W
e are often faced by such a question . Of course, a ready-made solution does not exist - everything depends on the specific circumstances. We can learn to understand this problem better if we see how it was solved by other players, and each time make a critical assessment of their actions. It is interesting to follow how a player's choice is influenced by his charac ter and style of play, when as a conse quence of individual preferences he is unable to decide on the objectively best course.We will begin with an analysis of two games by Mikhail Tal . He played the first when he was at the height of his powers, and the second many years later (I hope you will sense the d ifference). In the analysis of the games we will do some training by seeking replies to the difficult questions which invariably arise on the way.
Vasyukoy - Tal
29th USSR Championship, Baku 1 96 1 King's Indian Attack
1 e4 e6 2 d3 d5 3 ttJd2 ttJf6 4 ttJgf3 ttJc6 5 g3 dxe4 6 dxe4 .ic5 7 .ig2 e5 8 0-0 0-0 9 'iVe2 .ie6
In this set-up Black usually plays . . . a7-a5, . . . b7-b6 and . . . .ia6. By placing his bishop
on e6, Tal effectively condemned it to being exchanged for a white knight.
10 c3 a5
In such positions one cannot allow b2-b4, which secures White the initiative on the queenside.
11 ttJc4 1 2 ttJg5
ttJd7
Black is forced to part with his bishop. After the exchange he can recapture on e6 with a piece or the f-pawn . A third possibility is to give up the bishop for the knight on c4. Which would you prefer?
Of the two knights, the one on c4 is the more dangerous - it can subsequently be switched via e3 to d5 or f5. True, after the doubling of the pawns on e6 these points will be defended , and the f-file opened . Even so, after 1 2 .. . "e7 1 3 ttJxe6 fxe6 Black's posi tion looks dubious. Apart from playing for simplification with 1 4 .ie3, with the future hope of exploiting the weak pawns, 1 4 'iti>h1
followed by f2-f4 also deserves serious consideration. After the opening of the position the power of the two white bishops will tell.
1 2 . . . 1 3 'ii'xc4
i.xc41 'iVe7
Now Evgeny Vasyukov could have retu rned his knight to f3 and then played it to h4, provoking the reply . . . g7-g6, after which his bishop would have gained an excellent post at h6. However, he begrudged moving his knight back without a special invitation.
1 4 �d2
The plan is understandable: .l:tad 1 and then �c1 . 1 4 . . . 1 5 .l:tad1 1 6 'iVe2 .l:tad8 lLl b6 Of course, not 1 6 'ifb5? .l:txd2! .
1 6 . . . l:td6
Now 1 7 i.c1 ( 1 7 lLlf3!?) 1 7 .. . l1fd8 1 8 ':'xd6 'ifxd6 1 9 .l:te1 followed by 20 i.f1 is not bad , gradually taking control of the squares on the queenside. With this pawn structure the two bishops would have ensured White a small but lasting positional advantage. A good textbook example of how to handle such positions is the game Petrosian-Sax, played in 1 979 at the international tourna ment in Tallinn (cf. the addendum to the lecture).
17 i.h3
Think about what Tal might have played here.
(see diagram)
1 7 . . . l:txd2! ? A n unexpected combination. But not for Tal , who anticipated i t before his previous move. Here is his commentary:
A glance at the position suggests that play will continue on quiet positional lines, and that here there is no place for combinations.
This is not so. When I began considering my 1 6th move, I automatically looked at 16 .. . f1xd2, but it very soon became clear to me that with this move Black does not achieve anything, since after 1 7 'ikxd2 I1d8 1 8 'ifc1 1:!xd1 1 9 f1xd1 he remains the exchange down.
But now let's use our imagination a little. It turns out that, if the white bishop were not at g2, but at h3, a combination would be possible. I thought over my 1 6th move for 40 minutes. Initially I wanted to play 16 .. . 1:1.d7, provoking the reply 1 7 i.h3 (with gain of tempo!). But I decided that this would be too obvious. White can simply continue 1 7 �c1, retaining a minimal positional advantage. In the game there followed 1 6 .. . l1d6, and now Vasyukov, to my surprise, quickly played 1 7 i.h3. There immediately followed
1 7 .. . :txd2!. 1 8 'ii'xd2 1:!d8 19 'ii'c1 20 ':'xd1 l:txd 1 'iff6!
Now Black's idea becomes clear. He is threatening both 21 . . . h6 (the knight has no retreat square), and also 21 . . . 'ifxf2+ 22 'iti> h 1 �e3.
Tal's clever combination is certainly tempt ing. But is it correct? After all, even after the capture on f2 he will only have a pawn for the exchange.
Does it pay to sharpen the Play?
ltJ
87The grandmaster gives the variation 2 1 i.g2 'ii'xf2+ 2 2 �h 1 iLe3 23 ltJh3 i.xc1 24 ltJxf2 i.xb2 with the better endgame for Black. Another try, 2 1 i.d7 'ii'xf2+ 22 �h 1 i.e3 23 i.xc6, is refuted by 23 . . . i.xg5! with the terrible threat of 24 .. .'iVf3+.
However, it is possible to defend more strongly. Both the players and the commen tators overlooked the simple move 21 1:[d3! , taking control o f the important f3- and e3- squares. After 27 .. . 'ii'xf2+ 28 �h 1 ltJc4?! (28 .. . h6? 29 1:[f3 �e2 30 i.f1 , and the queen is trapped ) White has a pleasant choice between 29 1:[f3 'ii'xb2 30 'ii'xb2 liJxb2 31 .l:!.xf7 (31 ltJxf7! is simpler) 31 . . . h6 32 i.e6 �h8 (32 .. . hxg5 33 1:[f2+) 33 J::txc7 hxg5 34 l:txb7 and 29 'ii'f1 'ii'xf1 + 30 i.xf1 liJxb2 (30 . . . ltJd6!?) 31 .u.d7. I n both cases it is doubtful whether Tal would have been able to save the game. As Dvoretsky pointed out, Black's play can be improved by 28 .. . iLe7! 29 ltJc3 ltJc4, but in the ending arising in the variation 30 ltJd2!? i.g5 3 1 'iNf1 'iVxf1 + 3 2 ltJxf1 ltJxb2 3 3 J::td7 White's chances are better.
Vasyukov also devised a reasonable idea, but even so it was much inferior to 21 lid3!.
21 iLf5?! 96 22 b4!
The play has become much sharper. Tal now has a choice of three or four possibili ties. Which of them is the strongest?
In the game 22 .. . axb4 23 cxb4 i.e7 was played . But why not place the bishop on d4? It turns out that in the variation 23 .. . i.d4 24 b5 ltJa5 25 .1i.g4! 'ii'xf2+ 26 � h 1 i.e3 there is the defence 27 ltJh3! . White's position relies on this tactical nuance.
In which version is it better to retreat the bishop to e7, immediately or after the pawn exchange on b4? What is the difference? I n each case, i f he wishes, Black obtains two pieces for a rook, but it is important that the opponent should not be able to activate his forces. After 22 . . . iLe7!? 23 i.d7 for the moment the c-file is closed and there is the excellent resource 23 .. . ltJb8 ! . However, the consequences are far from clear: 24 i.g4 'iWxg5 25 'ii'xg5 i.xg5 26 bxa5 ltJc4 27 i.c8! ltJxa5 28 i.xb7! (28 .. . ltJxb7 29 l:tb1 ). [By playing 2B ... c6! 29 iLcB (with the threat of 30 :r:Lb 1) 29 .. . iLe 7!, Black retains the better chances, since he prevents the invasion of the rook and securely blockades the passed a-pawn. On the other, instead of 24 iLg4 White can try 24 i.b5!? (24 .. . c6 25 bxa5 ltJ 6d7 26 a6!? bxa6 27 i.xa6) 24 .. . 'ii'xg5 25 "fixg5 iLxg5 26 bxa5 ltJcB 27 f4! with a complicated and double-edged ending - Dvoretsky.]
22 . . . 23 cxb4
axb4 i.e7
[I think that the simple 23 .. . i.xb4 deserves serious consideration, with good compen sation for the sacrificed exchange - Dvo
retsky]
24 i.d7 ltJd4
Tal writes: 'Black does not want to simplify the position and he avoids 24 .. . 'ii'xg5 25 i.xc6 'ii'xc1 26 1:!xc1 bxc6 27 �xc6 iLd6, continuing to devote his main attention to the kingside. '
There is no point in going in for the variation 24 .. . ltJxd7 25 1:!xd7 'ii'xg5 26 'ii'xg5 iLxg5 27 :xc7 (stronger than 27 b5 ltJa5 or 27 a3 i.c1 ) 27 .. . ltJxb4 28 a4 b6 29 1:!b7, when White obtains a dangerous passed a-pawn .
25 'ii'xc7 26 'ifxb7
�d8 'it'xg5 Objectively, Tal's decision was risky - he has allowed his opponent not one, but two passed pawns on the queenside. However, he has kept the queens on. I n an endgame with rook and pawn against two minor pieces, a very important factor is the presence of an outside passed pawn , but in the middlegame there are chances of creating an attack on the king - after all, Black has one piece more. (However, it is not at all easy to include the knight at b6 and bishop at d8 in the attack.) I n addition, Tal always handled his strongest piece with great skill - it is sufficient to remember his famous game against Oscar Pan no from the 1 958 I nterzonal Tournament in Portoroz.
27 �e8 'ii'f6 28 a4 �f8!
It is important to drive away the bishop, in order to free the black queen for active play.
29 �b5 'ifd6 Attacking b4 and threatening 29 .. . lDf3+.
30 lIc1 ?
I n the time scramble White blunders his main pawn on a4. After 30 .l:.b1 lDxb5 31 axb5 'ifd3 32 l:tc1 Black would not have stood worse (32 .. . 'ifxb5?! 33 'it'b8 'iWd3 34
'ifxe5). Stronger, apparently, was the simple 30 �g2 'ifxb4? (as pointed out by Dvoretsky, 30 .. . �g7! 3 1 a5 'ifxb4 32 �e8 'ifb3 leads to a draw) 3 1 'iWb8 'ife7 32 a5.
30 . . . lDxa4! 31 �g2 31 �xa4 lDe2+. 31 . . . 32 l:[c5 lDb6 'iff6?
A mistake in reply. Here is Tal's explanation: 'Here White unexpectedly offered a draw. Somewhat confused, I forgot about the intended 32 .. . �g7, which would have given an easy win, and instantly replied 32 .. . "ikf6?. ' A conclusion about the importance of 'resistance to interference' suggests itself. I ncidentally, look at the position after 32 . . . �g7! . How well the black knights are placed - together with the remaining pieces they control all the invasion squares!
33 'iWb8t
White gains saving counterplay. 33 . . . 'ii'f3+? [ 'Mistakes never come singly! It was not yet too late to play 33 .. . �g7!, for example: 34 'ifxe5 (34 ':'xe5 �e7 with the threats of 35 .. . �d6 and 35 .. . �xb4) 34 .. . 'it'xe5 35 ':'xe5 �c7 36 :c5 �d6, retaining winning chances - Dvoretsky.] Draw. 34 �g1 35 �g2 36 �g1 37 ':'c6 38 �g2 39 ':'d6 40 �g1 41 �g2 'ifd1 + 'iff3+ lDe6 'ifd1 + 'ifd4 'ifxe4+ 'ifb1 +
Throughout the game Tal constantly took risks. First he d isrupted the balance, by conceding the advantage of the two bishops for the sake of rapid development. Then he
Does it pay to sharpen the Play?
ttJ
89decided on a dubious combination, and finally, in search of attacking chances he allowed the opponent to obtain two con nected passed pawns. Such was his style of play at that time!
Ribli - Tal
Candidates Tournament, Montpellier 1 985 Reti Opening 1 ltJf3 d5 2 g3 i..g4 3 �g2 c6 4 b3 ltJd7 5 i..b2 ltJgf6 6 0-0 e6 7 d3 i..c5 8 ltJbd2 0-0 g e4 dxe4 1 0 dxe4 e5
The structu re of the position is roughly the same as in the previous game. The only difference is in the placing of Black's queen's knight (there it stood at c6,whereas here this square is occupied by a pawn ) and White's dark-square bishop. These changes are rather to Black's advantage.
It is probable that on this occasion too Tal will have to exchange his g4-bishop for the
knight, conceding to his opponent the advantage of the two bishops. However, here this does not play a particular role.
11 h3
11 'ife2 followed by ltJd2-c4-e3 was prefer able. 11 . . . 1 2 'ifxf3 1 3 nad 1 ?! i..xf3 'iVe7
Another routine move, after which White is already in some d ifficulties. He should have placed his pawn on a4, preventing not only the exchange of the dark-square bishops by 1 3 .. . i..a3, but also 1 3 .. . b5, which deprives his knight of its lawful c4-square.
1 3 . . . b5! 14 h4?!
One mistake often leads to another. Appar ently Zoltan Ribli remembered about his light-square bishop and decided to bring it out to h3. But in so doing he weakens the g4-square.
White's primary objective is to rearrange his badly placed knight on d2. There is only one route available to it: via f1 to e3. This means that the correct move was 14 :fe 1 ! .
1 4 . . . as
The g4-square cannot be occupied immedi ately: if 14 . . . 'iVe6 there is the reply 1 5 'ii'f5! . Therefore for the moment Tal harasses his opponent on the queenside, by preparing 1 5 .. . a4.
1 5 c3
In order to answer 1 5 .. . a4 with 1 6 b4. But allowing the opening of the a-file would have been the lesser evil, since now another important square is weakened - d3.
1 5 . . . ltJb6 1 6 1:[fe1 ?
1 6 "'e2 or 1 6 i..h3 was better. (see diagram)
Black's last few moves have entirely fo cused the opponent's attention on the queenside - he has forgotten about possi ble diversions on the opposite side of the board and incautiously weakened his f2- point. This is immediately exploited by Tal , who, i t would appear, always remembers about the enemy king.
16 . . . 1 7 'ii'f5 1 8 .:te2 'ii'e61 li:)g4 .l:.ad8
White's position is already d ifficult. If 1 9 .th3 Tal was intending 1 9 .. J:td3! (with the threat of . . Jbg3+) 20 �g2 'ii'xf5 21 exf5 li:)xf2 22 ::'xf2 .txf2 23 �xf2 l::tfd8 24 �e2 e4. [In the event of 19 .th3? there is a simpler win by 1 9 .. . li:)xf2! 20 l1xf2 g6! - Dvoretsky.]
1 9 .ttl lld3! 20 �g2 li:)xf21 It should be said that Tal liked sacrificing two pieces for a rook. So that the previous game, in which completely the opposite balance of force arose, is rather an excep tion.
21 lbf2 .txf2 Can White somehow set up a defence?
21 'ii'xe6? fxe6 22 �xf2 l:tfxf3+ 23 li:)xf3 l:txd 1 is completely bad for White. After 2 1 �xf2? 'ii'd6 h e ends up in a mortal p i n on the d-file. The best chance of a defence was offered by 21 .te2! :xd2 22 :'xd2 .te3 (22 .. . .tc5 23 'iVxe6 fxe6 24 .tg4) 23 .l:.d3 .tc5 (23 .. . 'ii'xf5!? 24 exf5 .tc5 25 .tf3 %lc8) 24 'ii'xe6 fxe6 25 .tf3 (or 25 l:tf3). Black is a pawn up, but the win is still a long way off. It is amazing how great the safety margin is in chess - despite White's numerous errors, his position can still be held !
[The position can no longer be held! In the event of 2 1 .te2 the pretty stroke 21 .. . li:)d5!! is decisive: 22 �xf2 (22 'ii'xf2 li:)e3+ 23 �g 1 li:)xd1; 22 'ii'xe6 li:)e3+) 22 .. . li:)e3 23 .txd3 li:)xd1 +24 �e2 li:)xb2 - Dvoretsky.]
22 �xf2 'ii'd6
23 .tc1 g6
23 .. . 'ii'c5+ followed by 24 .. .... xc3 was also strong.
24 'ii'g5 f6!
Before the f-file is opened , the white queen must be driven away. 24 .. .f5 is unconvinc ing: 25 �g2 l:txf3 26 �xf3 "'d3+ 27 �f2 fxe4+ 28 �g 1 (Tal) 28 .. . e3 29 'ii'g4! (weaker is 29 'ii'xe5 li:)d7! 30 'ii'e6+ �g7) 29 .. . li:)d7!? (29 .. . e3 30 'ii'e6+ with a draw) 30 li:)e4! e2! 31 :e1 l:tf1 + 32 �g2 l:txe 1 33 'ii'e6+, and the battle ends in perpetual check (Dvo retsky).
Does it pay to sharpen the Play?
ttJ
9125 'ifh6 f5 26 'it;Jg2
If 26 'it;Je1 (e2), then 26 .. .f4 27 gxf4 ':xf4 is strong , while if 26 1:1e 1 - 26 .. . fxe4 27 ':xe4 l:lxc3.
26 . . . nxf3 1 27 lLlxf3
Now after 27 'it;Jxf3 'ifd3+ 28 'it;Jf2 fxe4+ 29 'it;Jg 1 Black has 29 . . . 'ifxg3+ 30 'it;Jh 1 ':f2 .
27 . . . 'ifxd1 28 lLlg5
And White resigned, since he is the ex change and a pawn down , and the mate threat is easily parried by 28 .. . 'ifhS or 28 . . . 'ifd7.
As you see, although the mature Tal had not lost his former resourcefulness, and he still liked to attack and make combinations, he did this on a strict positional basis, endeav ouring not to take the 'liberties' typical of his youth .
The following game is quite d ifferent in character to those examined earlier - it bears directly on the question of prophy laxis. This is not surprising - playing White was Tigran Petrosian.
Petros ian - Ivkov Olympiad , N ice 1 974 Nimzo-Indian Defence 1 d4 lLlf6 2 c4 e6 3 lLlc3 Jt.b4 4 e3 c5 5 Jt.d3 lLlc6 6 lLlf3 i.xc3+ 7 bxc3 d6 8 e4 e5 9 d5 lLle7 10 lLld2
prophylactic move 1 0 .. . h6!, so that if the knight moves he has the reply 11 . . . gS, preventing the opening of the position by f2-f4. The Sth game of Spassky-Fischer world championship match (Reykjavik 1 972) went 1 0 lLlh4 h6 11 f4 (hoping for 1 1 . . .exf4 1 2 i.xf4 gS 1 3 eS! with complications favourable to White) 1 1 .. . lLlg6! 1 2 lLlxg6 fxg6. Here Boris Spassky made a serious strategic mistake by exchanging pawns on eS, after which the position became static and the white bishops had no scope. Robert Fischer outplayed his opponent and went on to win .
1 0 . . . 0-0
Black also has another possibilities. The move made clearly shows his intention - to prepare . . . f7-fS.
1 1 lLlf1
Petros ian takes the opponent's plan into consideration and takes measures before hand , by transferring his knight to g3. Black should possibly have changed plan by playing 1 1 .. . lLlg6, in order to have the option of jumping with his knight to f4. White in turn can react flexibly to this move, by placing his knight on e3 and his pawns on g3 and f3, and then advancing his h-pawn , taking advantage of his delay in castling. The In the event of 1 0 0-0 Black makes the move order chosen by Petrosian is quite
venomous, but in recent times for some reason it has not been employed .
1 1 . . . 'iVa5
Black has decided to play on the kingside, and it is not clear why he moves his queen to the queenside. If he was going to develop his queen at as, he should have done this a move earlier, when , firstly, there was not the reply .i.d2, and secondly, he would have retained the option of castling on the queenside. 1 2 .i.d2 1 3 tbg3 14 exf5 tbe8 f5 tbxf5
Black has a difficult position after 1 4 .. . .i.xfS 1 S tbxfS tbxfS 1 6 'if c2.
1 5 'ii'c2 ! 1 6 0-0
g6 .i.d7
It is important to note that, in contrast to the King's I ndian Defence, Black's knight can not go to d4 - the square is defended by the white pawn.
17 tbe4
1 7 f4!? suggested itself, in order to open up the position and exploit the power of the two bishops. Possibly White was concerned about 1 7 .. . tbxg3 1 8 hxg3 e4!? ( 1 8 .. . exf4 1 9 .i.xf4, intending ':ae1 and at some point .i.xg6) 1 9 .i.xe4 tbg7! ( 1 9 . . . .i.fS 20 g4!?;
19 ... tbf6 20 .i.xg6!? hxg6 2 1 'iVxg6+ �h8 22 fS with a powerful attack). But the compen sation for the pawn is hardly sufficient. Black probably does better to reject the pawn sacrifice in favour of 1 8 .. . tbg7!? 1 9 fxeS dxeS.
Petrosian wants to preserve his knight from exchange and so he does not hurry to take decisive action. However, the opponent gains time to strengthen his position .
1 7 . . . tbf6 1 8 tbg5
The knight is very strongly placed here, since the attempt to drive it away by . . . h7- h6 leads to a weakening of the kingside.
1 8 . . . ':ae8 19 f3 !
A typical Petrosian move. Having taken control of the e4- and g4-squares, he is ready at a convenient moment to play g2- g4, depriving the enemy pieces of the fS point.
19 . . . 20 g4!
tbg7
Of course, the ex-world champion prevents the exchange of bishops by 20 . . . .i.fS planned by the opponent. All the black minor pieces are now shut out of play, and yet for the moment the situation remains unclear. For
Does it pay to sharpen the Play?
ttJ
93complete happiness White still needs also to cramp Black on the queenside with a2- a4.
20 . . . 'ifa4
Borislav Ivkov misses an excellent chance to complicate the play, pointed out by Petrosian: 20 .. . b5! 2 1 cxb5 c4 22 �xc4 �xb5.
21 'ifb31 :b8 22 .i.e21
23 a4
'ifaS
Thus, White has also succeeded in restrict ing the opponent's possibilities on the queenside. But even now the battle is not yet over.
23 . . . 'ife7 24 h3
White has to support the g4-pawn , to prepare f3-f4.
24 . . . a6
2S aS
Otherwise White would have had to reckon with 25 . . . b5, and after the capture on b5 with the c-pawn - . . . c5-c4.
2S . . . bS 26 axb6
What do you think, with which piece should Black capture on b6? In whose favour is the exchange of queens?
The answer is clear: in Black's favour, of course. In cramped positions you should exchange pieces! I think that after 26 .. . 'ii'xb6! (27 'ifa2 Wb2) he would have retained excellent drawing chances.
But now remember the situation before White's 1 7th move. I should like to ask: have Petrosian's subtle manoeuvres been justi fied? Wouldn't it have been simpler, by playing 1 7 f4!?, to immediately 'cut the Gordian knot'?
[At any event he should have struck in the centre, without waiting for the opening of lines on the queenside: 25 f4! (instead of 25 a5?!) 25 .. . exf4 26 :'xf4 or 25 .. . b5 26 axb5 axb5 27 fxe5 dxe5 28 .i.e3 with advantage to White - Dvoretsky.]
26 . . . lIxb6? 27 'ifa3
White switches his queen to the kingside for an attack, whereas the black queen lacks any prospects.
27 . . . Wd8 28 'ife1 We7
The rook on its own cannot do anything, and it is quickly driven off the second rank.
30 �d3 �e8
31 �e1 lIb3 32 .i.e2 :b6
The prophylactic work has been success fully accomplished . There now follows what is effectively the first active move in the game, and Black's position immediately collapses. 33 f4! h6 34 fxeS 3S 'ii'xeS 36 ttJe4 37 �a3 WxeS dxeS hS It is time to gather the harvest.
37 . . . ttJxe4 38 .:txf8+ 'iitxf8
39 i.xe4 40 i.xe5+ 41 'uf1 Black resigned . �b3 �e8
In conclusion I will take the liberty of offering one of my own games.
Tsariov - Vulfson Moscow 1 989 Sicilian Defence 1 e4 2 tbe3 3 f4 4 tbf3 5 d3 e5 tbe6 e6 d5
Apparently my opponent was satisfied with the endgame after 5 . . . dxe4 6 dxe4 . I was aiming for more complicated play.
5 . . . tbf6 6 e5
7 g3
tbd7 b5!?
Usually this advance has to be prepared , but here there is an opportunity to carry it out immediately.
8 i.g2 b4
9 tbe2 g6?!
It would have been better to continue in the