Art of Attack in Chess

by Vladimir Vukovic

  These variations do not exhaust the possibilities of the position, but they give a good illustration of the strength of White’s game after 16 dxc5.

  16 ... a6?

  An example of the inadequacy of passive defence. 17 Nb5, which is what Black was afraid of, is by no means White’s principal threat and is only effective in conjunction with a strong initiative. Alekhine is also at fault in his commentary when he claims that Nb5 represented a threat after 16 0-0-0.

  Instead of 16 ... a6? Black should have played 16 ... cxd4! and thus exploited the extra tempo in the centre, e.g.:

  1) 17 Nb5 Nc6 and now:

  1a) 18 f4 Rxf7 19 Nd6 and now Black has a satisfactory defence in 19 ... Qf8!.

  JN: Not so, as White wins by 20 Qg5+ Kd7 21 Nxf7 Qxf7 22 Rh7 and the g7-pawn falls.

  1b) 18 Nd6 Nxe5 (this gives back material and at the same time eliminates White’s threats) 19 Nxc8+ Rxc8 20 Qxe5 Qc7, and Black’s position is in no way inferior.

  1c) 18 Qg5+ Kd7 19 Qxg7 b6! 20 Nd6 Ba6 21 f4 d3 and Black obtains counterplay, e.g. if 22 c3 or 22 cxd3, then 22 ... Nd4, while Ne8 can always be answered by ... Ke7.

  2) 17 Rxd4 Nc6 18 Rf4 (sacrifices at this point are insufficient to break through) 18 ... Kd7 19 Rg4 Qe7 20 Rxg7 b6 and against any further activity by White on the kingside Black has enough counterplay based on ... Nd4 and eventually ... Rac8 and ... Qc5.

  17 dxc5!

  Now all is well again, and the position is won.

  17 ... Nd7 18 Rxd5! Qa5

  If 18 ... exd5 then 19 Nxd5+ Ke6 20 Nf4+ Ke7 21 e6! Nf6 22 Qe5 and mates.

  19 Qg5+ Kxf7 20 Rh7 Rg8 21 Rd4!

  There are a number of ‘subsidiary solutions’ here, but 21 Rxd7+? Bxd7 22 Ne4 does not work because of 22 ... Qe1#.

  21 ... Qxc5 22 Rxd7+

  But now, of course, this way is all right!

  22 ... Bxd7 23 Ne4 Qb4 24 Nd6+ Kf8

  25 Qf6+ gxf6 26 Rf7#

  In the past commentators used to end their work on games like this by praising the winner, but does not the loser in some way deserve credit too? Is it not he who ‘fulfils the ardent desires’ of the attacker, as is done in this case by the good natured matador from Ultima Thule...?

  Diagonals in the attack on the castled king

  Tartakower thought it suitable to bestow epithets of the following kind on the files: a-file - positional; b - covetous; c - far-sighted; d - prosaic; e - dangerous; f - tempting; g - sacrificial; and h - contentious. He might also have invented something similar (even if he in fact did not) to apply to the ranks. But the diagonals? Well, they are all somehow equally ‘elegant, subtle, far-sighted’, etc. On the long diagonals (a1-h8 and a8-h1) more kings have already died than in the whole of the world’s history; more tragedies have been played out on the classic diagonals (bl-h7 for White and b8-h2 for Black) than in all the world’s theatres, and on the developing diagonals (a2-g8 and a7-g1 respectively) there have been more mistakes made than in a large library full of books. Now that we have started on nicknames and epithets, let us bestow on the diagonals a4-e8 and h4-d8 (from White’s point of view; a5-e1 and h5-d1 from Black’s) the name of pinning diagonals. The reader can think over for himself how much material he has already lost on them.

  Just as the rook is the piece suited for play on open files and ranks, so the bishop is the master of the diagonals, indeed they are his raison d’être. At all stages of the game, and especially in an attack on the castled king, marshalling bishops on the diagonals demands great skill. A player’s skill, in fact, is often demonstrated by his play on the diagonals, and the impression is then given that ‘the slender bishop is subtler than the stout rook’. Putting facetiousness aside, let us now look at the relationship between the lines commanded by the rook and the bishop’s diagonals.

  In the first place, it should be stressed that the fundamental difference does not lie in the number or importance of the squares on the files or diagonals, nor in the weight of operations on them. The differences arise only from the different values of rook and bishop and also from the part played by pawn moves in opening up or blocking a line.

  It is harder to sacrifice the more valuable rook than the bishop, and the order of value is even more relevant when it comes to pinning: exchanging a bishop for a pinned rook wins the exchange, while bishop for knight is tit-for-tat, but a rook can only gain from pinning a piece when it takes the queen. The other difference is still more explicit: it is comparatively easy to open up or close a diagonal by the normal advance of one’s pawns, but achieving the same object on a file means that the pawn has to make a capture, and this naturally depends on the disposition of one’s opponent’s pieces; hence, there may be considerable difficulty in opening up a file. If, on the other hand, it has to be closed, an outpost as a rule has to be established; on the diagonals, however, outposts are the exception, since the ‘open-shut’ game is easily carried out by the pawns. It is in this, ultimately, that the whole ‘subtlety and far-sightedness’ of the diagonal lies. We shall open a series of examples with a case of pinning on a diagonal. The pinning here is in fact a temporary episode within the whole operation of ‘clearing up’ on the diagonal, and the end comes with the line being completely breached when the pinned piece is captured.

  This position comes from the game Tolush-Renter, Estonian Championship 1945. It is White to move.

  1 Nxg7

  This draws Black’s king out to g7 and at the same time opens up the diagonal of the white bishop on b2.

  1 ... Kxg7 2 Rd1 Qe7

  If 2 ... Qe6, then 3 f5.

  3 Qg4+ Kh8 4 Rd7

  In this way White wins the enemy bishop, but the correct method of exploiting the position would have been to aim consistently for a breakthrough on the square f6, e.g. 4 Qg5 Rc6 5 f5 Rg8 (or 5 ... Rd8 6 Re1 and if 6 ... Rdd6, then 7 Qh6 followed by Rf4 wins; otherwise the manoeuvre Rf4-g4 is at once decisive) 6 Qh6 Bc8 7 h4 b5 (Black has no better move; for example ... Rg7 and ... Ne8 will not save him because of f6) 8 h5 a5 9 Rfd2 b4 (the combination which White has prepared by advancing his h-pawn to h5 is also decisive against other moves) 10 Rd6 Rxd6 11 Rxd6 Qxd6 12 Bxf6+ Qxf6 13 Qxf6+ Rg7 14 h6 and White wins.

  JN: After 5 ... Rd8 6 Re1, Black might defend by 6 ... Rdd6 7 Qh6 Qd8, and if 8 Rf4 then 8 ... Rd1. In view of the uncertainties of this approach, Tolush’s method appears sounder.

  4 ... Qe6 5 Qg5 Rc6 6 Rxb7 and White won in due course.

  An important point is that, in addition to ‘clearing’ one’s opponent’s pieces from a diagonal, cases also occur where one’s own pieces prevent it being opened. It is extremely difficult to resuscitate a diagonal on which one has a blocked pawn of one’s own. Structurally, it is a similar problem to that of opening a file where a pawn blocks the path of a rook, and it is usually solved in the same way, i.e. by sacrificing a piece on a square controlled by the pawn. Here is an example.

  This position occurred in the game

  Alexander-Szabo. Hilversum 1947 . White pitched his knight into f6 with check so as to bring the long diagonal back to life.

  1 Nf6+! gxf6

  If 1 ... Kh8, then 2 Qh4 is very strong.

  JN: 2 Qh4 allows the defence 2 ... Qxd3; Instead, 2 Qe4 is correct, when 2 ... gxf6 3 exf6 followed by Ng5 forces mate.

  2 Qg3+ Kh8 3 exf6 Bxf6 4 Ne5! Bxe5

  White threatens 5 Rxf6, 5 Nxf7+, and indeed 5 Qh4 (5 ... Bxh4 6 Nxf7+ Kg8 and 7 Nh6#). Black no longer has any defence.

  5 Bxe5+ f6 6 Rxf6 1-0

  Since if 6 ... Rxf6, 7 Rf1 is decisive. The ‘clearing’ of the long diagonal was a victory for the bishop.

  As far as play on the diagonals is concerned, the strength of the central squares on the given diagonals is particularly important. These squares can assist the attacker in setting up outposts, in transferring pieces, and in other operations. Thus one should, as a rule, aim to have such squares protected or even over-protected. This is one of the maxims of positional play, and we shall see in the next example how Black pays for ignoring this maxim and how the attack on
the castled king suddenly erupts.

  This position arose in Stoltz-Guimard, Groningen 1946. White played:

  1 dxc5 Nxc5?

  A mistake, since it surrenders to White control over the important central squares d4 and e5, which means that the long diagonal occupied by his bishop on b2 is already as good as opened up (considering the gain of tempo by Nb5, which White always has in hand). Black should have played either 1 ... bxc5 or 1 ... Bxc5.

  2 Nb5 Qd8

  2 ... Qb8 is not good here because of White’s grip on e5, e.g. 3 Be5 Qa8 4 Nc7, and Black loses the exchange.

  3 Nxa7

  Black will, it is true, win back the a-pawn, but White will have succeeded, without loss of time, in strengthening the position of his knight at b5 (... a6 is no longer possible).

  3 ... Ra8 4 Nb5 Rxa2 5 Qb1 Ra8 6 Ng5

  Suddenly the course of the game turns against Black’s castled position; White threatens 7 Bxf6 followed by mate on h7.

  6 ... h6?

  Black should have consented to a further weakening of the diagonal by 6 ... g6, for that would have meant less trouble for him than the text move.

  JN: 6 ... Nce4 is perhaps the most sensible. Black blocks the b1-h7 diagonal without creating a further weakness on the kingside.

  7 Bxf6 hxg5 8 Be5

  By threatening 9 Bc7, White installs his bishop with gain of tempo at e5, a square that is now firmly under his control. His next intention is to play Qb2, after which the bishop’s whole manoeuvre takes on the purpose of clearing the diagonal.

  8 ... Qc8 9 Qb2 dxc4

  If 9 ... f6, which the tournament book (by Euwe and Kmoch) considers correct, White wins a pawn: 10 Bd6 Re8 (or 10 ... Qd7 11 Bc7 Ra6 12 Bf3, etc) 11 Bxe7 Rxe7 12 Nd6 Qc6 13 Nxb7 Rxb7 14 cxd5 exd5 15 Rxd5 and Black is in difficulties on a fresh diagonal (a2-g8).

  10 Bxc4

  Not 10 Bxg7? because of 10 ... Qc6.

  10 ... Qc6 11 f3

  11 ... Ba6

  The tournament book gives this a question mark and suggests 11 ... f6; however, after 12 Bd6 Rfe8 13 b4 Na4 14 Qb3 Kh8 15 Bxe7 Black loses his e-pawn.

  12 Nd4

  Making use of the other square Black surrendered with 1 ... Nxc5?. White has now created a battery consisting of queen and knight.

  12 ... Qb7 13 Bxa6 Rxa6 14 Bxg7!

  The various operations on the diagonal are concluded by a bishop sacrifice, the aim of which is to lure the king into the line of fire. Black did not submit to this and played 14 ... Rfa8, losing after a further sixteen moves. However, we are only interested in the continuation 14 ... Kxg7, which is followed by 15 Nxe6+ Kg6 16 Qg7+ Kf5 17 Nd4#. The plan also works with 15 Nf5+, in which case mate is achieved by 17 g4#.

  The next example is devoted to the problem of recognizing strong diagonals and combining their ‘clearing’ with an effective use of the bishop.

  Here we have a position from the game Lundin-Tartakower, Groningen 1946. It is clear that White is the better placed, and that Black is weak on the a2-g8 diagonal, but it is less obvious that the white bishop will put in such a quick appearance on that diagonal and that the game will be decided on the b1-h7 diagonal. White won as follows:

  1 e4! fxe4

  Otherwise 2 e5 is too strong.

  2 Ng5 Rxf1+ 3 Bxf1!

  The bishop is on its way to c4, the knight on g5 is already casting glances in the direction of f7, and the threats on the e-file see to it that the initiative remains in White’s hands.

  3 ... Qg6 4 d6

  This opens the diagonal with gain of tempo (the threat is 5 Qd5+ and then Qxb7).

  4 ... Rf8 5 Bc4+ Kh8 6 Qd5 h6

  If 6 ... Bc6, then 7 Nf7+ Kg8 8 Qxc6! bxc6 9 Ne5+ Kh8 10 Nxg6+ hxg6 11 Rxe4 g5 12 Bxa6, etc.

  7 Nf7+ Kh7 8 Ne5 Qe8 9 Qxe4+ Bf5 10 Qxb7 Nb8 11 Bf7 Qd8 12 Qf3 Qf6 13 g4 Bb1 14 Qd1!

  Now Black’s bishop does not have a single square left on the b1-h7 diagonal. Black lost the exchange and the game a few moves later.

  Here is an example showing the strength of two bishops working together.

  White offers a pawn in order to open up the long diagonal for his b2-bishop.

  1 d5! Nd7

  Black does not accept the sacrifice. After 1 ... exd5 2 Rxe8 Qxe8 3 Ng5 both 3 ... g6 4 Bxg6! fxg6 5 Qd4 and 3 ... h6 4 Bh7+ Kh8 5 Qh5 give White a decisive attack.

  2 dxe6 Rxe6 3 Rxe6 fxe6 4 Qc2 Bxf3

  4 ... g6 is met by 5 Qc3, while if 4 ... h6, then 5 Bh7+ Kh8 6 Nh4.

  5 Bxh7+ Kh8 6 gxf3

  Having gained a pawn and weakened Black’s position, White won the game in ten more moves.

  In this position, which arose in Geller-Kotov, USSR Championship (Moscow) 1955, White sacrificed a pawn to obtain control of the long diagonal and thereby gain the possibility to break through with a winning attack.

  16 b3! Nb4 17 Bb1 Nxe4?

  Black accepts the pawn sacrifice, a decision which not only costs him a series of tempi but also causes him to yield control of the long diagonal to White.

  18 Bb2 Bb7 19 d5 c4 20 bxc4 bxc4 21 Bxe4 Rxe4 22 Ng5 Re7 23 Qh5 h6 24 Nf5! Rxe1+ 25 Rxe1 hxg5

  25 ... g6 fails against 26 Qxh6! Bxh6 27 Nxh6+ Kf8 28 Nh7#. If 25 ... c3, White wins by 26 Re7! cxb2 27 Nxh6+.

  JN: After 26 Re7, Black has the surprising defence 26 ... Bxd5!, when 27 Rxc7 cxb2 28 Qd1 Bxa2 promotes the pawn. Therefore, 26 Ne7+ Qxe7 27 Rxe7 Bxd5 28 Nxf7! is better, when White should win.

  26 Re3 Bc8

  There was a threat of 27 Rh3 f6 28 Qg6, etc. If 26 ... f6, then 27 Qg6! (not 27 Bxf6 because of 27 ... Qf7!) 27 ... g4 28 Nh6+ Kh8 29 Nxg4, etc. If 26 ... g4, then 27 Nh6+ gxh6 28 Qxg4+ with mate in three moves.

  27 Bxg7! Bxg7 28 Re8+ 1-0

  For 28 ... Bf8 is answered by 29 Rxf8+ and 30 Qh8#.

  8 Pieces and pawns in the attack on the castled king

  Our examination of the attack on the castled king has so far (in chapters 4-7) been mainly directed towards the spatial aspect of the attack, and it is now time to examine the parts played by particular pieces and the significance of material sacrifices. Some of these points have already been touched on, so in this chapter we shall concern ourselves only with certain themes where the material aspect is paramount and which have not yet been fully dealt with.

  The queen in the attack on the castled king

  The queen is undoubtedly the most important piece in the attack on the castled king, and indeed without it such an attack rarely comes into consideration. Admittedly, it is not impossible to undertake a direct attack without the queen, but the necessary conditions for it cannot easily be created.

  The queen’s great mobility, its main characteristic, is a useful factor in the attack on the castled king, but its effectiveness at short range is even more striking. If it is in occupation of the focal-point g7, the enemy king on g8 has every square taken from it; moreover, if it is protected on that focal-point and the focal-point itself has been cleared of the influence of the opposing pieces, then the king is mated. The queen is, as it were, ‘made for focal-points’ , and particularly for play on compound focal-points. Therefore, the queen’s chief strategy in an attack on the castled king consists in playing on the focal-points and discovering ways in which to attack as effectively as possible at close range. Thus the queen, which in an endgame on an open board feels such an aversion towards the opponent’s king that it will not allow it nearer than two paces away, is filled in attack with some kind of dark desire and dreams only of how it can steal into the castled dwelling and there fold the king in a close and deadly embrace.

  The reverse side of the queen’s positive characteristics is its great material value; this means that the queen is not lightly sacrificed and is not as a rule given up for smaller gains, such as preventing castling or clearing a focal-point. However, there are a number of different kinds of queen sacrifice which occur in the attack on the castled king; these can be reduced in the main to two fairly sharply differentiated types: the case where the sacrifice forces mate, ‘check upon check’, as it were, and the case where it produces a great positiona
l weakness on the opposing side.

  There are also unusual and curious queen sacrifices; we shall give three examples as variations on this theme, ranging over a period of 40 years.

  In this position, from the game Fox-Bauer, Washington 1901, White, to move, set his opponent a trap.

  1 Ndc4 dxc4?

  Black in his greed accepted the sacrifice, which he should have refused.

  2 Qxg6!

  A correct sacrifice of the queen. which forces the opening of either the h-file or the a2-g8 diagonal.

  2 ... hxg6

  Or 2 ... fxg6 3 Bxc4+ Kf8 4 Nxg6+ and 5 Rh8#.

  3 Nxg6 fxg6 4 Bxc4+ Kf8 5 Rh8#

  A position from the game Lewitzky-Marshall, Breslau 1912. Marshall won renown in the following way:

  1 ... Rh6 2 Qg5 Rxh3 3 Rc5

  White saw that 3 gxh3 would not do because of 3 ... Nf3+ and intended, on 3 ... Qa3, to play 4 Rc7, etc.

  JN: Although even this fails after 4 ... Ne2+ 5 Kh1 Rxh2+ 6 Kxh2 Qd6+, etc.

  However, Black took him by surprise with:

  3 ... Qg3!! 0-1

  4 hxg3 Ne2# and 4 fxg3 Ne2+ 5 Kh1 Rxf1# fail at once, while after 4 Qxg3 Ne2+ 5 Kh1 Nxg3+ 6 Kg1 Nxf1 7 gxh3 Nd2 White is left a knight down.

  This position is from a blindfold game of Alekhine’s at Tenerife in 1945 (Alekhine-Supico) and acts as a counterpart to Marshall’s combination above.

  Alekhine, in playing

  1 Qg6!

  forced Black to resign (1-0). If Black does not take the queen but plays instead 1 ... Rg8, White wins by 2 Qxh7+ Kxh7 and 3 Rh3#; if Black plays 1 ... hxg6 the rook mates immediately, while if 1 ... fxg6, then 2 Nxg6+ hxg6 and again the rook mates on h3.