Chapter 40
The attack on the uncastled king
A king that has not castled is almost always bad. In the opening and the middlegame, the battle usually takes place in the centre, and if the most important object of attack is located there, with the rooks unconnected and his own piece coordination disrupted, then it is bound to be dangerous. In many games, players are willing to part with material to keep the enemy king in the centre for longer.
221 *
Paul Keres
Istvan Bilek
Leipzig ol 1960 (4)
Black has not managed to evacuate his king (the queenside looks relatively the safest), and the game is now decided by energetic play along the central files:
15.♗xd5! exd5 16.♘xd5 ♕d6 17.♘xc6
White also keeps a fierce attack after 17.♘b3.
17…♗xc6 18.♗c5! ♕xc5
18…♗b5 19.♕e1 ♘d3+ 20.♖xd3 ♕xc5 21.b4 ♕d6 22.♘xe7+–.
19.♕xe5 f6 20.♘xf6+ ♔f7 21.♕e6+ ♔f8 22.♖he1 ♖d8 23.♖xd8+ ♗xd8 24.♘d7+
1-0
222 *
Dariusz Swiercz
2585
Etienne Bacrot
2705
Germany Bundesliga 2011/12 (13)
Black has played the opening poorly, leaving his king in the centre, and trying to confuse White from the opening moves. He could have easily been punished for this:
16.fxe5!
In the game, there followed 16.♕c6, which also prevents Black from castling and retains the advantage, but it is rather less convincing.
16…♘xe5 17.♘xf6+ exf6 18.♕c5
Now, castling is impossible, and the queen on a5 is just a spectator to the white mating attack.
18…♘d3+ 19.♖xd3 ♖xd3
20.♘d4!
This move closes the black king’s path to the queenside.
Less convincing is 20.♖e1+ ♔d8 21.♘d4 ♕c7 22.♕d5+ ♕d7 23.♕a8+ ♕c8 24.♕e4 ♕c4 25.♔b1
analysis diagram
25…♖xd4 (quite hopeless is 25…♔c7 26.♕e7+ ♔b8 27.♕d6+ ♕c7 28.♕xa6 ♖xd4 29.cxd4) 26.cxd4 ♕e6 27.♕a8+ ♕c8 28.♕a7 ♖e8 29.♕b6+ ♔d7 30.♖c1 ♕b8 31.♕xf6 ♕b7 32.♕xf7+ ♖e7 33.♕f6, and White still has some work to do.
20…♖e3 21.♘c6
21.♘f5 wins even more effectively.
21…♕xa2 22.♕xe3+ ♔d7 23.♖d1+ ♔xc6 24.♕e4+ ♔c7 25.♕e7+ ♔c8 26.♕c5+ ♔b8 27.♕b6+ ♔a8 28.♕c6+ ♔b8 29.♖d7
And mate.
223 **
Konstantin Sakaev
2600
Andrei Kharlov
2605
Budapest tt 1996 (3)
1.d4 d5 2.♘f3 ♘f6 3.c4 dxc4 4.♘c3 a6 5.e4 b5 6.e5 ♘d5 7.a4 e6 8.axb5 ♗b4 9.♕c2 ♘b6 10.♕e4 ♕d5 11.♕g4 axb5 12.♖xa8 ♕xa8 13.♗e2
White has a serious initiative for the pawn, thanks to Black’s inability to castle. Even so, the latter needs to save his kingside and keep the character of the position closed, as in this case, it will hard for White to get at the black king. Instead, the attempt to create counterplay…
13…♘d5?
turns out badly.
It was essential to play 13…♔f8 or even 13…g6.
14.♕xg7 ♖f8
Black had calculated that White could not play 15.♗d2, because of the check at a1. But White simply ignores the queenside, evacuating his king to a safe place, and then begins an attack on the kingside:
15.0-0! ♗xc3 16.♘g5!
Black had only counted on 16.bxc3 ♘xc3 17.♗d1 ♕e4 18.♗a3 ♘d7, with an unclear battle.
16…♗b4 17.♗h5
White breaks through to the black king, who even fails to escape to relative safety on the queenside.
17…♔d7 18.♘xh7 ♖e8 19.♗xf7 ♖e7
20.♗xe6+! ♔c6
20…♔xe6 21.♘f8+ ♔f5 22.♕g6#.
21.♗xd5+ ♔xd5 22.♕g3
The black king is fighting on his own against the white army, and the outcome of the game is decided.
22…♖f7 23.♘f6+ ♖xf6 24.exf6 ♕a6 25.♕f3+ ♔e6 26.♗g5 ♕b7 27.♕g4+ ♔d5 28.♕f4 ♘d7 29.f7 c3 30.bxc3 ♗d6 31.♕f3+ ♔c4 32.♕e2+ ♔d5 33.♗e7 ♕b6 34.f8♕ ♘xf8 35.♗xf8 ♗xf8 36.♕f3+ ♔c4 37.♕xf8 ♗b7 38.♕b4+ ♔d5 39.♖e1
1-0
224 **
Evgeny Bareev
2655
Alexei Dreev
2645
Russia tt 1996 (4)
Black has dawdled in the opening and delayed castling. This allows White to launch an assault:
16.d5! exd5 17.♗e3!
A very important move, allowing White to complete his development with tempo, without giving Black a breathing space.
17…♗c5
17…♕d8 18.♗b5+ ♔f8 19.e5 ♘e4 20.♘d4 – because of the undeveloped ♖h8, Black’s position is hopeless.
18.♗xc5 ♕xc5 19.♖ac1 ♕b6 20.♗b5+ ♔f8 21.e5 ♘e8?
Very passive. Black hopes to transfer the knight via c7 to e6, but does not succeed in doing this.
More tenacious was 21…♘e4, although after 22.♖fd1, followed by ♘f3-d4, White also has a clear advantage.
22.♖fe1!
White is ready for a further opening of lines.
22…♘c7 23.e6! ♘xe6 24.♘e5 ♔g8 25.♘xf7! ♔xf7 26.♖c6 ♕d8 27.♕xe6+ ♔f8 28.♖d6
1-0
225 **
Zaven Andriasyan
2615
Alexei Dreev
2711
Rogaska Slatina tt 2011 (6)
In an open position, White has two splendid bishops, and Black has not yet castled. This gives rise to the idea of saving a tempo on retreating the rook. Indeed:
24.♖xd7!
In the game, White retreated with 24.♖d2, and after 24…0-0 25.♗b2 ♘f6 26.♖xd8 ♖xd8 27.♕d2 ♖e8 28.♗f3 ♕g5 29.♕xb4 h5, Black obtained a slightly worse, but defensible position, which he duly held confidently.
24…♖xd7 25.♗f5 ♖d8 26.♕e4+ ♔f8 27.♕xb4+ ♔g8 28.♗b2
The white bishops have managed to obtain great operating space. Black is playing without his ♖h8, and White will easily be able to develop his initiative subsequently. Given the additional weakness of the back rank, Black has no chance of saving himself, for example:
28…♕g5 29.♕c5 h6 30.f4 ♕h4 31.♕e5 ♕f6 32.♕xf6 gxf6 33.♗xf6
With a winning position.
226 **
Alexei Dreev
2630
Evgeny Bareev
2675
Elista 1998 (5)
The white king is in the centre, and with his next blow, Black cuts the board into two sections, finally destroying the coordination of the white forces, and opening the e-file for his rooks:
15…e4! 16.c4 e3!
In the event of the unhurried 16…♖fe8, White would have chosen 17.♘f1, intending to put his knight on e3 – the game would have become unclear (of course, not 17.e3, which weakens the light squares. Black could exploit this immediately with 17…♘b7!, heading to c5).
17.fxe3
17…♖fe8!
He needs in the first instance to attack the weaknesses with those pieces which are not taking part in the game, and bring them into the game with tempo.
17…♕e5 is weaker, not just out of general considerations, but also for a concrete reason: after 18.0-0!, none of the possible captures brings Black any benefit.
18.e4 ♕e5 19.♔f2
More tenacious is 19.0-0, although after 19…♕xg3 (19…♗xh3 20.♗xh3 ♕xg3+ 21.♗g2 ♕e3+ 22.♔h1 ♖e5 23.♖f5) 20.♕a3 ♕g6,
analysis diagram
Black has the advantage anyway – e4 hangs, and the rook threatens to transfer via e5 to the kingside.
19…♘h5 20.♕a3
20…♕d4+!
Forcing the advance of the pawns, and thus blocking the white pieces.
21.e3 ♕e5 22.g4 ♕f6+ 23.♗f3 ♕h4+ 24.♔g1 ♘xc4 25.♕d3 ♕g3+ 26.♗g2 ♘f4 27.♕f1 ♕xe3+ 28.♔h2 ♘xd2 29.♕d1 ♘xg2 30.♗xd2 ♕xe4 31.♔g3 ♘e3
0-1
227 **
Krum Georgiev
2465
Garry Kasparov
2595
Malta ol 1980 (5)
Obviously, White’s hopes are bound up with play in the centre, where the black king is caught. But he needs to play exceptionally energetically and accurately, as he has already sacrificed one piece:
13.♘xe6!
There is no time to prepare this blow, as after 13.♕h3?, there is the strong 13…b3!, seizing the initiative.
13…fxe6 14.♕h3!
On another square, the queen would be attacked with tempo: 14.♕g4 ♔f7 15.f5 ♘xe5!
14…♔f7
Completely bad is 14…♗d5 15.♖xd5.
15.f5 ♗e4
Other replies are also in White’s favour:
A) 15…♘xe5 16.fxe6+ ♔g8 17.e7 ♗xe7 18.♕e6+ ♘f7 19.♗xe7;
B) 15…b3 16.fxe6+ ♔g8 17.♕xb3 ♘c5 18.♕c4 d5 19.♖xd5;
C) 15…♔g8 16.fxe6 ♗e4 – 15…♗e4 16.fxe6+ ♔g8.
16.fxe6+ ♔g8
On 16…♔g6, there follows 17.♗d3.
17.♕b3 ♗xc2
Nor does the following variation offer any chances: 17…♘c5 18.♕c4 d5 19.♖xd5 ♘xe6 20.♖c5! ♕f7 21.♖f1.
18.♕xc2
An even easier win was 18.exd6 ♗xd6 19.♕xc2.
18…♕xc2+ 19.♔xc2 ♘xe5 20.e7 ♗xe7 21.♗xe7
White wins material, and retains much the more active pieces. The rest is easy.
21…♘bc6 22.♗xd6 ♖a5 23.♖d5 ♖xb5 24.♖xb5 ♘d4+ 25.♔b1 ♘xb5 26.♗xe5 ♔f7 27.♖c1 ♔e6 28.♖c5 ♘d6 29.♔c2 ♖f8 30.♗d4 ♖f1 31.♖e5+ ♔d7 32.♗c5 ♔c6 33.♗xb4 ♖f2+ 34.♔c1 ♘b7 35.♖g5 g6 36.a3 ♖e2 37.♔b1 ♖f2 38.♔a2 ♘d8 39.♖c5+ ♔b7 40.♖d5 ♘c6 41.♗c3 ♖xg2 42.♖d7+ ♔b6 43.♖xh7 ♔b5 44.♔b3 ♖g4 45.♖d7 ♖h4 46.♖d2 ♖h3 47.♖g2 ♘d4+ 48.♔a2 ♘f5 49.♗e5 ♖e3 50.♗b8 ♖e6 51.♖g4 ♘e7 52.♗g3 ♘c6 53.a4+ ♔c5 54.♔a3 ♖f6 55.b3 ♔d5 56.♖g5+ ♔e4 57.♖c5 ♖e6 58.a5 ♔d3 59.b4 ♘d4 60.♗e5! ♘f3 61.♗b2 ♘d2 62.♖c3+ ♔e2 63.♔a4
Black resigned.
228 **
Alexander Beliavsky
2635
John Nunn
2615
Wijk aan Zee 1985 (2)
Structurally, Black is not doing so well, so he has to rely on dynamics, exploiting his lead in development:
11…♖xf5!!
If White delays, Black has numerous ways to develop his initiative: in the first instance, …♕d8-f8 is strong, ensuring h6 for the bishop, and strengthening the pressure on the f-file. Also possible is the simple retreat of the rook to f8, followed by the breaks …e7-e6, or …a7-a6, or …b7-b5.
12.g4
What else? He needs somehow to develop, but he cannot, and White decides to take the piece.
On 12.♘e4, there is the strong reply 12…♖f4!, with threats of …♗g7-h6 and …♗c8-f5.
Black also develops a decisive attack after 12.0-0-0 ♖f4! 13.g4 ♘xc4 14.♗xc4 ♖xc4 15.gxh5 ♕a5 16.♘ge2 ♕xa2.
12…♖xf3! 13.gxh5 ♕f8! 14.♘e4
The knight is already developed, and so it is wrong to move it again, without extreme necessity.
14.hxg6 hxg6 15.♖d1 ♗f5 16.♖h2
16…♖e8!!. When conducting an attack, it is especially important to bring in undeveloped pieces. Only an understanding of this principle can enable one to find such a move. 17.♖g2 e6 18.♗e2 ♖xf2 19.♖xf2 ♗h6 20.♖f4 ♘xc4 21.♗xc4 e5 – for the piece, Black has the advantage of the two bishops, a pair of pawns, and a strong attack. Even so, White can manage to defend.
14…♗h6
Black could have won immediately by 14…♕f5! 15.♘g5 ♖f4 16.h4 h6.
15.♕c2
15…♕f4?
Black misses the chance to break effectively into White’s camp: 15…♗e3! 16.hxg6 hxg6 17.♕e2 ♖xf2! 18.♘xf2 ♕f4, with a fearful attack.
16.♘e2?
In the first instance, one should look at chances to eliminate attacking units. And there was such a chance here, but White passes by it, trusting his opponent that the rook cannot be taken: 16.♘xf3! ♘xf3+ (if 16…♕xf3 17.♘g3!, the attack fizzles out) 17.♔d1 ♗f5 18.♗g3 ♕e3 19.♕d3 ♗xe4 20.♕xe3 ♗xe3 21.♗g2 ♖f8 22.♖f1. Black is pinned, and has lost all his attacking potential – the roles have changed, and White has the advantage!
16…♖xf2! 17.♘xf2 ♘f3+ 18.♔d1 ♕h4 19.♘d3 ♗f5
All the black pieces are attacking, whilst White’s forces totally lack coordination. The extra white rook is not felt at all.
20.♘ec1 ♘d2! 21.hxg6 hxg6 22.♗g2 ♘xc4 23.♕f2 ♘e3+ 24.♔e2 ♕c4 25.♗f3 ♖f8 26.♖g1 ♘c2 27.♔d1 ♗xd3
White resigned.
229 **
Vasily Ivanchuk
2739
Parimarjan Negi
2615
Greece tt 2010 (1)
Black has wasted many tempi in the opening on non-developing moves, even ‘forgetting’ to castle. The knight has already moved three times to get to c5, which is an impermissible luxury, as Black’s development is far from complete. The white pieces are all mobilised and ready for the attack.
14.e5!
Starting with a piece sacrifice was inaccurate: 14.♘dxb5 axb5 15.e5 b4! leads to unclear consequences.
14…♗xe5
If 14…dxe5 15.♘dxb5 ♕xd2+ 16.♖xd2, little changes – the queenside pawn armada decides the game.
15.♘dxb5! ♘b3+
15…axb5 16.♗xc5 ♘c6 17.♗xd6 ♗xd6 18.♕xd6 ♕xd6 19.♖xd6 b4 20.axb4 also leads to a winning position for White, although it prolongs the game somewhat.
16.cxb3 axb5 17.♗xb5+ ♗c6 18.♗d4
White has a safe king and an attack, everything he could want.
18…♕c8 19.♗xe5 dxe5 20.♔b1 ♖g8 21.♘e4
Black resigned.
230 ***
Leonid Stein
Semyon Furman
Moscow ch-URS 1969 (18)
White has already mobilised almost all of his forces, whilst Black is some way behind with his development. The sacrifice on d5, followed by play on the e-file, is tempting, but before playing it, White first prepares a target on the queenside, where the black king will soon end up.
10.a4! b4
If 10…bxa4, 11.♘d5 gains further in strength.
11.♘d5! exd5
If Black declines the knight, then after 11…♘xd5 12.exd5 e5 13.a5, the pawn on b4 is cut off from its own camp and will be lost.
12.exd5+ ♔d8 13.♗g5 ♘bd7 14.♕e2
White had an alternative means of developing the attack: 14.♘c6+ ♗xc6 (14…♔c8 15.♗xf6 ♘xf6 16.♗h3+) 15.dxc6 ♘c5 16.♗e3 ♔c8 17.a5!, with the threat of c2-c3.
14…♔c8
15.c3! b3 16.♘c6
He also wins with 16.♕e8+ ♕d8 17.♕xf7 ♘e5 18.♗h3+ ♔b8 19.♖xe5 dxe5 20.♘c6+ ♗xc6 21.dxc6 ♖a7 22.♕xb3+ ♔a8 23.♖d1.
16…♗xc6 17.dxc6 ♘e5 18.♖a3
Playing in the centre with 18.♖ad1 also looks good, retaining the possibility of transferring the rook to the queenside via d4.
18…d5 19.♖xb3 ♗d6 20.♗xf6 gxf6 21.♗xd5 ♘xc6
21…♖a7 22.♖b7 ♖xb7 23.cxb7+ and ♕xa6.
22.♕g4+
Going into an endgame allows Black to prolong his resistance.
A more convincing way to win was 22.♕f3 ♖b8 23.♖xb8+ ♘xb8 24.♖e4, with a decisive attack.
22…♕d7 23.♕xd7+ ♔xd7 24.♖b7+ ♗c7 25.♗g2 ♖ad8 26.♗h3+ ♔d6 27.♖d1+ ♔c5 28.b4+ ♔c4 29.♗f1+ ♔b3 30.♖xc7 ♖xd1 31.♖xc6 ♔xa4 32.♔g2 a5 33.bxa5 ♖hd8 34.♖xf6 ♖8d7 35.a6 ♖c1 36.♗d3! ♖xc3 37.♗xh7 ♔b4 38.♗e4 ♖a3 39.♗b7 ♔c5 40.h4 ♔d4 41.♖f5 ♖e7 42.h5 ♖e5 43.♖f4+ ♔c5 44.h6
Black resigned.
231 ***
Konstantin Sakaev
2625
Mikhail Ulibin
2580
Dubai 2000 (8)
1.e4 e6 2.d4 d5 3.♘c3 ♘f6 4.♗g5 ♗e7 5.e5 ♘fd7 6.h4
The Chatard Attack was introduced into practice by Alekhine. White sacrifices a pawn, trying to get the chance to develop an initiative on the dark squares.
6…c5
Black declines the sacrifice, trying instead to break up the white centre.
More popular is 6…♗xg5 7.hxg5 ♕xg5 8.♘h3, with compensation.
But the most common move is 6…a6, taking b5 from the white knight and preparing …c7-c5.
7.♗xe7 ♔xe7
This extravagant king recapture is Black’s idea. Naturally, 7…♕xe7 is bad because of 8.♘b5.
8.dxc5
The black king is in the centre, so White tries to open it up, not minding sacrificing a pawn.
8…♘xe5 9.♕e2 ♘bc6 10.0-0-0 ♔f8 11.f4 ♘d7 12.♘f3 ♘xc5
‘Quiet’ moves / prophylaxis in the middle of an attack
None of the developing continuations bring White any special dividends, so he makes a useful move, which he cannot do without in the long run:
13.♔b1!
‘Just in case’, the king gets off the c1-h6 diagonal on which he could be checked. Furthermore, if the ♘c3 moves away to take part in the attack, the a2-pawn will now be defended against a black queen raid to a5. And finally, after 13.♕e3, there could follow 13…♕b6, with the threat of 14…♘b3+. It is always important to choose the correct moment for such prophylaxis – in this case, the single tempo that Black is given is not enough to enable him to consolidate his position.
13…b5
An active attempt to start counterplay. It is not clear what else Back can do, because he cannot castle by hand on the kingside.
14.♕e3
White is not in a hurry to take on b5, which gives Black free development.
14…♕b6
15.♖xd5! exd5
The tactical trick does not help Black; after 15…♘e4 16.♕xb6 axb6 17.♖xb5 ♘xc3+ 18.bxc3 ♖b8, White has a pawn advantage on the queenside.
16.♘xd5 ♕d8 17.♕xc5+ ♘e7 18.♘c3 a6
On 18…♕b6, there follows 19.♕b4, keeping the pin on the knight. White can take the b5-pawn in comfort, since after 19…a6 or 19…a5, he wins with 20.♕xe7+.
19.♗e2!
This is simpler than the ‘more active’ 19.♗d3, since here the bishop does not obstruct the d-file, and is ready later to exert pressure on the centre from the f3-square.
19…♗e6 20.♖d1 ♕c8 21.♕b4 h5
Black finds a way to bring his king’s rook into the game, but even so, it is not so well placed on h6, as it would be on an open file.
22.♘d4 ♖h6 23.♗f3 ♖a7
The best defence was 23…♖b8!, with the idea of …a6-a5!. After 24.f5 (White is only a little better after 24.♘c6 ♗xa2+ 25.♔xa2 ♖xc6 26.♗xc6 ♕xc6 27.♘d5 ♕d7) 24…a5 25.♕xa5 ♘xf5 26.♘cxb5 ♘xd4 27.♘xd4 ♗g4 28.b3, White has the advantage, but Black retains counterplay.
24.f5!
Despite the fact that the position is sharp, the decisive factor is the restricted situation of Black’s light-squared bishop.
White’s advantage is not so great after 24.♘xe6+ fxe6 25.♕a3 ♖d7 26.♖xd7 ♕xd7 27.♕xa6 b4 28.♘e4 e5.
24…♗c4 25.♘c6 ♖xc6 26.♗xc6 ♔g8 27.♗e4 ♘xf5 28.♕a5
28…♖d7
28…♔h7 29.a4! (in the endgame arising after 29.♕d8 ♕xd8 30.♗xf5+ g6 31.♖xd8 gxf5 32.b3 b4, Black retains drawing chances) 29…♖e7 30.b3 ♗e6 31.axb5 axb5 32.♘xb5 – the two passed pawns on the queenside should in time bring White victory.
29.♖xd7 ♕xd7 30.b3 ♗f1 31.♕xa6
The position gradually simplifies, and White’s queenside pawn majority on the queenside settles things.
31…g6 32.♘d5 ♕d6 33.♕xd6
Carelessness; 33.♘f6+ ♔h8 34.♕c8+ ♔g7 35.♘e8+ wins the queen.
33…♘xd6 34.♘e3 ♘xe4 35.♘xf1 ♔g7 36.♔b2 ♔f6 37.♔a3 ♘c3 38.♔b4 ♘xa2+ 39.♔xb5 ♔e5 40.♔c5 ♘c3 41.b4 ♘e4+ 42.♔b5 ♘c3+ 43.♔c4 ♘e4 44.♔b5 ♘d6+ 45.♔c6 ♘f5 46.b5 ♘d4+ 47.♔c5 ♘e6+ 48.♔b4 ♔d6 49.c4 ♘c5 50.♘d2 f6 51.♘b3 ♘d3+ 52.♔a5 g5 53.c5+ ♔c7 54.b6+ ♔b7 55.♔b5 gxh4 56.c6+ ♔b8 57.♘c5
Black resigned.
232 ***
Veselin Topalov
2640
Evgeny Bareev
2685
Linares 1994 (11)
Does the ♗f6 need to be moved away to avoid being exchanged? No, because the position is open, and every tempo allowing a piece to be developed is worth its weight in gold.
12…♕d5!
This move is strong, not only because the white ♖h1 would be hanging after the exchange on f6, but also because it destroys the coordination of the white forces and prevents White’s queenside castling.
13.a3?
White strives to castle queenside anyway, but he ought to have sensed the danger. Or maybe he did feel it, but was too upset to go over to a defensive posture so early in the game, something which is always difficult psychologically.
It was essential to complete his development, even if it means the white pieces having to take up passive squares. 13.♘e2 ♖e8 14.0-0 ♗h3 retains a small stable initiative for Black, but no more than that. Now, however, White’s position collapses.
13…♘xf4!
Maybe, Black did not at this stage calculate the final mating variation to the very end, but his intuition told him that his group of excellently-coordinated pieces ought to be able to mate the lone white king.
14.♘xf6+
It is already impossible to bail out at small cost, since after 14.0-0-0 (14.gxf4 ♗h4+ and 15…f5) 14…♘xd3+ 15.♕xd3, there follows 15…♗g5+! allowing him to avoid doubled pawns, with a technically winning position (on 15…♕xd3?, White has a zwischenzug: 16.♘xf6+! gxf6 17.♖xd3, with good chances of a draw).
14…gxf6 15.♗xh7 ♔g7 16.♕e4
Dynamics – deflection
16…♖e8!
The main defender of the white king is drawn off into the corner of the board, where it is just a spectator.
17.♕xe8 ♗f5!
This move also shuts off the white bishop, which was protecting the e4- and d3-squares in his own camp.
18.♕xa8 ♕e4 19.♔f2 ♕g2 20.♔e3 ♘d5 21.♔d4 ♕d2 22.♔c5 ♕e3 23.♔c4 ♘b6
White resigned, because after 24.♔b3 ♕e6, he is mated.
233 ***
Rashid Nezhmetdinov
Mikhail Tal
Baku ch-URS 1961 (15)
Castling kingside is impossible because of the ♘h6, so Black plans to castle queenside. White’s position has some holes that are no less serious than those in Black’s, and so he needs to play energetically:
17.♖xf6!
Other lines do not bring any benefit, and result only in unclear play:
A) 17.g5 ♗xh6 18.gxh6 ♘d4∞;
B) 17.♘d5 ♘xd5 18.exd5 ♘d4 19.♕f2 (19.♕d1∞) 19…f6 20.d6 ♕d7 21.♗xb7 ♕xb7 22.♗e3 0-0-0 23.c3∞.
17…♗xf6 18.♘d5
18…♕d8
No better is 18…♗xd5 19.exd5 ♘d4 20.♕f2.
19.♕f2 ♘f4
19…♗xd5 20.exd5 ♘f4 21.♗xf4 exf4 22.♕xf4; or 19…♘f4 20.♗xf4 exf4 21.♕xf4 ♗xd5 22.exd5.
20.♗xf4 exf4
21.♕xf4!
In the game, White played differently and ended up winning, but Black had a chance. After 21.e5, he had to play 21…♗h4! (the game ended: 21…♗xe5 22.♖e1 f6 23.♘xf6+ ♕xf6 24.♕d4, and White won) 22.♕d4 ♖f8 23.♖d1 ♖c8, with an unclear position.
21…♗xd5 22.exd5 ♕b6+ 23.♔h1 0-0-0
If now 24.♘xf7 ♗xb2 25.♖f1 (25.♖e1 ♖he8 26.♖xe8 ♖xe8 27.♘d6 ♔d8 28.♘xe8 ♔xe8 29.a4 ♕d4=) 25…♖hf8 26.a4 ♗a3 27.g5 ♕c7 28.♕f6 ♖xf7 29.♕xf7 ♕xf7 30.♖xf7 bxa4 31.♖f4 ♗d6 32.♖xa4 ♔b7, and Black has excellent chances of holding.
Surprisingly, one problem-like move ensures White the advantage:
24.c3!
Restricting the dark-squared bishop. After…
24…♖he8 25.g5 ♗e5 26.♕g4+ f5 27.gxf6+ ♔c7 28.♕f3
White keeps a strong initiative.
Additional material
Fischer-Tal, Bled/Zagreb/Belgrade 1959
Tal-Taimanov, Kislovodsk 1966
Tal-Velimirovic, Riga 1979
Kramnik-Shirov, Cazorla match/9 1998 – Black’s 21st move
Timman-Geller, Hilversum 1973
Alekhine-Junge, Prague 1942
Jussupow-Portisch, Tunis 1985
Tal-Larsen, Bled match/6 1965
Spielmann-Alekhine, Carlsbad 1911 – White’s 17th move
Kasparov-Salov, Barcelona 1989