~g6 18 'iVh3. Ermenkov-Zakhariev, Bulgaria 1989: 9 J.g5 de (9 ...0-0!? at once is more accurate) 10 de 0-011 J.b3 'iVd7 12 h3 b6 13 ~2 ~g614 %ladl J.a6 15 %lfel. In both cases, White has a minimal edge.

This strong move with the f-pawn occurred for the first time in Kindermann-Tatai, Budapest 1987. Another three moves, and Black's p0sition was critical: 8 ...d5 (8 ...~g6 9 'iVh5!) 9 f5 f6 10 'iVh5+ ~ 11 ~hl! (preparing c2-c3).


In Kindermann-Lorenz, West Berlin 1987, facing the author of the innovation, Black varied with 9 ... g6. After a few more moves - 10 f6 ~g8 11 ed bS 12 .el+ WfB 13 b4! .i.b6 14.tb3 cd IS .i.xdS! .tfS (lS ...•xdS 16.th6+) 16 .i.xaS - Black had to resign. A better decision is 9 ...de, but the game Spassky-Barua, New York 1987, showed that here again Black has considerable problems: 10 de 0-0 11 .i.b3 .i.d6121rhs! d3 (the immediate 12....i.e5 is more precise; then after 13 ~d2 White is only slightly better) 13 cd .i.eS 14 ':f3!. 8 ••• IS An improvement on Black's play in previous games, but this too fails to solve all his problems. 9 .i.b3! In Blatny-Malaniuk, Warsaw 1989, the chances were equal after 9 1rhs+ g6101rh6~g8! 11 .g7 .f612 .xf6~xf613 eS ~g414.i.b3 d6 15 ed Wd7 16lM2 ':e8!. The immediate retreat of the bishop, which was tried for the first time in the present game, is much stronger.

9

•••

cIS

This appears to be the only reply. 10 eel liJxclS 11 ':el+ Wf8 If 11 ....te7, then 12 .e2 is unpleasant. White also has the advantage after Il...Wt7 12 ~ ':e8 131rhs+ Wf8 14 ':xe8+ .xe8 15.xe8+ Wxe8 16 .i.xdS cd 17 ~f3 .i.d7 18 b3 ':c8 19 .i.b2 .i.b6 20 ':cl, winning a pawn (Kamsky). 12 'irhS! White's advantage is clear, and he gradually increases it. 12 g6 13 'irh6+ Wg8 14 ~ .ttS 15 'irh3 .i.g7

16

M

h6

Kamsky recommends 16...Wf7!? 17~!

17...Wh7 is unplayable on account ':xeS, White is obviously better. 18 .txclS+ 19 b3 20 .i.b2 21.f3

.16 (122) of 18 ~t7. After 17....i.xeS 18

.6 cd Wh7

':e8


122 W

22

Wa

22 WxdS is also good.

22 23 24

2S

:XeS i.xd4 i.xb6 i.xf2

i.xeS :XeS :e2

lhf2

26 The queen exchange has reduced Black's defensive chances. 26 ~g7

27

a4

28

as

i.e6 86 ~

29

j,d4+

30

~

:c8

31 32 33

:&2!1

g5

j,eS

~g6

~e3

~bS

34 ~d2 The immediate 34 g3, or 34 ~d4, would have been even stronger. 34 :g8

3S 36 37 38 39 40 41 42

g3 ~e3

~g4 ~b3

Ab2

d4+ :d8+ i.xb3 j,dS

c4

1.12

c3 ~d4 ~e3

:b6!

An accurate move preventing 42 ...~xh2, which would be met by the decisive 43 ~f2! j,c6 44 d4 ~h3 45 d5.


42 ••• J.n 43 lbh6+ He could have won at once with 43 cilf2 J.xd3 44:xb6+ cRg4 45 h3 mate. 43 cRgl 44 d4 g4 45 cS J.c4 46 lU6 1-0 If 46...cRxh2 then 47 cilf2, with the decisive threat of :b6 mate. If 46 ...:h8, then 47 :xfS :xh2 48 cRxg3 49 f5+ wins.

:f8

Game No. 24 TImman-Korcbnoi Candidates Quarter-Final, 3rd game Brussels 1991 1 e4e5 200~c63J.bS 3 •••

IS The Schliemann Defence is one of the sharpest variations of the Spanish, and of course it is impossible to examine all its nuances within the context of a single game. This counter-gambit rarely occurs in grandmaster contests, but I selected it for inclusion because just recently it has cropped up twice at the very top level - in Candidates Matches. In both games the White side was played by Timman, and in the first one - against Speelman - he lost. It may have been for that very reason that Korchnoi too, in his match against the Dutchman, resorted to the Schliemann at a stage when a win was essential. But this time things turned out better for Timman. White's main reply to the gambit, and also his strongest, is 4lOc3. In general, the modest 4 d3 and the straightforward 4 ef are not particularly unpleasant for Black to meet. However, 4 d4 is worth scrutinising in more detail; this very sharp line, involving a piece sacrifice, can be recommended to all lovers of a tactical mel6e. Piskov-Yandemirov, Moscow 1984, went 4 d4 fe 5 ~xe5 (the old move 5 J.xc6, already analysed in the last century, has long since been relegated to the archives; after 5...dc 6 ~ 'lrh4 7 "e2 1Of6, Black has no problems at all) 5 ...~e5 6 de c6 7 lOc3 (if the bishop retreats, Black has 7.....a5+ and 8..."xe5, when White has no compensation

160 Bird, Schliemann and Berlin Defences for the pawn) 7 ...cb 8 ~e4 d5 9 ed W6 (123)

10.ig5 (inadequate alternatives are 10 0-0 ~xe4 11 1Vit5+ g6 12 "e5+ ~n 13 "xh8 ~f6, and 10 "d4 ~xe4 11 "xe4+ ~n 12.if4 "e8 13 .ie5 .ixd6 14 "d5+ "e6 15 "xd6 "xd6 16 .ixd6 :e8+ 17 ~f1 .ifS, with advantage to Black) 1O.....a5+(after 10....ifS11 .ixf6 Sf 121Vit5+ .ig6 13 "xb5+ ~n 14 "xb7+ White has a winning attack, as Bardeleben established as long ago as the turn of the century) 11 1Oc3! b4 (in a game Bushuyev-Chudakov, USSR 1983, White gained a substantial plus after 1l....ie6 12 .ixf6!? gf 131Vit5+ ~d7 14 0-0-0 :c8 15:be1 :xc3 16 :xe6!; if 16....ih6+, White wins with 17 "xh6! :xc2+ 18 ~xc2 "a4+ 19 ~d2 'iVb4+ 20 ~e2) 12 .ixf6 gf 13 00(124)

13 ...b3+ (again 13 ....ie6 is bad: 141Vit5+~d815 0-0-0:g8 16 d7! .id6 17 :bel .ie5 18 :xe5 fe 19 1Vit4+ ~xd7 20 "e7+ ~c6 21 "xe6+ ~b5 221Oc3+ 1-0 Glek-Yandemirov, Moscow 1983; though at move 15, Black had the more accurate 15 ...b3!) 14 c3 .ie6! (this time


the bishop move is quite acceptable) IS!Dc,7+ ~d7 160-0 .txd6 17 18 g3 "xe6, with a double-edged position. In the game, White played 19 :el?! "fl 20 ab :be8 21 :xe8, and now 21...:xe8 22 :xa7 ~c7 23 b4 "e6 would have given Black some advantage. However, it was worth considering 19 ab ~c7 20 :el (20 b4!? is also interesting) 20 .....fl21 "d4, with chances for both sides (analysis by Mikhail Tseitlin and Igor Glazkov). 4 M fe ~xe6

"e5

5

lOxe4 (125)

In the primary game, Black played S...~6 here. Another wellknown move is 5...dS. The latter was played in Timman's other game, which it is now time for us to examine. Timman-Speelman, London 1989: S ... dS 6 ~eS de 7 ~c6 "gS In general, this variation has a dubious reputation from Black's viewpoint; 7 .....dS is played much more often. In preparing for the match, Speelman must have unearthed a game which his opponent played a long time ago against Bohm. Timman was no doubt satisfied with that game, but the English grandmaster discovered something interesting in it. 8 "e2li¥6 9 f4 "xf4 (126) The game Timman-Bohm, Wijk aan Zee 1980, to which I referred, ended quickly with 9 ...~4+ 10 g3 ~3 11 ~+ c6 12.tc4 .tcS 13 d3 ~g4 14 ~f7 .tf2+ 15 ~dl e3 16"f3 ~6 17 "e4+ ~f8 18 .txe3 .tg4+ 19 ~d2 :e8 20 ~ 1-0. Black should of course have taken off the f-pawn, as Speelman does. 10~S+

Sharp play results from 10 ~xa7+ .td7 11 .txd7+ ~xd7 12 d4!?


"f5 13 ""5+ "xb5 14 ~xb5 c6 15 ~3.tb4 16 ~e2 .txc3 17 be :a4; Adams-Lautier, Terrassa 1991. 10... 00 11 d4 "h4+ 12 g3"h3 13 .tc4 .te6 14 .tg5 Another possibility is 14.tf4. 14... 0-0-0 150-0-0 .td6 16 ~fl The ending arising from 16 "flllbe8 17 "xh3 .txh3 18 ~f1 :d7 19 ~d6+ :xd6 20 .tf4:d7 is harmless for Black, for example: 21 libel 00 22.te5 .tg4 23 :d2 e3!? 24 :d3 ~b6 25 .tb3 e2 26 :e3 (26 h3?.till 27 :Xe2 .tfl) 26...a5 27 a4 c5! 28 c3 (28 :c3 :xe5!) 28 ...c4 29 .tc2 00, with a fine game; Ernst-Inkiov, Gausdal 1989. However, 16 g4!? .txc417 "xc4 is interesting, for example 17 ....txe5 18 de :xdl + 19 :Xdl "xg4 20 .te3 ~7 21 ~xe5 22 "e7 ~g6 23 "xg7 :d8 24 :xd8+ ~xd8 25 "xb7 ~7, with about equal chances; Donchev-Inkiov, Bulgaria 1989. 16... .txp 17 .txp :hj8 18 .tc4 Black comes off well from 18 .tb3:de8 19 c4 "g4 20 "e3 W 21 :dfl :xf1+ 22 :xfl h6 23 c5 .txg3 24 hg "xg5; Velimirovic-Klinger, Palma de Mallorca 1989. 18...:de8 19 d5?! Not a very effective move. The d-pawn will restrict the activity of White's bishop. A better line is 19:hfl h6 (19 ...~g4? 20:xt8.txfS 21.tfl, and wins}20.tf4.txf421 :xf4:d8 (2l...:e7! Tseitlin) 22 c3 ~b8, with complex play (but not 22 ...:te8? 23 .tfl! :e7 24 .tg6, with a large plus for White; Yudovich-Boey, corr. 1972-76). 19... c5 20 :hfl ~b8 21 j,{4 :d8 A draw would suit White in view of the match situation, but Black is in a determined mood. 22 .tg5 a6! 23 .tx.f6 gf 24 "xe4 "xh2 25 :hI? A serious mistake. After 25 :n the position is roughly equal, since

"fl

=

Bird, Schliemann and Berlin Defences 163 it is hard to demonstrate the superiority of the 'good' black bishop over the 'bad' white one; 2S ....txg3 can be met by 26.txa6. 25..."xg3 26 f1xh7 :fe8! White's pieces tum out to be most awkwardly placed.

27"/5 If 2711h4, Black has the decisive 27 ..."f3, with the double threat of 28 ....tf4+ and 28 ...:e4; for instance 28:n "e3+ 29 Wbl :e4 etc. On

the other hand, 27 "d3 is very strongly answered by 27 ...:e3 28 :deS! 29 .txa6:el 30 :xb7+ Was, and wins. 27... b5! 28.tfl Or 28 .td3 c4 29 .te4 "f4+. 28... :el 29 "h5 "f4+ 30 Wbl "xfl 0-1. 5

•••

"f1

lM6

6 "e2 The immediate exchange on f6 probably promises White more. 6 ••• cIS White has a lasting initiative after 6 ..."e7 7 0-0 dS 8 lOxf6+ gf 9 d4 e4 10 1Oh4 fS 11 c4!? 7 lOxf6+ If 8 d4 .tg7 8...e4 is met by 91Oh4!. 9 de 0-0 10 e6(127) Theory recommends the preliminary 10 .txc6 be, and only then 11 e6, but after 11...:e8 120-0 :xe6 the game is about equal.

10

••.

:e8

10... lOes is sounder. In the following example, Black is played by Mikhail Tseitlin, a noted authority on the Schliemann Defence.

164 Bird, Schliemann and Berlin Defences Renet-Tseitlin,PalmadeMallorca 1989: 1O...~ 11 0-0c612.td3 (not 12.ta4 .txe6 13lbd4 .td7 14 f4 'iVb6, with advantage to Black; Popovi~-Kurajica, Sarajevo 1985) 12....txe613lbd4 .tg4! 14 f3 .td7 15 ~5?! IOxd3 16 "xd31i'b6+ 17 Whl.txf5 18 "xf5 :ae8 19.tf4 :e2!, and Black won. An improvement for White is II.tf4 c6 12.td3 IOxd3+ 13 "xd3.txe6 14lbd4:e8 150-0-0 "d7 16 h4 :ad8 17 h5 c5 with equality; Psakhis-Grosar, Yugoslavia 1987. It appears that for some reason Korchnoi didn't like the knight move to e5; after forty minutes' thought he moved his rook.

H 12

0-0

"d3

..txe6

In Dvoirys-Morvay, Budapest 1989, White obtained a big advantage with 12 .tf4 .tg4 13 "d3 :e4? 14 .txc7!, but the right reply is 12....tf7 with a complex struggle. 12 "d6 13.tel 86 14 .txc6 14 .ta4 b5 15 .tb3lOe5 161Oxe5 fe is bad for White. 14 "xc6 15 lOd4 "d7 16 lOxe6 "xe6 17:adl :adS 18 .td4 1i'd6 Black has not managed to secure full equality; 18 .....e2 19 "c3 (and if 19.....e7, then 20 :fel) would be unpleasant for him. as 19 c3 After 19...c5 20.te3 b5 21 :d2 :d7 22 :tdl:cd8 23 "f5! "e5 (on 23 .....c6, the white h-pawn advances) 24 "xeS fe 25 .tg5, the pawn on d5 is lost, but F~nik has recommended the sounder move 19...c6!? 20 h3 b6 21.tel cS 22:rel :e7 22...d4 is not better: 23 cd cd 24 "c4+ "e6 25 :xd4 :Xd4 26 "xd4 "xa2 27 "xb6. 23 :e2! "e6?! After 23 .....c6 White would only have kept a minimal plus. Now his initiative is substantial.

24""5 2S :ed2

:d6

Bird, SchliemaM and Berlin Defences 165

2S £xcS "xe2 26 "xe2 lhe2 27 £xd6 l:txb2 28 l:txdS l:txa2 29 l:tbS fS is not so clear. 2S ••• ltd8(128) In time-trouble Korchnoi misses a drawing chance: 2S ...fS 26 'iVb3 d4 27 "xe6+ l:texe6 28 cd cd 29 £xd4 £xd4 30 l:txd4lhd4 31lhd4 l:tel+ 32 ~h2 l:te2. 128 W

The material is equal, but Black's pawn structure is extremely vulnerable, and White strikes a powerful blow at it.

26

M!

ab

Or 26...d4 27 be be 28 "xcS.

27

cb

d4

If 27 ...cb, White doesn't play 28 "xb6 "xb6 29 .txb6 l:tdd7 30 l:txdS l:txdS 31 l:txdS l:te2 with drawing chances for Black, but 28 "xb4! l:td6 29 ""3 l:ted7 30 £f4 with a big advantage. 28 be be 29 "xeS l:tdS 29 ... l:ted7Ioses by force to 30 l:txd4 l:txd4 31 l:txd4 l:txd4 32 £xd4 "xa2 33 "c8+ ~f1 34 "d7+ ~g6 3S "g4+ ~f1 36 'ii'hS+ ~g8 37 "e8+ £f8 38 £xf6 (Timman). 30"c4 l:taS

31 32

"xd4 "d8+

l:txa2

rM7

He can't save himself with 32...l:te8 33 l:txa2lhd8 34lhd8+ ~fl 3S l:ta7+ ~g6 36 l:tdd7 £f8 37 lhh7, or with 32...£f8 33 £h6:n 34 l:txa2 "xa2 3S l:td7!. 33 l:txa2 34 £eS

35

]:tel


1-0

GameNo.2S Short-Gelfand

Candidates Quarter-Final, 5th game Brussels 1991 1 e4 e5 2 00 l2X6 3 .i.b5 3 ••• lM; This book concludes with the so-called Berlin Defence, which may not occur in tournaments too often, but has lasted several decades without disappearing from the scene. It is highly reminiscent of the Open Variation (if the standard capture ...~f6xe4 is made), but there are also some essential differences. Black's omission of the standard ...a7-a6, driving away the 'Spanish bishop' , means that White cannot retain this piece; on the other hand, the black knight on e4 lacks its usual support and has to waste time retreating. More often than not, an early queen exchange takes place, leading to an ending with some advantage to White. 4 0-0 lOxe4 (129) The continuations 4...d6 5 d4 and 4 ....i.e7 5 1Dc3 d66 d4 lead to the somewhat passive Steinitz Defence, while 4 ....i.cS gives the Classical Defence - variations we are excluding from this book. 129

W

5 d4 The most thematic line, and by far the most popular continuation for

Bird, Schliemann and Berlin Defences 167 White.The alternative replies 5 :el and 5 "e2 have virtually gone out ofuse. S ••• lOd6 5 ...£e7 is rarely played here. As an example of it, let us take the second game of my match with Korchnoi, Merano 1981. A well-known theoretical position arose after 6 "e21O:16 7 £xc6 be 8 de li:lb7 9lLlc3 0-0 10 :el lLlc5 11 £e3 lLle6 12 lZadl d5 13 ed cd 1410:14 £d7 15 li:lfS d5 16li:lxe7+ "xe7, whereupon I played the new move 17"d2 (instead of 17 lLla4). Soon afterwards - 17...1Vh4 (l7 ...1Vb4 18 li:lxd5!) 18lLle2lZfe819b3:e7 20li:lg3! "f621 1'3 £e8 22lLle2h623 £f2 "g624lLlcl d4 2510:13 "f6(25 ... c5 26li:lxc5!) 26£g3-Black found himself in a highly dubious position.

6

hc6

6 de leads to sharp play: 6...li:lxb5 7 a4li:lbd4 8 lLlxd4 lLlxd4 9 "xd4 d5 10 lLlc3 c6. In Nuon-Salov, Haifa 1989, White continued with the novelty 11 as (instead of the previously known 11 f4 or 11 £e3), but after 11...£fSI2f4 "d713lZf2h5!? 14£e3lZh615lLla4lZg6B1ack's chances were no worse. The harmlessness of 6 £g5 was demonstrated a century ago in the game Pillsbury-Lasker, St Petersburg 1895: 6...£e7 7 £xc6 £xg5 8 de dc 9lLlxg5 "xg5 10 ed cd 11 lZel+ £e6 12 "xd6 lZd8, with a fine game for Black. Nor does White achieve anything with 7 £xe7 "xc7 8 £xc6 dc 9 de li:lfS 10 lLlc3 £e6 11 "d2 lZd8 12 "f4 0-0 13lLle4 Wh8 14 c3 c5; Noon-Smejkal, West Germany 1989. 6 ... de 7 de 7...lLle4 is played more rarely. In TseshkOVSky-Malaniuk, Alma-Ata 1989, White quickly gained the advantage with 8 "e2 £fS 9 £e3 "e7 10 lZel £g6 llli:lbd2li:lxd2 12 "xd2. 8 "xd8+ cRxd8 (130) The queen exchange has produced what we may consider as a key position - the evaluation of the whole system depends on it. With further piece exchanges, the defects of Black's pawn chain will increasingly make themselves felt. But his possession of the bishop pair does to some extent compensate for the pawn inferiority. Incidentally, the less popular Exchange Variation (3 ... a6 4 £xc6 dc) is characterised by these same nuances. From the diagram, White has three continuations at his disposal- 9 lZdl+, 9lLlc3 and 9 b3 - which often transpose. Let us begin by looking at the interesting game Karpov-Miles, BieI1990:

m


130 W

9~c3We8

There is no point in 9...h6; Yudasin-Rogers, Manila IZ 1990, continued 10 :d1+ We8 11 h3! £e7 (11...£e6 12 g4 ~7 13 ~4 £d7 14 Wg2, with f2-f4 to follow) 12 ~2! g5 (12...£e6 13 ~4 g5 14 ~xe6 fe 15 g4) 13 b3 £e614g4~g715~g3c516~:d817 £a3! £d518 ~f6+ £xf6 19 ef ~ 20 ~5 b6 21 c4, with a big advantage to White. In Dolmatov-Smejkal, Polanica Zdroj 1991, White varied with 10 h3 We8 11 b3 £e6 12 £b2 £b4 13 ~2 as 14 ~4 £d5 15 ~xd5 cd 16 g4 ~7 17 e6 fe 18 £xg7:b7 19 £b2 ~g6 20 ~5 ~f4, and Black held out. 9...h5 is somewhat better: 10 £g5+ £e7 11 :001+ We8 12 ~2 ~6 13 h3 £f5 14:d2 :d8 15 :idl £xg5 16 ~g5 We7 17 ~4 £c8 18 c4 f6 19 ef+ Wxf6 20 h4, and White had a minimal edge in Ljubojevi~-Salov, Linares 1990. Often Black plays 9...£e6. Then 10 b3 £b4 11 £b2 £xc3 12 £xc3 £d5 13 :ad1 We7 14 £b4+ We8 15 ~4 gives White a slight initiative. All this happened in Timman-Gelfand, Tilburg 1990. There followed 15 ...~xd4 16 :xd4 as 17 £d2?! (17 £el! is correct, preparing f4 and £h4) 17 ...£e6 18 :a4 b6 19 b4 :d8 20 £g5 :d4 21 c3 :c4 22 :a3 Wd7 23 ba ba 24 lhaS lhc3 25 :bl?! (25 £e3) 25 ...c5, and Black had solved all his problems. In reply to 9 ...£e6, it is worth considering 10 ~g5 £c4 11 :dl+ We8 12 b3. If play then proceeds 12...£b4 13 £b2 £xc3 14 £xc3 £d5, the position is about level. But an important novelty was introduced in Dvoirys-Aleksandrov, Podolsk 1991: 13 be! £xc3 14 :b1 b6 (not 14...£xe515 :e1 f616f4) 15 g4~416f4h6(16 ...c517Wf2) 17 ~ £as 18 :b3 h5 19 h3 hg 20 hg ~g6 21 Wg2 :d8 22 :bd3 :xd3 23 cd, and White had a considemble endgame advantage. lOb3

Bird, Schliemann and Berlin Defences 169 Instead, 10 :dl would transpose into the 'primary' game Short-Gelfand, in which White did without the pawn move b2-b3 altogether. Unclear play arises from 10 h3 .i.e6 11 g4liJe7 12 ~gS (12 ~4 .i.d7!) 12....i.c4 13 :el ~ 14 ~1 h6 15 ~ ~b4 16liJe3 .i.d5 17 ~g3 .i.e6 18 a3 ~S 19 ~g2 a6 20 f4 .i.cS+ 21 cRh2 h5!? 22 ~xh5
1990. 10... h5 After W...aS 11 .i.b2 (11 h3.tb4 12 .i.b2 .i.xc3 13 .i.xc3 c5 equalised in I.Polgar-Dautov, Bmo 1991), we reach a position from the first game of the Women's World Championship Match Xie lun-Chiburdanidze, Manila 1991, which continued: 11.. ..i.e6 12 :Cdl .i.e7 13 h3 h5 14 a4 f61SliJe2.i.d5 16liJel rIif7 (it was worth exchanging on e5: 16...fe 17 .i.xeS .i.d6) 17 ~4 :ad8 18 c4 '*.e6 1900 .i.c8 20 :el gS 21 e6+ cRe8 22 ~g6 :g8 (22 ...:h6!) 23 ~xe7 cRxe7 24 g4 hg 25 hg ~g7 26 ~4 cS 27 ~S+ ~fS 28 gf :h8. Black clearly has counterplay, and peace was shortly concluded. 11 :d1.i.e7 The threat is ...~4. An inferior line for Black is 11...'*.e6 12 ~gS .i.b4 13 ~e4 .i.d5 14 c4. 12 .i.g51 (131) 131 B

12... ~613h3 The position would immediately level out after 13 .i.xe7 cRxe7 14 :d2.i.g4!. 13... .i.f514 Le7cRxe7 15 ~4 :ad816fld2 .i.g6 Not 16... .i.xc2? 17 ~xc6+ be 18 :xc2, and Black has a difficult position. But 16....i.c8 was better, preparing to return the knight to fS. 17:ad1 h4

170 Bird, Schliemann and Berlin Defences Alternatives would also have left White with. the better chances: 17 ...~5 1800 lhd2 (or 18...h4 19lbe4, heading for cS) 19 :xd2 :d8 20 lhd8 ~xd8 21ltle2, or 17...cS 18llklbS :xd2 19 :xd2 :d8 20 00+ :xdS 21 :xdS c6 22 :xcS cb 23 :c7+ ~e6 24 :xb7 .i.xc2 25 LbS .i.d3 26 a6 27 f4 h4 28 ~f2 ffi 29 :CS. 18 b4! fQf5 191Oce2 ~4 20 ~4 f6 21 ef+ ~xf6 22 tDb3! The threat of lbas gives White a tangible plus in the ending, which I gradually succeeded in exploiting. I will give the game in full: 22... Ltl2 23lhi2 b6 24 :d7 :c8 25 f1d4 Not as strong as 2S c3. 25... Lc2 26 flxh4 :e8 27 :f4+ ~e5 28 :p Lb3 29 ab ~d4 30 Lg7 ~c3 31 Lc7 ~xb4 32 L6 ~xb3 33 f4 ';,f8? More stubborn resistance was offered by 33 ...bS!. 34g4! But now it is allover. 34...'4xf4 35 g5 b5 36 g6:/8 37 g7:'8 38 :c7 as 39 h4 a4 40 h5 a3 41 h6 a2 42:a7 1-0. We now return to the primary game.

:as

9

:dl+

~e8

10 M In this case White dispenses with b2-b3. 10 .i.e6 (132) In Matulovi~-Karaklaji~, Yugoslav Ch 1991, Black preferred 10...~7. White reacted correctly with 11lbd4! (11 h3.i.f5 would give equality) 1l...~S (11...h6? loses to 12llklbS! cb 13/t)xbS 00 14 :xdS c6 15 lOc7+ ~e7 16/t)xa8 cd 17lOc7 - Matulovi~) 12lbde2 .i.e7 13 b3 .i.d7 14 .i.b2:d8 15 lbe4 hS 16 c4 as 17 :d3 .i.c8 18 :adl Ld3 19 lhd3. White has some advantage in the ending. 000

132 W

Bird. Schliemann and Berlin Defences 171 The diagram position can arise from various move-orders. The position in Dvoirys-Aleksandrov which we have already discussed arose from the diagram via 11 ~g5 .tc4, and we recall that White obtained the advantage.

11

h3

Another interesting line is 11 ~2 .td5 12 ~1 h5 13 ~f4 :d8 14 b3 1.e7 15 .tb2 gS 16 ~2. In Anand-Salov, Reggio Emilia 1991/92, there followed 16...:g8 (Anand recommended 16 ...cS. with equality) 17 c4 .te6 18 ~2 as 19 ~3 :d7 20 ~ cS 21 ~fl ~d8 22 ~e2 ~c8 23 lbd7 ~xd7 24:dl + ~c6 2S a4 b6 26 ~1 ~b7 27 g31Oh6 28 ~3 1.g4+ 29 f3 1.e6 30 00 .td8 31 ~3 .te7 32 .tc3 c6? 33 ~lc2 .td8 34:hl! g4 35 f4 ~fS 36 :dl .tc7 37 ~xf5 .txf5 38 ~3 .tc8? 39 ~d3. After several errors by both sides, White has obtained a won position. 11 ••. 1.b4

12.tdl

as

A new move. After 12...:d8 13~, White has a definite initiative.

13 14 15

lOe2 .tc5 lbf4 1.d5 IDel (133)

133 B

Black has obtained a roughly equal position. A long war of manoeuvre now commences. 15 .tb6 IS ...h5 was also playable. 00'

16

g4

1De7

16...~g3?! is dangerous on account of 17 1.c3 and 18 1.d4. 17 .tel Setting up the threat of eS-e6.


17 18 19

.te6 ~S

:g8

lOcI3 19 00 also looks quite good. 19 ... lOds 2O.td2 cSt With this move, the possibility of c2-c4-c5, with pressure for White, is nipped in the bud. Black is now ready to go into action on the queenside himself. 21 lOdf4 If21 c4, then 21...1Ob4!? 22 .txb4 (22IOxb4 cb 23 b3 .td4 favours Black) 22 ...cb with unclear play. But 21 a3!? was a move worth considering.

21 22

23

10M c3 :e1

lDc6

23 IOxe6 fe 24 J.f4 looks attractive for White.

23

•••

a4?!

In Short's view, 23 ...:d8! 24lOxe6 fe 25 .tgS :d3 would have led to a double-edged struggle. Now, however, White acquires the better chances.

24 2S

:e4! a3

c4

Preventing 25 ...a3, but 25lOxe6 was even stronger.

2S 26

..• lOxe6

:d8 fe?!

It was worth considering 26...:xd2!. Then 27 lOexg7+ ~f8?! 28 IOrs J:xf2 29 ~hllOe7! 30 lObg3! and 31 :dl would be good for White, but Gelfand gives 27 ...~d8 instead. 27 J.gS :d7

28 29

l:ael :lle2

30

~g2(l34)

lOsS

:m

White has a solid plus; his kingside pawns are ready to advance. Black now makes things worse with a time-trouble error. 30 ••• h6?! 30....tc5 was essential.

31 32 33

lOxg7+ J.xh6 J.xf8

lbg7 :lgf7 ~


Generally speaking, two pieces are enough for a rook and two pawns, but in this case Short's pawn phalanx is too dangerous. 34 b4 -*.cS 35 ~g3 ltd7 36 b5? 36 ltf4+ would have decided the game quickly. 36 ..• ltd3+ 37 f3 37 ~h4!? is also interesting. ltd1 37 38 ltf4+ ~g8 39:t'6 ~b3

40

lhe6

With the win of a third pawn it looks as if the struggle is almost over, but it continues nonetheless. 40 l2X1 41 lte4 ~3 42 ~b4 ibxb2 43 ~g5 :d8! Not 43 ......xa3? in view of 44 lteS+ ~f7 45 ltcS. 44 ltg6+ ~h8 45:t'6! Or 45 ......e7 46 ~f5.

"'xa3

46:n

~1

47 ltd4 ltxd4 48 cd lDe3 49 ltd7 ~g2! But not 49 ...c3? at once, because of 50 ~g6. 50 e6 c3


51

lbe7

In time-trouble Short misses a quick win with 51 e7! l.xe7 52 :'xe7, for example: 52...c6 53 :'e2 b5 54 :'xg2 b4 55 ~g6 ~g8 56 :'e2 ~f8 57 h6, or 52 ...a3 53 ~f6lbf4 (53 ... a2 54 g5 a1" 55 :'e8+ ~h7 56 g6+ ~h6 57 ':h8 mate) 54 :'e8+ ~h7 55 g5lilxh5+ 56 ~f7 a2 57 :'e1!.

51 52 53

lbc3

b5 l.e7+

~g6

lilf4+

54

rkI

lilds ~h7

55:'clI+ 56:as 57

a3 lilb6 (135)

f4

135 W

58 :'e8?? Throwing away the win. After 58 g5!, the g-pawn goes straight through to queen. 58 ••• lOOi?? In the time-scramble Black returns the compliment. After 58 ...a2! 59 :'g8 a1" 60 :'g7+, the game would have ended in perpetual check. Incidentally, on the straightforward 59 g5, Black has the clever retort 59 ...1.xg5! 60 fg a1" 61 g6+ ~h6 62 ':h8+ ~g5 63 g7 "a7+, when Gelfand would even have won. b4 59 ~e7 60 gS ~g7

61

fS

1·0

Index of Variations

Closed Spanish: Zaitsev Variation I e4 e5 2lbf3lbc6 3 i.b5 a64 i.a4lbf6 5 0-0 i.e7 6 l:el b5 7 i.b3 d6

8 c3 0-0 9 h3 i.b7 10 d4 l:e8 IIlbbd2 i.f8 U~

n

12 a4 12 .....d7 13 ab ab 14 l:xa8 i.xa8 15 d5 lbas 15 d5 lbe7 12 ... h6 13 i.c2 13 ...lbb8 13 ...ed 14 cd lbb4 15 i.bl 15 ...c5 16 d5 lbd7 17 l:~ 17...c4 18 ab 18lbd4 17 .. .f5 18 ef 18 ...lbf6 18 ... i.xdS 18 l:ae3 18 ...lbf6 18...f4 18lbh2 15 ... ba 16 l:xa4 as 17 l:~ l:a6 18lbh2 18 l:ae3

11

14 18

25 30

37 42 45 52 57 63 68

Closed Spanish: 9•••lbd7 9 ...lbd7 10 d4 i.f6 II a4 i.b7 12lb~

82

12 ab

90

Open Spanish 1 e4 e5 2 M t2k6 3 .tbS a64 .ta4lill6 S 0-0 lLlxe4 6 d4 bS 7 .tb3 dS 8 de.te6 9 lLlbd2 t2ks 10 c3 10...d4 97 10....te7 106 9 c3 112 9.te3 121 Marshall Attack 1 e4 e5 2lLlf3 t2k6 3 .tbS a6 4 .ta4lLlf6 S 0-0 .te7 6 :el bS 7 i.b3 0-0 8 c3 dS 9 ed lLlxdS 10 lLlxeS lLlxeS 11 :xe5 c6 12d4.td6 13 :el 128 13 :e2 137 12 d3 144 Bird, Schliemann and Berlin Defences 1 e4 e5 2 lLlf3 t2k6 3 .tbS 3...lLld4

3...fS 3...lill64 0-0 lLlxe4

ISO IS9 166

WINNING WITH THE SPANISH )

Throughout his long career, Anatoly Karpov has tortured countless opponents with the Spanish, one of the oldest and most famous of all chess openings. White develops quickly in true Classical style, applying immediate pressure to Black's position. This opening leads to the most rich and varied positions, appealing to players of every style of play: tactical players revel in the sharp gambit variations, while those with more positional ambitions are able to select from a number of more sedate, but equally testing, alternatives. Karpov's own personal account shows you all the tricks of the trade. He draws extensively on his own experience on both sides of the Spanish to illustrate the main strategic themes in this most vibrant of openings. • Up-to-date coverage of the critical variations • Features Karpov's analysis of his World Championship games against Korchnoi and Kasparov • The complete guide to learning, playing and winning with the Spanish - as White or Black!

Anatoly Karpov was the World Chess Champion between 1975 and 1985, during which time he achieved a tournament record unsurpassed by any previous title holder. After losing his title his play has, if anything, improved and he has challenged Garry Kasparov three times, narrowly missing out on each occasion. In 1993 he re-established himself as FIDE World Champion by defeating Jan Timman.

Other titles in the 'Winning With ... ' series include: Winning With the Dragon

Winning With the Naidorf

Chris Ward

Daniel King

Winning With the

Winning With the Modern

La _

_ _ _ _ ...1__

David Norwood

9 With the Torre Jufeld and tsko

ISBN 0-7 tl. I. tlatstord LTd

4 Fitzhardinge Street London Wl H OAH

I II

9 780713

Winning With the Spanish

Recommend Documents