Game No. 8 Shirov-Akopian
Tbilisi 1989
1 d4 lDf6
l c4 g6
3 lDc3 dS
4 lDf3 �g7
5 cd lDxdS
6 e4 lDxc3
7 be c5
s : bt o-o
9 �el
It is interesting that the 1 976 edition of ECO gave this set-up no more than a cursory mention. Yet in the last few years, the system with 8 : b 1 has virtually become the most popular choice against the Griinfeld Defence, perhaps even surpassing the old variation in which the bishop is brought out to c4.
Another method of developing the white pieces, behind that strong pawn centre which characterises the Exchange Variation, involves an early �c1-e3. In this case, the light-squared bishop is brought to e2 (later than usual), and the queen's rook may occupy either c1 or d1 (after 1Fd1-d2). This variation was an object of dispute in my last match with Kasparov - see Games No. 14 and 1 5.
I would add that the popularity of the variation with : b 1 and .te2 is above all associated with the names of the young grandmasters "Boris Gelfand and Alexander Halifman.
9 1FaS
The main line for Black is 9 ... cd 10 cd, and only then 10 . . . 1Fa5 +. We shall deal with it in Games 1 1 - 1 3. The fairly popular moves 9 . .. b6 and 9 ... l0c6 are examined in Games 9 and 10.
Modern Exchange Variation 53 As for 9 . . . .i.g4, it has gone out of use. White gains the advantage, either with the immediate 10 J:l. xb7, or after sacrificing the d
pawn : 10 0-0 cd 1 1 cd .i.xf3 1 2 ..txf3 .i.xd4 ( 1 2 . . . 11fxd4 1 3 11fxd4 ..txd4 1 4 J:l. xb7) 1 3 J:l. xb7.
10 0-0
Of course, sacrificing the pawn on a2 is part of White's plan.
Whether he will be able to work up a dangerous initiative in return is another question. Incidentally, at the beginning of the 1 980s White used to prefer 10 J:l. b5 11fxc3 + ( 1 0 . . . 11fxa2 1 1 J:l. xc5 is worse for Black) 1 1 .i.d2 11fa3 1 2 J:l. a5 11fb2 13 J:l. xc5, and at last Black would take the pawn with 13 . . . 11fxa2. White would then continue 14 J:l. a5, 14 11fc1, or 14 0-0, leading in all cases to a complex game with chances for both sides.
10 11fxa2 (26)
At this point, capturing the c-pawn is extremely risky. White's best reply to 10 . . . 11V xc3 is 1 1 d5. Let us see how play might develop: 1 1 . . . 11fa5 (if 1 1 . . . l'hd7, then 1 2 .i.g5 J:l. e8 1 3 11fa4 is unpleasant, while 12 . . . .i.f6 loses virtually by force : 1 3 .i.d2 11fa3 14 11fc2 l'he5 1 5 J:l. b3 l'hxf3 + 16 gf 11fa4 1 7 .i.b5 11fd4 1 8 ..tc3) 1 2 .i.g5 11fc7 1 3 Wf c l (another good line is 1 3 11fd2 .i.g4 14 J:l. fc1 b6 1 5 h3 ..txf3 1 6 ..txf3 e5 1 7 d6 11fc6 1 8 -*.e7 J:l. e8 1 9 .i.e2 l'hd7 2 0 .i. b 5 11fb7 2 1 11fc2; Legky-Veingold, Lvov 1 984) 1 3 . . . ..tg4 14 .i.f4 11fc8 1 5 e 5 t2J d 7 1 6 J:l. e 1 (or 16 11fe3, connecting the rooks and increasing the pressure in the centre) 1 6 . . . l'hb6 1 7 d6 ed 1 8 ed, and in return for the pawn White has a highly promising position; Chiburdanidze-Malaniuk, Odessa 1 982.
26 w
11 .i.gS 11fe6
An interesting alternative is 1 1 • • • t2Jd7 12 .i.xe7 J:l. e8 13 .i.d6 cd 14 l'hxd4 l'hf6 15 f3 l'hd5!? 1 6 J:l. a 1 11f b2 1 7 11fa4 .i.f5 1 8 ed
54 Modern Exchange Variation
.i.xd4+ 19 lrxd4 .D. xe2 20 g4 .D. d2 2 1 lrb4 .i.d3 22 lr xb2 .D. xb2 23 .D. fd 1 , with equality; Vaiser-Dvoiris, Bamaul 1984. Vaiser suggests 15 .J:l a1 lrb2 16 .lObS!? .ll:lxe4 17 .i.a3 ird2 1 8 .ll:lc7.
12 ird3!?
A valuable idea, first employed in Epishin-Henkin, Bamaul 1 988. Formerly, White used to advance one of his central pawns here. For example : 12 eS .D. d8 1 3 lra4 lrc6 ( 1 3 . . . .i.d7 14 dS .i.xa4 1 5 de f6 1 6 ef ef 1 7 .i.c4! .D. e8 1 8 .i.f4 b6 19 .i.dS .i.c6 20 c4! gave White a clear advantage in Yusupov-Tukmakov, Moscow 1 983) 14 irb3 .i.e6 ( 1 4 . . . lrc7 1 5 .i.c4 .D. f8 1 6 e6 f6 1 7 .i.h4 .ll:lc6 1 8 .i.g3 ird8 1 9 lra2 cd 20 cd b 6 2 1 d S .ll:leS 2 2 .D. fd 1
�h8 2 3 .ll:lxeS fe 2 4 d6! gave White a large plus in another game Yusupov-Tukmakov, Erevan 1 982) 1 5 c4 cd 1 6 .i.xe7 .D. c8 1 7 .ll:lxd4 .i.xc4! 1 8 .ll:lxc6 .i.xb3 1 9 .ll:lxb8 .i.e6 20 .i.d6 .i.f8 2 1 .i.f3 .i.xd6 2 2 .i.xb7 .i.xb8 2 3 .i.xa8 .i.xeS 24 .i.b7 .J:l c2 2 5 .D. fc 1 .J:l a2 2 6 .l:t d 1 .i.c7 2 7 .J:l a 1 .D. b2 2 8 .l:t db 1 .l:t d2 2 9 .l:t d 1 .l:t b2 1-1; Damjanovic-Schmidt, Athens 1984. This one game serves to illustrate the harmlessness of 1 2 eS.
12 b6
After 12 • . . lrd6 1 3 lre3, or 12 • • • .ll:ld7 1 3 lre3 .ll:lf6 14 .ll:leS!, the initiative is with White. On the other hand, 12 . . • .D. d8!?
deserves to be tried.
13 d5
In this case, 1 3 lre3 is no good : 1 3 . . . .i.a6! 14 dS 1rd6 1 5 c4 e5!, and White has nothing for the pawn.
27 w
13 ird6 (27)
14 e5!
The point of White's play. He now obtains the two bishops and
Modern Exchange Variation 55 a big advantage in space and development. Still, Black does have two extra pawns. 1Wc3, but Henkin has indicated a line in which Black again has noth
ing to fear: 25 . . . .te2 26 f3 bS! 27 cb .txbS 2S : xeS + : xeS, 29 1Wf6 : e 1 + 30 �f2 : e2 + 3 1 �g3 c4 32 1Wxd6 a6. But then, the text move should not have brought White any concrete gains either.
56 Modern Exchange Variation
19 Wf6
20 d6 : b8
21 : bd1 (28)
2 1 de is bad : 2 1 . . . :Z. xe7 22 Wd2 .!LieS 23 ..tgS .!Llxf3 + 24 gf :Z. d7! 25 Wxd7 WxgS +, and Black wins.
28 B
21 eS?
A serious mistake. He could not have solved his problems with
21 . • . ..tb7 22 de :Z. xe7 (22 . . . ..txf3 23 :Z. xd7 ..tg4 24 : xa7
:Z. bc8 25 :Z. e l ) 23 :Z. xd7! : xe3 24 fe ..txf3 25 ..txb8 W xc3 26 :Z. xf3 Wel + 27 : fl Wxe3 + 28 ..thl We8 29 :Z. fd l ! and wins. But the essential move was 21 . . • e6!, and already after 22 ..tc6 Wd8 23 :Z. fe l bS it is White who has to think about equalising.
22 ..tg5 Wg7
If 22 . . . Wf5, then 23 ..te4 We6 25 ..tdS WfS 25 f4!, while 22 . . . Wh8 is well answered by 23 ..tc6.
23 ..th6 Wf6
24 ..tc6! :Z. e6?
24 . . . Wd8 25 f4 ..tb7 26 ..txd7 Wxd7 27 fe f5! was more tenacious, although after 28 ..tgS ..te4 29 ..tf6, White has an advantage due to the subsequent h4-h5.
25 ..tg5 Wg7
26 f4 h6
If 26 . . . f6, then 27 fS! gf 28 ..th6 Wg6 29 :Z. f3 f4 30 :Z. xf4!
27 ..te7 ef
28 Wxf4 g5
29 Wa4 .!Oe5
30 ..td5 ..tb7 31 Wxa7 .!Lld7 32 ..txb7
1-0
Modern Exchange Variation 5 7 Game No. 9
Halifman-Pribyl Leningrad 1989
1 d4 ltlf6 2 c4 g6 3 ltlc3 d5 4 ltlf3 J.g7 5 cd llJxd5 6 e4 ltlxc3 7 be c5 8 l:l. b1 0-0
9 J.e2 b6
10 0-0
Another familiar plan is 10 J.e3 J.b7 1 1 e5 llJc6 1 2 h4, with a kingside attack. However, in a game Nemet-Korchnoi, Switzerland 1985, Black seized the initiative with ll . . . cd! 12 cd ltla6 1 3 1rd2 ltlc7 14 h4 ltld5 1 5 h5 llJxe3 1 6 fe e6 1 7 .i.d3 l:l. c8.
There is also another possibility (after 1 1 . . . cd 12 cd): 1 2 • • • J.e4!
13 l:l. c l llJc6 14 h4 l:l. c8 1 5 llJg5 J.xg2 1 6 l:l. h2 J.d5 1 7 h5 llJb4 1 8 hg hg 1 9 a3 llJa2, and Black is clearly better; Vaiser-1. Sokolov, San Bernadino 1 989.
In Cebalo-1. Sokolov from the same tournament, White launched his attack at once with 10 h4!?. His opponent's ineffective response unexpectedly led to a quick and crushing defeat : 1 0 . . . J.g4 ( 1 0 . . . J.b7 is more logical) 1 1 '/o>fl !? cd 1 2 cd llJd7 (and here, the right continuation is 1 2 . . . llJc6 1 3 d5 J.xf3 14 gf ltld4) 1 3 e5! l:l. c8 14 J.e3 llJb8 1 5 h5! 1rd5 1 6 hg hg 1 7 1rd2 J:Hd8 1 8 l:l. h4! .i.xf3 19 gf J.xe5? 20 l:l. b5! 1re6 2 1 l:l. xe5, and i t i s all over.
29 B
10 J.b7
1 1 1rd3 (29)
1 1 J.a6
In earlier games, Black played 1 1 . . . cd 12 cd e6 13 .i.g5 1rd6, but as a rule failed to equalise. Here is one example : 14 'iWe3 l:l. c8 1 5 l:l. fd 1 llJd7 1 6 h4 l:l. c2 1 7 e5 1rc7 1 8 J.d3 l:l. c3 ( 1 8 . . . l:.Xa2 19
58 Modern Exchange Variation
l:t bc 1 ) 19 'lre2 .*.d5 20 : be l h6 2 1 .*.e3 l:t dS 22 h5! with a big advantage; Lputian-Lalic, Sarajevo 1 9S5.
1 2 'lre3
1 2 'lrc2 promises White less. Chernin-Dorfman, Moscow 19S4, continued 1 2 . . . cd 1 3 cd 'IreS 14 'lrdl .*.xe2 1 5 'lrxe2 'lra6 1 6 21 'lre4, Black has no prospects whatsoever; Lerner-Gorelov, Ivano-Frankovsk 1 9S2.
In the last few years, Black has tried moving his queen to cS or d7 without exchanging pawns in the centre. These experiments have been unsuccessful, for example : 12 • • • 'IreS 1 3 d5 .*.xe2 1 4 'lrxe2 .*.xc3 1 5 e5 'lrf5 1 6 : b 3 .*.a5 1 7 lbh4 'lr d 7 1 S .*.h6, with advantage; Halifman-Lau, Rotterdam 1 9SS. Or 12 . . . 'lrd7 1 3 de be ( 1 3 . . . .*.xe2 14 'lrxe2 be is safer) 14 l:t xbS!? l:t axbS 1 5 .*.xa6 'lr a4 16 'lre2, with a won position; Gelfand-I. Sokolov, Yugoslavia 1 9SS.
13 cd 'lrd7
A recommendation of Grandmaster Gavrikov. A much earlier game Browne-Martz, USA 19S2, went 1 3 . . . .*.xe2 14 'lrxe2 llJc6 'lrxc4 : xc4; Lputian-Zilberstein, Blagoveshchensk 1 988. How
ever, Salov's innovation 14 d5!? deserves attention. After 14 . . . .*.h6!, as in Gelfand-Malishauskas, Vilnius 1 988. However, Pribyl gives IS : fd! as more accurate.
Modern Exchange Variation 59 15 'irxe2 .l:l. e8
Black is not in a hurry to bring his knight out. In Gelfand
Groszpeter, Palma de Mallorca 1989, Black played 1 S . . . c!Llc6, and after 16 dS c!LleS 17 c!LlxeS -*.xeS 1 8 f4 �d6 19 .i. b2 e6 20 .1:1. bd 1 White gained the initiative and soon won.
16 .l:l. fcl e6 17 .l:l. c3 .l:l. c8
1 7 . . . c!Llc6? loses to 1 8 'irbS .l:l. ac8 1 9 .I:!. bel a6 20 'irxb6.
18 .l:l. xc8+
1 8 .l:l. bc1 .l:l. xc3 1 9 .l:l. xc3 c!Llc6 20 "irbS .l:l. c8 2 1 .l:l. c4 'irb7 leads to equality.
18 'irxc8
19 .l:l. c l 'irb7
An alternative is 19 . . . 'ira6 20 "ire3 c!Lld7 2 1 .l:l. c7 c!Llf6 (2 1 . . . jfa4 22 h3) 22 c!LleS .l:l. c8! 23 .l:l. xf7 "irbS 24 g3 jfbl + 2S �g2 jfxe4+ 26 jf xe4 c!Llxe4 27 .l:l. xa7 �xeS 28 de .l:l. c2, with full equality; Dzhandzhgava-Malishauskas, Uzhgorod 1 988. A better line for White, according to Halifman, is 20 jfxa6 c!Llxa6 2 1 �fl.
with some advantage.
30 w
20 'irc4 c!Lla6 (30)
This position had first occurred only a few days earlier, in the game S. Ivanov-Pribyl from the same international tournament (Leningrad 1989). After 2 1 dS ed 22 ed .l:l. d8 23 d6 c!LlcS! 24 �xeS be 2S 'irxcS 'ira6 26 h4 .l:l. xd6 27 c!LlgS h6, the players agreed a draw. The main game we are examining was to end in victory for Black, but that was hardly the rightful outcome. For the moment Halifman doesn't want to settle for a draw, and adds fuel to the flames.
60 Modern Exchange Variation
21 h4 .l:l. d8
22 e5!? h6
23 h5 gh
24 .l:l. c3 b5!
25 'irc6 'irxc6 26 .l:l.xc6 lt!b8 27 .l:l.c7 .U8 28 ..txf8 �xf8 (31)
3 1 w
29 d5!
White's last chance - he hopes for success by sacrificing another pawn. As Pribyl has pointed out, a forced draw results from 29 .l:l. xa7 lt!c6 30 .l:l. b7 lt!xd4 3 1 lt!xd4 .l:l. xd4 32 .l:l. xb5 .l:l. d l + 33 �h2 .l:l. d2.
29 ed
30 lt!d4 .l:l. e8!
31 f4 .l:l. e7 32 .l:l. c5 lt!a6 33 .l:l. xd5
In the event of 33 .l:l. c6 lt!c7 34 .l:l. xh6, Black can save himself with 34 . . . �g7! (but not 34 . . . lt!e6? 35 lt!f5 .l:l. c7 36 .l:l. h8 mate) 35 .l:l. xh5 lt!e6 36 lt!f5 + �g6 37 lt!xe7 + �xh5 38 lt!xd5 �g4.
33 lt!c7
34 .l:l. d6 lt!e6 35 lt!c6 .l:l.c7
36 f5 lt!g5
37 f6?
Now all White's previous work comes to nothing. After 37 e6!
Black would still have to struggle for the draw.
Modern Exchange Variation 61 'ifxa2; Paunovic-Ristic, Vrnjacka Banja 1 983. In comparison with the variation 9 . . . cd 10 cd 'ifa5 + , which we examine in Games Black has numerous problems; Kasparov-Natsis, Malta 01 1 980) 1 3 h4 (White may also castle, with e4-e5 to follow) 1 3 . . . J.g4 14 h5 J.xf3 1 5 gf! (not allowing the knight onto c4). Despite the simplification, White's attacking potential is very noticeable, for example : 1 5 . . . e6 1 6 'if h4 'iff6 1 7 hg 'if xg6 1 8 'ifh2 'it>h8 1 9 'it>d2!
Ghinda, Halle 1 987. Alternatively Black has an interesting counter
sacrifice of a pawn : 12 • • • f5 1 3 ef J.xf5 14 .:t xb7 'if d6 1 5 .:t b3.
62 Modern Exchange Variation
.i.xh 1 2 1 11fe2 l:tf5 22 f3 .J:I. af8 23 .i.h6 .J:1. 8f6 24 11fe4, and Black has worries. A move that looks fairly logical is 12 • • • 11fc7, covering the pawn on b7 and preventing White from castling. A correspondence game Danner-Nesis ( 1 983-5) continued 1 3 h3 f5 1 4 ef .i.xf5 1 5 .J:I. b3 ..td4 1 6 0-0 e5, and Black achieved a sound position.
32 B
In the first half of the 1980s, Black almost invariably played 1 2 . . . e 6 here, and after 1 3 f4 h e retreated his bishop t o g 7 o r h8. I do not plan to give a thorough study of the positions which then arise, but it is worth reminding the reader of the difference that the choice of retreat square makes - or more exactly, of why it is essential to withdraw the bishop into the corner.
Consider the variation 12 . . . e6 13 f4 .i.g7 14 c4 .J:I. e8 1 5 e5 f6 1 4 d6 fe 1 7 .i.b2 ef 1 8 .i.xg7 <t;xg7 1 9 0-0 (33) .
33 B
Practice has abundantly shown that 19 ... .J:I. f8 20 .J:I. xf4 .J:I. xf4 2 1 11fxf4 secures White a distinct plus. But at move 19 Black
Modern Exchange Variation 63 cannot play 19 . . . eS, on account of 20 d7!, and the pin on the 7th rank will be lethal.
However, if earlier Black plays 13 . . . .th8, and accordingly 1 8 .txh8 �xh8 occurs, then 1 9 0-0 can be answered effectively by 19 . . . eS!, giving Black a solid position after 20 :t fd l 'it'd7 2 1 'it'dS :t b8 22 'it' xcS b6; Bikhovsky-Polovodin, Irkutsk 1983.
All the same, I should point out that after 13 . . . .th8 14 c4 :t e8 15 e5 j6, White doesn't have to play 1 6 d6; he has the stronger 16 f5! (34) .
16 . . . ed
34 B
Other pawn captures are no better: 1 6 . . . ef 1 7 e6! followed by 1 8 0-0 and 19 g4; 1 6 . . . fe 1 7 fg ( 1 7 fe 'it' h4+ 1 8 �d 1 doesn't look bad either) 17 . . . hg 18 0-0 ed 19 cd 'it'd6 20 :t b3 .tg7 2 1 :t g3 etc.; 1 6 . . . gf 1 7 :t b3 :t e7, and now 1 8 .tb2, 1 8 d 6 and 1 8 'it'h6 are all highly unpleasant for Black.
1 7 fg
Practice has also seen 1 7 e6 d4 1 8 g4 b6 19 .tf3 : b8 ( 1 9 . . . .ta6!?) 20 'it'g2 'flc7 2 1 0-0 gS 22 .tdS 'it'e7 23 h4, with the initiative; Vaiser-Pribyl, Sochi 1984.
17 . . . hg, and now White has two lines :
(a) 18 0--0 : xeS 1 9 .tf3 d4 20 :t e l 'it'e8 2 1 .ta3 .tfS 22 : xeS 'it' xeS 23 .tdS + 'it' xdS 24 cd J..xb l 2S J..xcS, and Black is in a bad way; Cebalo-Raicevic, Citorle 1984.
(b) 18 cd : xeS 19 0-0 b6 20 d6 J..e6 21 .tf3 :t c8 22 J..b2 fS 23 'it'h6! and White has a won position; Agzamov-Pribyl, Sochi 1984.
ll b6
A comparatively new idea. Black limits the effect of his
64 Modern Exchange Variation
opponent's rook on the b-file, and makes b7 available to his own bishop.
Considerably stronger than 15 o-o, which Black used to meet with 15 . . . l% e8. But 15 . .. i.d4+! is better; after 16 �h1 l% e8 1 7 i.d3 ed 1 8 ed i.f5! 1 9 i.xf5 gf, Black obtained good play in Novikov-Krasenkov, Odessa 1989. The game continued very sharply and ended in a draw; here is how : 20 i.b2 1rf6 2 1 i.xd4
Modern Exchange Variation 65 after 2 1 .I:!. xeS 11Ff6!. There followed 22 11Fel (22 .*.xd5 + was more to the point, giving equality after 22 . . . .*.xd5 23 11Fxd5 + �h8) 22 . . . .l:l. xe5! 23 11Fxe5 (23 fe 11Ff4) 23 . . . 11rf7 24 .l:l. e2 (better is 24 cd .l:l. e8, when Black has a minimal edge) 24 . . . de 25 .*.g4 11Fg7!
and White's position is not easy.
35 w
21 fe!
A significant refinement.
21 11Fc7
22 .*.xd5+ .*.xd5 23 'it'xd5+ �g7 24 e6 .l:l. ad8 25 11Ff3 .l:l. e7
He should have blockaded the pawn with his queen, although after 25 . . . 11Fe7 26 .1:1. be2 White still has the more pleasant position.
26 .1:1. £2 11Fd6
27 11Ff6+ �g8
28 11Fg5!
This accurate move casts doubt on Black's set-up. After the immediate 28 h4 .l:l. de8, White would have to acquiesce in a draw with 29 11Ff7+ �h8 30 11Ff6+. But now 28 . . . .l:l. de8 is answered by 29 11Fd5! .l:l. d8 30 .l:l. d2! with a won position.
18 �g7
29 'it'f6+ �g8
30 11Fg5 �g7
31 h4! .l:l. de8 32 11Fd5 .l:l. d8 33 11Fg5
Of course not now 33 .l:l. d2? 11Fg3!, and it is Black who wins.
66 Modern Exchange Variation
33 :tde8
34 :t dl 11fc6?
As Henkin has shown, the only defence was 34 0 0 0 11fc7. Then 35 h5?! :t xe6 36 h6+ �f8 37 :t xe6 :t xe6 38 :t d8 + �f7 39 :t h8 11fd6 40 :t xh7 + �g8 4 1 :t g7+ �h8 42 :t xa7 :t e l + ! 43 �f2 :t e5 44 :t a8 + would lead to a draw. Still, after the quiet 35 :t del White would retain some positional advantage.
36 w
35 11fe5+ �g8 (36)
36 :t d7!
Black's pieces are stalemated.
36 b5
36 0 0 0 h5 doesn't help : 37 :t e3, followed by 38 :t g3. If 36 0 0 0
11fa4, White's simplest course would be 3 7 h5 11fxc4 38 :t xe7 :t xe7 39 11f b8 + �g7 40 h6+ ! �xh6 42 11ff8 + :t g7 43 e7.
37 cb 11fxb5
38 hS 11fb6
39 :t xe7 :txe7
40 h6 c4+
41 �h1 11fd8
42 :t b1 ! :tc7 43 11ff4!
After 43 11fd5 :t c8!, or 43 11fd4 'tiffS!, Black could still hold on.
43 g5
The threat was 44 e7!
44 11fxg5+!
1-0
Modern Exchange Variation 67 Game No. 1 1
Vaiser-Pein Budapest 1989
l d4 lt!f6 2 c4 g6 3 lt!c3 dS 4 lt!f3 ,j_g7 S cd lt!xdS 6 e4 lt!xc3 7 be cS 8 l:t bl 0-0
9 ,j_el cd
10 cd 'lfaS+
At the present time, the exchange on d4 followed by the check with the queen is Black's most popular choice in this variation.
l l ,j_dl
Formerly, 1 1 'lfd2 was seen more often, but in the resulting endgame White can hardly count on a plus. For that reason, the search for the initiative now proceeds on different lines, and involves a pawn sacrifice. This variation of the Griinfeld occurred only once in my encounters with Kasparov - in the 1 3th game in Seville, where 1 1 'lt'd2 'lfxd2 + 12 ,j_xd2 was played (37) . Let us briefly examine that game.
3 7 B
1 2 . . . e6 (in answer to 1 2 . . . b6 at once, White has various choices : 13 ,j_d3, 1 3 d5 or 1 3 l:t c 1 , but best of all is 13 0-0. Then after 1 3 . . . ib7 14 d5 ,j_a6 1 5 l:t fe l ,j_xe2 16 l:t xe2 lt!a6 1 7 .i.e3, White has the better chances. If 17 . . . f5, Gutman recommends 1 8
�d4! fe 19 l:t xe4 if6 20 lt!e5!. I f 1 7 ... l:t fe8, he recommends 18 l:t c2!, instead of 1 8 lt!d4 �xd4 19 .i.xd4 e6 20 de, which led to a draw in Halifman-Tseshkovsky, Minsk 1 985) 1 3 0-0 ( 1 3 .i.c4, 1 3 l:t c l and even 1 3 h4 have been seen, but there i s n o point in starting active operations with incomplete development) 1 3 . . . b6 1 4 l:t fd l (it is hard to decide how to arrange the rooks on the first rank; 14 l:t fe l , 14 l:l. bc l and 14 l:l. fc l have also been played.
68 Modern Exchange Variation
As the d-pawn is due to advance, I place a rook behind it) 14 . . . ..tb7 1 5 d5 ed 1 6 ed lDd7 ( 1 6 . . . ..txd5 loses to 1 7 ..tb4 .l:l. d8 1 8 ..tc4) 1 7 ..tb4 .l:l. fc8 1 8 ..te7 (neither 1 8 tDd4 nor 1 8 ..tb5 gives White anything. But now his initiative looks fairly substantial -for example 1 8 . . . ..tf8 1 9 d6 ..txe7 20 de tDf6 2 1 .l:l. e 1 !. However, Black manages to find a powerful retort which solves all his opening problems) 1 8 . . . ..tf6! 1 9 d6 �g7! 20 .l:l. e 1 (20 ..txf6+
is answered by 20 . . . �xf6, but not 20 . . . lDxf6 2 1 1 d7 .l:l. d8 21 lDe5!) 20 . . . .l:l. c5 2 1 ..tb5 ..tc6 22 ..txc6 .l:l. xc6 23 .l:l. bd 1 ..tc3 24 .l:l. e3 f6 25 g4 g5 26 h4 h6 27 hg hg 28 lDd4 (now Black seizes the initiative; the correct line was 28 .l:l. c 1 .l:l. ac8 29 ..td8! with equality) 28 . . . ..txd4 29 .l:l. xd4 .l:l. h8 30 .l:l. e l .l:l. c2 (Black could have continued quite actively with 30 . . . .l:l. c3 3 1 �g2 �f7; but now peace ensues quickly) 3 1 a4 a5 32 f4! �g6 33 fg �xg5 34 .l:l. fl
�g6 35 .l:l. f2 .l:l. hc8 36 .l:l. df4 .l:l. xf2 t-t.
1 1 "Wxal
12 0-0 b6 (38)
The main variations are 1 2 . . . tDd7 and 1 2 . . . "We6; these will feature in Games 12 and 1 3. There is no point in exposing the knight to a pawn attack : after 12 . . . lDc6 1 3 d5 lDe5 14 lDd4, White has clearly the better chances. But then, in the last few years, practice has shown that 1 2 . . . b6 is also dangerous for Black. The present game will convince us of this.
38 w
13 "Wet! "We6
The alternative is 1 3 . . . ..tb7 14 ..tc4 "W a4 1 5 ..tb5 "Wa2 (now of course White can force a draw by repetition, as occurred for instance in Chemin-Dvoiris, Lvov 1990; but that is not much of an achievement) 1 6 .l:l. e l ! (but not 1 6 "Wet? "Wc2! 1 7 d5 ..ta6! 1 8
Modern Exchange Variation 69 We3 .txb5 1 9 J: xb5 �a6 20 l:l. bb1 �c5 2 1 .tb4 aS!, and Black seizes the initiative; Yusupov-Korchnoi, Reykjavik 1 988) 1 6 . . . : c8 1 7 Wd1 e6 1 8 We2 �c6 1 9 We3 l:l. d8 ( 1 9 . . . e5 is no better : 20 d5 �d4 2 1 �xd4 ed 22 Wf4 .tf8 23 l:l. a l Wc2 26 .i.d7! with more than enough for the pawn; Halifman-Epishin, Vilnius 1 988) 20 : a 1 Wb2 21 .txc6 .txc6 22 .tc3 W b3 23 J:l. eb1 Wc2 24 �d2 .th6 25 W xh6 W xc3 26 �f3 f6 27 l:l. c 1 'lfb2 28 'Wf4! 'it;>g7 29 'Wc7 + .td7, and now instead of 30 'Wd6, as played in Epishin -Mark Tseitlin, Leningrad 1 988, Epishin gives 30 : cb 1 ! 'We2 3 1 J:l. e 1 'Wb5 3 2 : xa7 : xa7 3 3 'Wxd8 .te8 3 4 e 5 J: f7 35 h3, with the advantage.
14 .tc4!
15 :e1 16 .tb4 17 J: xe6!
'Wxe4 'Wb7 .te6
A typical exchange sacrifice giving White a dangerous initiative.
39 B
17 fe
18 �g5 (39)
18 �c6
A novelty which, however, does not solve all Black's problems.
In Gelfand-Dorfman, Minsk 1 986, Black played 18 • . • �h8, which gave him a won position after 19 �xe6 �d7 20 .txe7?
: fc8. In Vaiser-Andrianov, Naberezhnie Chelny 1 988, White's play was strengthened : 19 J: b3! �d7 20 l:l. h3 �f6? (loses at once.
A few days later, in a game Konyashkin-Titlyanov, Black replied 20 . . . h5!, and after 2 1 'W b 1 :rs 22 �f7 + �h7 23 .txe6 �f6 24 .txf5 gf 25 'Wxf5 + 'itfg8 26 'Wg6 'We4 27 �h6 + �h8 28 �f7 +, the game had to end in perpetual check. However, the right line was 2 1 'Wd 1 ! �f6 22 'Wb1 'Wc6 23 .txe6, with numerous threats)
70 Modern Exchange Variation
Things also go badly for Black in the endgame that results from 25 . . . 'irf3 26 g5 'irh5 27 'irxh5 gh 28 f4 .l:l. gg8 29 wf2 .tg7 30
Modern Exchange Variation 71 13 .tb4
The natural move, which is usually played almost automatically.
However, in Nemet-Kozul, Liechtenstein 1 989, White decided to postpone this active bishop sortie and drive the queen away : 1 3 J:l. a 1 'ire6 ( 1 3 . . . 'irb2 i s also playable) 14 'ir b 1 'irb6 1 5 'ird3 'ird8 16 l:tfcl b6 17 .tb4 ltlf6! (but not 17 . . . .tb7 18 e5!) 18 ltle5 .tb7 19 ltlc6 .txc6 20 .J:I. xc6 aS 21 .td2 (other bishop moves are answered by 21 . . . b5, when White has inadequate compen
sation for the pawn) 21 . . . J:l. c8 22 .J:I. ac 1 l:l. xc6 23 l:l. xc6 'ira8 24 d5 ltld7. Black's chances are better; here is how the game ended : 25 'ifa3 .td4 26 .th6 J:l. e8 27 .tb5? ltlc5! 28 .J:I. c7 'ird8 29 .J:I. a7 'irb8 30 .txe8 'ifxa7 3 1 'irh3 'irb8 32 .tc6 ltlxe4 33 d6 ltlxd6 34 'ird7 ltlf5 35 .td2 'ire5 0-1 .
40 w
13 ltlb6 (40)
Before proceeding to the game continuation (14 ltle5), let us look at some alternatives.
The exchange of the pawns on e4 and e7 took place in Vaiser
Kozul, Ptuj 1989: 14 .txe7 J:l. e8 1 5 J:l. a 1 (or 1 5 .tc5 l:l. xe4 16 .td3 J:l. e8 1 7 J:l. a 1 'ir b2 with a draw, but Black also has the interesting reply 15 . . . ltld7!) 1 5 . . . 'ire6 16 .tc5 'ifxe4 1 7 ltle5 .txe5 18 .tf3 .txh2+ 19 �xh2 'iff4+ 20 �g1 ltld7 21 .txa7 ltlf6 22 'irc1 1Wf5 23 J:l. e 1 J:l. xe 1 + 24 'ifxe 1 'iff4. White has the advantage of the bishop pair, which compensates for the pawn minus.
An equal game results from 14 .tb5 .td7 1 5 .txd7 ltlxd7 1 6 .t.xe7 l:l. fe8; Riemersma - Conquest, Dordrecht 1 988. I t is worth considering 14 d5, with the threat of 1 5 .txe7.
72 Modern Exchange Variation
The game Halifman-Gavrikov, Moscow 1 988, took the follow
ing interesting course : 14 'fldJ!? r1 e8 15 liJg5!
Exploiting the unfavourable placing of the enemy queen, White sets up the threat of 16 f1 a l . The tactical justification, given by winning the pawn back and retaining a positional advantage.
26 . . . liJc4 27 'flg4 �g7 28 .td4+ J-J.
4 1 w
Modern Exchange Variation 73
: a3, as in S. Ivanov-Novik, USSR 1 989, and now 2 1 . . . .tc4!
22 de .txfl 23 �xfl , again with an equal position.
Tukmakov played 15 d5??, not even suspecting the danger awaiting him : 1 5 . . . lllxd5!, an attractive stroke which decides the issue at once. After 16 ed .tf5, not only the rook on b1 but also the bishop on e2 is left under attack, and White could very well cease resistance. This actually happened 12 moves later.
14 llle5 (42)
14 : a 1 "ife6 1 5 "ifc2 is interesting, but has yet to be tried out in practice. It was Neverov who suggested jumping to e5 with the knight, and the move was first played in a game Alterman
Rogozhenko, USSR 1 989.
42 B
14 f6
In the original game with this line, Black decided to eliminate the knight at once, and was quickly crushed : 14 • • • .txe5 1 5 de
"ife6 1 6 "ifd4 .td7 1 7 f4 f6 ( 1 7 . . . .tc6!? was more tenacious;
after 18 f5 "ifd7, the play isn't so clear) 18 ef "ifxf6 19 "ife3! : ac8
74 Modern Exchange Variation
20 .tc5 .tc6 21 f5 'ife5 22 l:l. xb6! ab 23 .tc4+ l:l. f7 24 .txf7 + ! and White won.
The retreat of the queen to e6 in these circumstances led to disaster in Ivanov-Maslov, Leningrad 1990 : 14 . . • 'ife6 16 fe f5 1 6 .tc5 fe 1 7 .txb6 l S de ab 1 9 'ifd4 'ifc6 20 l:l. fc l (or 20 .tc4+
.te6 21 .txe6+ 22 l:l. xb6 'iff5 23 l:l. xb7, with the better chances) 20 . . . l:l. a4 (he should have played 20 . . . 'ifd7 2 1 'ifxb6 'ifdS, with a complex struggle) 2 1 'ifd l ! 'ifd7 22 : xeS! and wins.
14 . . . .te6 is also dangerous for Black: 1 5 .txe7 .: feS 1 6 .th4!
(stronger than 16 .tc5 .txe5 17 de .tc4 lS .txc4 'ifxc4, with equal chances; Halifman-Kindermann, West Berlin l 9S9) 16 . . . g5 ( 1 6 . . . .:t eeS may be better) 1 7 .txg5 (or 1 7 .tg3 : adS l S .tb5 .: rs, as in Sakayev-Novik, USSR 1990, and now 1 9 'ifh5!
f6 20 d5 fe 21 de 'ifxe6 22 'ifxg5 with a clear plus) 17 . . . f6 1S .:t a l 'ifb3 ( 1 S . . . "ifb2 1 9 .!Lid3 'ifxd4 20 .te3 'ifxe4 was better) 19 'ifd2 l:l. acS 20 .td1 'ifc3 2 1 'ifxc3 l:l. xc3 22 .td2, and White wins.
15 .!Lic4!
In the game Shevelev-Lagunov, USSR 19S9, in which 14 . . . f6
In the game Shevelev-Lagunov, USSR 19S9, in which 14 . . . f6