Art of Attack in Chess

by Vladimir Vukovic

  9 Qg5 Nb3 10 Re1 Rc7

  Clearly, 10 ... h6 is bad because of 11 hxg6.

  11 Kg2 Raa7 12 Reh1 Rab7 13 Rh4! Rd7 14 Rf4 Rdc7

  Black is unable to parry the combination which White has prepared. If White had previously made an incorrect exchange on g6, ... Qe8 would now be the answer to everything; in the present position this move would be answered by 15 Qh6 Qf8 16 Qxh7+ Kxh7 17 hxg6+ Kxg6 18 Rg4+ Kf5 19 Rh5+ Kxg4 20 Rg5#.

  15 hxg6 fxg6 16 Qxg6+! hxg6 17 Rh8+ Kf7 18 Bd8+ and White finally wins.

  It is well worth remembering this analysis, especially the method of delaying the attack with the h-pawn is and how the winning position is prepared by a calculated deployment of the pieces. The reader will be reminded of this extremely important practical technique in the section on the role of the rook files in the attack on the castled king.

  As for the part played by the network of weak squares in attacks on the castled king, one more diagram and one interesting master game will be given as examples.

  This position is from the game Euwe-Flohr, Amsterdam 1939. White’s bishop has reached the square f6, but the network is still guarded by the opponent’s dark-squared bishop, which is preparing to effect a delayed fianchetto by ... Bg7. For this reason forceful play by White is indicated.

  1 Rc5!

  With the idea of eliminating the dark-squared bishop. If Black plays 1 ... Bxc5, then 2 dxc5 uncovers an attack on the bishop at d7, which Black cannot defend, e.g. if 2 ... Qc7 the manoeuvre Qe3-h6 is decisive.

  1 ... Qxa2 2 Rh5! e5

  If 2 ... gxh5, then 3 Qg3+ wins, while 2 ... Bg7 is answered by 3 Bxg7 Kxg7 4 Rxh7+ Kxh7 5 Qxf7+, followed by 6 Rd3. Also interesting is 2 ... h6 3 e5! Kh7 4 Qe3 Qxb2 5 Rd3 Qb1+ 6 Bf1 g5 (if 6 ... gxh5, then 7 Qe4+ Kg8 and 8 Rg3+ wins; otherwise Black is mated by 7 Rxh6+ Bxh6 8 Qxh6+ Kxh6 9 Rh3#) 7 Qe4+ Kg8 8 Rxg5+ hxg5 9 Rh3, etc. The move played parries the threat of 3 Qh3.

  3 dxe5 Be6 4 Qf4 Qxb2 5 Bf1 Be7

  He still cannot play 5 ... Bg7, because of 6 Bxg7 and 7 Rxh6+.

  6 Qh4

  Euwe notes in his commentary that Black’s queen is fairly well placed on b2 (by controlling f2 and f6 it prevents Qh6 being played at once) and cites as more precise 6 Rb1 Qd4 (if 6 ... Qxb1, then 7 Qh6) 7 Rb4! Qd8 8 Qh4, etc. 6 Rd3 would also win more quickly than the move played.

  6 ... Bc5! 7 Rh6

  The great defensive master’s last trump: 7 Rxh7? is now no good because of 7 ... Qxf2+. Even so, Black is lost; the process merely takes a little longer.

  7 ... a5

  If 7 ... Bg4, then 8 Rd3 with the threat of 9 Re3!.

  8 Rd3 Bxf2+

  Otherwise White will play 9 Rf3.

  9 Qxf2 Qxf2+ 10 Kxf2

  10 ... a4

  The situation now becomes critical; White must get his rook to h3 before Black’s pawn promotes to a queen.

  If 10 ... Ra6, there is insufficient time for 11 Be2 because of ... Rb6-b3; nor is 11 Rf3 Rb6 12 Rf4 strong enough owing to 12 ... g5! 13 Rf5 Rc8, and the black king escapes to d7. Strongest, after 10 ... Ra6, is 11 g4 Bxg4 12 Bh3!, etc.

  11 Be2

  White can also win by 11 g4 Bxg4 12 Bh3. After the text move Black cannot reply 11 ... a3 because of 12 g4 a2 13 Rxh7, so he tries to prolong the game by a desperate rook sacrifice.

  11 ... Ra5 12 g4 Rxe5 13 Bxe5 Bc4 14 Rdh3 Bxe2 15 Bf6 Re6 16 e5 Bxg4 17 Rxh7 Rxf6+ 18 exf6 1-0

  The rook has no difficulty in controlling the pawns, e.g. 18 ... Bxh3 19 Rxh3 b5 20 Rc3 b4 21 Rxc6 b3 22 Rb6, etc.

  Pilnik - Najdorf

  Mar del Plata, 1942

  Caro-Kann Defence

  1 e4 c6 2 d4 d5 3 Nc3 dxe4 4 Nxe4 Nf6

  Sounder is 4 ... Bf5 or 4 ... Nd7 followed by Ngf6.

  5 Nxf6+ exf6

  The doubling of the pawns on the f-file represents a slight weakness in Black’s position whichever pawn he recaptures with.

  6 c3 Bd6 7 Bd3 0-0 8 Qh5

  Black could answer this with 8 ... f5 (9 Bxf5? g6) and so ward off White’s attack without weakening his castled position. Indeed, 8 ... f5 not only avoids weakening Black’s position, it also falls in with his correct positional plan, i.e. the front f-pawn is used to control squares on the e-file, while the back one (at f7) stays at home and guards any weaknesses which the front one leaves behind it. If the front pawn is exchanged in some tussle over a square, Black has got rid of the doubled pawns and the fear of an inferior ending.

  8 Qc2 was sounder.

  8 ... g6

  Though less precise than 8 ... f5 this is still not a mistake.

  9 Qh4

  9 ... c5?

  Fundamentally inappropriate, since White is stronger on the queenside and in particular has a firm hold on d4. Black should have left any operations on the queen’s wing to White and instead prepared for ... f5 and ... Nd7-f6, but without exchanging queens and without relinquishing the d8-h4 diagonal to White. Correct, therefore, was 9 ... Nd7 10 Ne2 Be7! 11 0-0 f5 12 Qg3 Nf6, etc.

  10 Ne2 Nc6 11 Bh6 f5?

  Black should have played 11 ... Re8. At this point ... f5 is weak.

  12 Bg5!

  By exchanging queens White could have obtained the better endgame prospects. However, why should he play for an endgame when he has the opportunity to gain control of the diagonal h4-d8, thereby obtaining the necessary conditions for an attack on the castled king based on a network of weak dark squares?

  12 ... Qb6

  The diagonal cannot be defended: if 12 ... Be7, then 13 Bxe7 Nxe7 14 dxc5 Qxd3 15 Qxe7; while 12 ... f6 13 Bc4+ Kh8 14 Bh6 is equally unacceptable.

  JN: It is hard to see why this latter line is unacceptable. Black may continue with 14 ... Re8, threatening 15 ... Re4, and thanks to the move ... f6 Black’s kingside dark squares are not especially weak.

  13 0-0 cxd4 14 cxd4

  14 ... Re8

  Some commentators have marked this move out as the decisive mistake and have suggested 14 ... Be6 (preventing Bc4), but this would make the win easier, e.g. 15 Bf6 Rfe8 (or 15 ... h5 16 Qg5 Kh7 17 d5 Bxd5 18 Bxf5) 16 Nf4! Nxd4 (precluding 17 d5; if 16 ... Bxf4 17 Qxf4, Black will find himself unable to defend the focal-point g7, having given up his dark-squared bishop) 17 Nh5 Nc6 18 Bc3 Be5 (18 ... Be7 comes to the same) 19 Nf6+ Bxf6 20 Bxf6 h5 21 Qg5 Kh7 22 Be2 and White decides the game by Bxh5.

  In this analysis, as in other phases of this game, it should be noted that a method has been used which we have already seen in the variations arising from the diagram; namely, in creating threats against a dark-square network, White repeatedly makes use of sacrifices on squares of the other colour. These sacrifices fulfil the task of breaking up clusters of pawns which otherwise impede the attacker’s access to the weak squares on the network itself. The linkage between the operations on the networks of different colours should be noted; on one network the intention is to keep the attacking pieces in the strongest and most permanent positions possible, while on the other they are sacrificed when the situation demands.

  It should finally be pointed out that, instead of the move played, 14 ... Qxb2? would have been bad because of 15 Rfb1 Qa3 16 Rb3 Qa5 17 Bf6 Re8 18 Bc4, when White’s prospects for an attack are greatly improved.

  15 Bc4 h5

  There was a threat of 16 Bxf7+ Kxf7 17 Qxh7+ Ke6 18 d5+ Kxd5 19 Rad1+; If 15 ... Be6, then 16 d5 would naturally follow.

  16 Rae1?

  After this move Black gains drawing chances. In such an attractive position one would expect to find a safe and clear way to victory. White here has all the necessary ingredients for success: control over a dark-square network; advantage in development; two pieces to break up the cluster on the light squares (bishop at c4 and knight at e2); and as well as all this, a pawn in the centre that prevents Black from using his e5 square as a post for his pieces!

  Stronger and simpler would have been 16 Bf6! as the following lines show:

  1) 16 Bf6 Be7 17 Qg5 Bxf6 18 Qxg6+.

  2) 16 Bf6 Be6 17 d5 Bxd5 18 Bxd5 Rxe2 19 Qg5 Kf8 20 Rae1 Qb5 (or 20 ... Nd4 21 Qh6+ Ke8 22 Bc6+, etc) 21 Bf3 Rxe1 22 Rxe1 Ne5 23 Qh6
+ Ke8 24 Bxe5 and 25 Qh8+.

  JN: 20 Rae1 is refuted by 20 ... Qxf2+!! and Black forces mate himself. However, Vuković’s attacking plan is basically correct, except that White has to exercise a little more care: 20 Qh6+ Ke8 21 Qg7 Nd8 22 Rae1 and now 22 ... Qxf2+ 23 Rxf2 Rxe1+ 24 Rf1 Bc5+ fails, because after 25 Bd4 the bishop is defended.

  3) 16 Bf6 Re4 17 Qg5 Kf8 18 Nc3! and the black rook has no good square to go to: 18 ... Rxd4 is inadequate because of 19 Rfe1, while 18 ... Rg4 is answered by 19 Qh6+ Ke8 20 Bxf7+ Kxf7 21 Qg7+ Ke8 (or 21 ... Ke6 22 d5#) 22 Rae1+ and mate is unavoidable.

  16 ... Re4 17 Nf4 Qxd4 18 Rxe4 fxe4?

  This pawn blocks the fourth rank to White’s advantage. Black should have played 18 ... Qxe4, when 19 Nxg6 Qxh4 20 Nxh4 leads to an ending in which White is better placed, but Black can still salvage a draw by 20 ... Be6 21 Bd3 (or 21 Be2 Kg7! when 22 Bxh5 fails against 22 ... f6 and 23 ... Rh8) 21 ... f4, threatening both the a-pawn and 22 ... Kg7 followed by 23 ... f6.

  19 Nxh5!

  Breaking up Black’s light-square complex.

  19 ... gxh5 20 Bf6! Qc5

  Forced If 20 ... Qxc4 then White wins by 21 Qxh5, while if 20 ... Be5 then 21 Qg5+ is decisive.

  21 Rd1! Kf8

  The only reply. If 21 ... Be6 then 22 Rd5! Qxc4 (or 22 ... Bxd5 23 Qxh5) 23 Qg5+ Kf8 24 Rxd6 and mates, while 21 ... Be7 22 Rd5 Bxf6 loses to 23 Qg3+.

  22 b4!

  Obliging the knight to abandon its watch on d8. That White’s bishop on c4 is still ‘taboo’ can be seen from 22 ... Qxc4 23 Qxh5 Ke8 24 Rxd6 Qc1+ 25 Rd1, and Black loses his queen.

  22 ... Nxb4 23 Qg3! Bg4

  Clearly the only answer, for 23 ... Bxg3 cannot be played on account of 24 Rd8#.

  24 Rxd6 Nd3

  Or 24 ... Qxc4 25 Qf4 and mates.

  25 Bxd3 Qc1+

  Not 25 ... exd3 because of 26 Qf4.

  26 Bf1 Rc8 27 h3 Qxf1+ 28 Kh2! Qc1 29 hxg4 hxg4 30 Qxg4 Qh6+ 31 Kg3 Rc3+ 32 f3 1-0

  The first part of this tense game is instructive as an example of the gradual creation of the preconditions necessary for an attack on the castled king, while from the twelfth move onwards, by which time White’s control of the dark-square network is assured, the course of the game and its analysis provide many useful lessons on how to carry out an attack of this kind.

  6 The classic bishop sacrifice

  JN: More commonly called the Greek Gift sacrifice, although the reason for applying this term specifically to the sacrifice on h7 is unclear.

  This sacrifice deserves a chapter to itself, because it is the oldest and most explored of all the sacrifices involved in the attack on the castled king and also because it provides particularly good illustrations of the role of h7 and f7 as mating and secondary focal-points.

  The earliest instance of the sacrifice of the bishop on h7 followed by Ng5+ occurs in Gioacchino Greco’s handbook of 1619, but the first systematic review of it was made by E. Voellmy in his article ‘On attacking the castled king’ in the Schweizerischen Schachzeitung in 1911. Some writers give it the name of Greco’s sacrifice, for which there are some grounds, while others call it Colle’s sacrifice, for which there is in fact no justification, as Colle never played a good game which contained the sacrifice in spite of the fact that he often used it. I think the term ‘classic bishop sacrifice’ more appropriate, since this indicates both its early origins and also the basic simplicity of the idea underlying it.

  Greco reaches the position below in his handbook after making six not particularly intelligent moves.

  Now comes the blow: 1 Bxh7+! Kxh7 2 Ng5+ Kg8 (if 2 ... Bxg5, then 3 hxg5+ Kg6 4 Qh5+ Kf5 5 Qh3+ Kg6 6 Qh7#; 2 ... Kh6 3 Nxe6+ costs Black his queen; thirdly, after 2 ... Kg6 3 h5+ there is a choice between 3 ... Kh6 4 Nxe6+ and 3 ... Kf5 4 g4#) 3 Qh5 Re8 (or 3 ... Bxg5 4 hxg5 f5 5 g6) 4 Qh7+ Kf8 5 Qh8#

  Let us now consider the next, rather more natural, position, where the classic sacrifice is again successful.

  1 Bxh7+ Kxh7 2 Ng5+

  Here, as in other examples of the classic bishop sacrifice, there are three main variations arising from the sacrifice, i.e. those beginning with ... Kg8, ... Kh6, and ... Kg6. These three variations must be carefully examined on every occasion, for if any one of them contains a loophole, the whole combination falls to the ground.

  1) variation with ... Kg8

  2 ... Kg8 3 Qh5 Re8 4 Qxf7+ Kh8 5 Qh5+ Kg8 6 Qh7+ Kf8 7 Qh8+ Ke7 8 Qxg7#

  This variation contains, as its main feature, a typical example of an attack using h7 and f7 as secondary focal-points with the queen finally mating on g7, a common enough pattern after this sacrifice.

  2) variation with ... Kh6

  2 ... Kh6 3 Nxe6+ and 4 Nxd8, etc.

  This naturally succeeds because Black’s queen is on d8, but even with the queen on a5 the move 2 ... Kh6 would have no prospects. In that case White would win by, e.g. 4 Qg4 Rh8 (otherwise 5 Qh4+ Kg6 and 6 Qh7#) 5 Nxe6+ and Qxg7#.

  3) variation with ... Kg6

  2 ... Kg6 3 h4

  This is the strongest move here; in other cases ... Kg6 can be best met by 3 Qg4. Black’s reply is forced.

  3 ... Rh8 4 h5+! Rxh5 5 Qd3+ f5 6 exf6+ Kxf6 7 Qf3+ Ke7 8 Qf7+ Kd6 9 Qxh5 and White wins.

  One could now produce a series of examples, pursuing different variations according to the various elements in each position, but it is more to the point at once to pose the question as to how to judge the correctness of the sacrifice, first with reference to White’s position and then to Black’s.

  Necessary conditions for the classic bishop sacrifice

  White must firstly have a queen, a bishop and a knight. The light-squared bishop must be able to reach h7 in order to force the tempo of the attack, though it is not essential that it should put Black in check or take a pawn in so doing. The knight should be within easy and safe reach of the square g5, and the queen within reach of h5, though in some cases it is enough for it to be able to get to some other square on the h-file.

  As far as Black’s position is concerned, there should be two pawns standing intact at f7 and g7 (g7 may on rare occasions be occupied by a bishop instead of a pawn); the h-pawn should be on h7 (on h5 in exceptional cases), but it may be that there is no h-pawn at all. The position of Black’s queen on d8 and a rook on f8 points to, but does not absolutely guarantee, the correctness of the sacrifice. What is more important is that Black’s knight should not be able to reach f6 and that neither his queen nor bishop should be able to occupy the h7-b1 diagonal unharmed.

  These are the basic conditions which need to be taken into account. In the examples which follow we shall examine to what extent each of these conditions is fulfilled; the positions are chosen in such a way as to show which is the critical continuation in each example, i.e. the one that is in doubt because a certain condition is apparently unfulfilled.

  The variation with ... Kg8 is critical

  In the following position the sacrifice is correct:

  1 Bxh7+ Kxh7 2 Ng5+ Kg8

  The continuation 2 ... Kh6 obviously fails; while if 2 ... Kg6, then 3 Qd3+ Kf6 (3 ... f5 4 Qg3! wins, for Black cannot play 4 ... Kf6 because of 5 Rxe6#) 4 Qe4 threatens both 5 Nh7+ and 5 Nxe6, and 4 ... Ke7 does not work because of 5 Qxe6+ fxe6 6 Rxe6#.

  3 Qh5 Nf6

  An example of the defence of the focal-point h7 by the knight, but in this case White’s position is strong enough (thanks to the activity of his bishop on f4 and rook on e1) to withstand the loss of time.

  4 Qh4 Re8

  If 4 ... g6, then 5 Qh6!, followed by 6 Be5 and Re3-h3, wins, while 4 ... Bc7 would be met by 5 Be5, etc.

  5 Be5 Qe7

  If 5 ... Kf8, then 6 Qh8+ and Qxg7 is decisive.

  6 Re3 Bd8 7 Rh3 and White mates.

  In the position in the next diagram an important condition regarding the ... Kg8 variation has apparently not been fulfilled, since the diagonal from c8 to f5 is open to Black’s bishop. However, White has the possibility of Qe4+ as well as other active factors in his favour, and therefore the sacrifice still succeeds.

  1
Bxh7+! Kxh7 2 Ng5+ Kg8

  Both 2 ... Kh6 and 2 ... Kg6 are bad, the latter being met by 3 Qe4+ f5 4 Qh4 f4 5 Qh7+ Kxg5 6 h4+ Kg4 7 f3+ Kg3 8 Qg6+ Kxh4 9 Kf2, etc.

  3 Qe4!

  But not 3 Qh5? Bf5!.

  3 ... f5 4 Qh4 Rf7 5 Qh8+!

  JN: 5 e6 is even stronger, when 5 ... Qxg5 is forced to avoid mate.

  5 ... Kxh8 6 Nxf7+ Kg8 7 Nxd8 and White wins.

  Here the boxed-in position of Black’s rook and the possibility of Qe4+ followed by Qh4 fully compensate for the absence of one of the basic conditions.

  In this position Black has an extra tempo in which to defend himself after ... Kg8 because his rook is already on e8, but White’s position is so strong in other respects that his sacrifice still leads to a successful breakthrough.

  1 Bxh7+ Kxh7 2 Ng5+ Kg8 3 Qh5 Nf8

  A difficulty for Black is that he cannot defend the f7-square; if, for example, 3 ... Qe7, then 4 Qh7+ Kf8 and 5 Qh8#, while if 3 ... f6, then 4 Qf7+ Kh8 5 Re3, etc.

  4 Qxf7+ Kh8 5 Re3 and White wins.

  These three examples represent only a small section of a large number of possible cases in which the crucial continuation is ... Kg8. How large the variety of these positions is and how sensitive they are to minute changes can be illustrated by the following observation: if, in any of the previous three diagrams, White’s h-pawn is moved to h4, the bishop sacrifice is no longer correct. In the first two diagrams this is because the queen is prevented from getting to h4, while in the diagram above it is because White’s rook can no longer create threats along the h-file. Having the pawn on h4 rather than h2 is an obstacle in the variation with …Kg8; on the other hand, we shall see how it assists White in the variation with …Kg6.

  The variation with ... Kg6 is critical

  Here the critical continuation is ... Kg6, while that with ... Kg8 involves less difficulty.