Art of Attack in Chess

by Vladimir Vukovic

  The third phase of an attack is of course characterized by the execution, either in the form of mate, of obtaining a decisive material advantage, or of being able to call off the actual attack on the king and obtain a positional advantage or favourable ending.

  Types of commitment

  To the above-mentioned types of preconditions necessary for an attack there also correspond various types of commitment which the attacker can, and indeed must, enter on, if he is going to attack.

  The commitments which correspond to the first phase are, of course, minor ones, usually a question of tempi. A player, for instance, spends a tempo improving his prospects for an eventual attack, transferring a piece from one side of the board to the other or the like; admittedly, the degree of commitment involved is not great, but it can become significant if the attack on the king is never carried out. It is natural that contemporary masters, in an era of precise opening play, aim first of all at sound moves, which give nothing away to their opponents and constantly strengthen their general positional build-up; the result of play according to this recipe on both sides is that a drawn position is reached. It should also be said that Alekhine (JN: and Kasparov!) knew how to instil a little aggressiveness into his moves even in the opening and create some conditions for attack with a minimum of commitments. This is precisely where his skill lay, and it is this which is lacking in many masters.

  A somewhat greater degree of commitment is entailed in the second phase. It may be a question here of a sacrifice accompanied by considerable compensation, of the dislocation of a number of pieces, or, in particular, of the advance of the wing pawns, which as a rule represents a fairly far-reaching positional commitment. Still, in this phase too the sensible player aims to conduct his game with a minimum of commitments, unless he is undertaking an uncompromising mating attack that is certain to succeed. It is especially important that the attacker should avoid a pattern of thinking whereby an unclear attack is reckoned as part of his compensation. For instance, he may sacrifice a knight for two pawns and an ‘attack’; when that ‘attack’ peters out, his reckoning naturally falls to the ground.

  Assessment of preconditions and commitments

  One of the attacker’s most important tasks is assessing the extent to which the preconditions for an attack are present and the degree of commitment which may be incurred. The following positions provide examples of this task.

  This position is from the game

  Botvinnik-Chekhover, Moscow 1935 . It is White’s move, and he has to decide on how to pursue what is already a promising initiative. He has the upper hand in the centre, a strong knight on e5, and well placed bishops; all these factors together mean that he has some of the preconditions for an attack on the king, without yet having entered on any commitments. He can therefore be said to be on the threshold of the first phase and must consequently look for further moves which will prepare an attack, while committing himself as little as possible. Thus, it would clearly be out of the question to make a pawn advance with h3 and g4, since Black would then find it easy to parry the main threat, after which White would be left only with pawn weaknesses, to say nothing of the dangers which would arise on the diagonal a8-h1.

  A slightly lesser degree of commitment would be entailed by White’s manoeuvring his a1-rook to h3 via e1 and e3, but it can be shown that such a displacement of the rook would in fact be premature; Black could easily parry the threats involved and then himself prepare for the counter-thrust ... b5 (e.g. by ... Nf8 and ... Rac8) thereby opening up the c-file, over which the white rook on h3 no longer has any influence. Because of this Botvinnik decided on a third plan, which involved the least degree of commitment: this consisted of transferring the c3-knight via d1 and f2 to h3 and thence on to g5. Admittedly, there is a definite commitment here too – in the shape of the number of tempi used; however, Black cannot really take advantage of this, and there is the additional point that the threat posed by the position of the knight on g5 is quite genuine and cannot be easily parried. Now here is the course that the game took:

  1 Qe2! Nf8

  First White thoroughly eliminates the possibility of Black’s playing ... b5.

  2 Nd1 Ra7

  2 ... Ne8 would probably have been better.

  3 Nf2 Qb8 4 Nh3 h6

  Now there is a weakness in the black king position as well. White still presses on.

  5 Ng5! hxg5 6 fxg5

  This is of course already the second phase, since White is committed materially; at the same time, his attack breaks through. If now 6 ... N6h7 7 Nxf7 Nxg5 (or 7 ... Bxg5 8 Bxh7+ Nxh7 9 Qxe6) 8 Qh5 Ngh7, then 9 d5 exd5 10 Nh6+ Kh8 11 Qf7 Nf6 12 Qg8+ and 12 ... Nxg8 13 Nf7#.

  JN: In the 7 ... Bxg5 line, 9 Qxe6 allows 9 ... Bc8, when no win is apparent, so White should prefer 9 Nxd8 Qxd8 10 Qxe6+ Kh8 11 Rae1 with a large advantage.

  6 ... N8d7 7 Nxf7

  In fact, it would have been simpler and quicker to play 7 Nxd7 Rxd7 8 gxf6 Bxf6 9 Rxf6 gxf6 10 Qg4+ Kf8 11 Ba3+ Rd6 12 Qf4 Ke7 13 c5.

  JN: Recent analysis has shown that this further sacrifice is quite wrong, in that Black could later play 7 ... Kxf7 8 g6+ Kf8! with fair defensive chances. Therefore White should have chosen 7 Nxd7, as given by Vuković.

  7 ... Kxf7 8 g6+ Kg8 9 Qxe6+ Kh8 10 Qh3+ Kg8 11 Bf5 Nf8 12 Be6+ Nxe6 13 Qxe6+ Kh8 14 Qh3+ Kg8 15 Rxf6! Bxf6 16 Qh7+ Kf8 17 Re1 Be5 18 Qh8+ Ke7 19 Qxg7+ Kd6 20 Qxe5+ Kd7 21 Qf5+ Kc6 22 d5+ Kc5 23 Ba3+ Kxc4 24 Qe4+ Kc3 25 Bb4+ Kb2 26 Qb1#

  This position arose in the game

  Alekhine-Gilg, Semmering 1926 . It is White’s move, and he decided on an attack, although the preconditions were only partially fulfilled. Black’s king position is weakened and some of White’s pieces are well placed, but there is as yet no communication between his rooks. Meanwhile, it is of decisive importance that Black is able to launch a counterattack based on his control of d4 and the exposed position of the white king; the general instability in the centre is also to Black’s advantage. We shall see how Alekhine’s attack came to grief in this position.

  1 g4?

  Seeing that the preconditions for an attack have not been fulfilled, this additional commitment by White turns against him. The correct move, objectively speaking, was 1 Be3, which would have led to approximate equality but not to an attack on the king.

  1 ... Nf6 2 Nxf6+ Qxf6 3 f5 Nd4 4 g5 Nxf3+ 5 Qxf3 hxg5 6 h4 Re8! 7 Kd1

  Black threatened 7 ... Qxf5.

  7 ... gxh4 8 Kc2 Bd7

  It is clear that White’s attack has been repulsed and that he must now defend himself against an action by Black.

  9 Bd2 Ra6! 10 Qh5 Ba4+ 11 Kc1 Rb6 12 Ra2 Bd4 13 b4 Be3 14 Bxe3 Qc3+ 15 Bc2 Qxe3+ 16 Kb1 Bxc2+ 17 Rxc2 axb4 18 Qxh4 bxa3+ 19 Ka2 Qh6!

  Since the exchange of queens is now forced (if 20 Qe1, then 20 ... Qf6 would be decisive) the danger on the h-file disappears and Black’s superiority in the endgame is decisive. White lost the game in due course. The game is also, of course, a good example of indirect defence.

  This position is from the game Gligorić-Kostic, Zagreb 1947. White was afraid that, if he allowed ... cxd4 cxd4, he would not be able to maintain his backward pawn at d4, so he played 1 d5 exd5 2 Bxd5, intending to support the centralized bishop on d5, if necessary, by c4.

  However, this is not a good plan, since it leaves White with an inflexible position as well as the danger of an inferior endgame. On the other hand, the backward d-pawn is not a great drawback, as long as White strives to obtain compensation in an attack on the king. The preconditions are in fact there, for Black’s position has been weakened by ... h6, and White can rapidly post his queen and bishop on the b1-h7 diagonal and so force a further weakness in the shape of ... g6. It should also be noted that Black has no minor pieces on the kingside, while his knight and bishop on the queen’s wing are obviously not suitably placed for defence. In the centre Black can only exchange pawns on d4, while a pawn advance on the queenside is too slow. The preconditions for an attack on the castled king are therefore fulfilled to a considerable measure; White
can play for an attack – indeed he must, for otherwise he may well find himself in difficulty over the backward pawn.

  1 Qe2! Nc6

  If Black plays 1 ... Rd8 2 Qe4 Nd7 in order to have ... Nf8 in reserve, then the continuation 3 Rfd1 cxd4 4 cxd4 Nf8 5 Bb3 leaves him at a loss to complete his development, whereas White can either play on the c-file or manoeuvre with Nd2-c4-d6.

  2 Qe4 cxd4 3 cxd4 Rd8

  After 3 ... Na5 4 Bd3 g6, White first of all plays 5 Rac1, whereupon counterplay commencing with 5 ... Bd7 can be answered by 6 d5 exd5 7 Qxd5. If 3 ... Nb4 4 a3, Black loses a tempo, since his knight cannot go to d5.

  4 Bd3 g6 5 Rac1 Nb4 6 Bb1 Nd5 7 h4 h5

  Otherwise 8 h5 is strong.

  8 g4 hxg4 9 Qxg4, and White has a promising kingside attack.

  The creation of preconditions

  The preconditions for a successful attack on the castled king can arise in a number of different ways. Some arise out of the normal process of building up a position; some are created by the player’s own moves, which are specifically aimed at such an attack; some occur unnoticed, as it were, in the course of the game; and others are the result of risky or even faulty play by the opponent. Let us leave aside those cases where the preconditions are created ‘unnoticed’ or through faulty play and turn our attention to those whose origin is conscious and calculated. In this connection we are already acquainted with a rule which has few exceptions.

  Of all the different preconditions for an attack on the king, one should create first those which entail the lesser degree of commitment, i.e. those which are of use not only in the event of an attack on the king, but also go to strengthen one’s position in general.

  Good posts for one’s pieces, spatial advantage, strength in the centre, the security of one’s own king – all these are factors of use in an attack on the castled king but which do not commit one to it. Commitment occurs gradually as weaknesses in the castled position are induced and as a flexible central formation is transformed into a fixed one (e.g. when, with pawns at e4 and d4, one plays e5). In other words, weaknesses in a castled position can be full of point if it comes to an attack, but otherwise they may be without significance or even useful to the side which has incurred them, e.g. ... h6 may provide a flight square for the king; or ... g6 may prepare the advance ... f5. Equally, a premature advance in the centre can entail a general positional commitment and so profit the opponent in various ways. More gradual commitments also arise from a regrouping of minor pieces outside the castling area but with the sole aim of attacking the castled king. It is already more difficult when it comes to transferring a rook to the third rank, since this rook will need some time to return and reopen communication with the other one, should the attack not succeed. The engagement of minor pieces in the vicinity of the castling area (e.g. as a result of moving a knight to g5) also involves a heavier commitment in terms of tempi. However, the heaviest commitments are those incurred in an attack carried out with the aid of pawns. Of these the heaviest is usually incurred by advancing the g-pawn, next comes the advance of the h-pawn, while advancing the f-pawn entails the least commitment. Now we have a small graduated scale for the degree of commitment incurred by the various operations which are necessary to create the preconditions for an attack on the castled king; it need not be emphasized, perhaps, that this scale has only a general and approximate validity. Let us now examine the following example from practical play.

  This position is from the opening stages of the game

  Alekhine-Weenink, Prague 1931 . It is White to move.

  1 Bd3!

  The textbooks gave 1 Bc4; Alekhine’s move takes into account the fact that Black is about to castle kingside and threatens e4 and e5 followed by Bxh7 (whether Black castles or not). As a result, Black has no better move than ... h6, which means that a weakness in his castled position has already been brought about, and without any special expenditure in time by White, since Bd3 is a natural developing move in this position. Inducing weaknesses like this is naturally correct in any circumstances.

  1 ... h6 2 e4 Qd8 3 0-0 Nbd7 4 b4

  With his second move White occupied the centre, with his third he made his king secure, and now he safeguards his central position against the counter-thrust ... c5. Apart from this, b4 prepares the ground for piece regrouping on the queenside.

  4 ... 0-0 5 Qe2

  This contains the threat of 6 e5 followed by 7 Qe4 but, at the same time, does not yet commit White in the centre.

  It would be a mistake to play 5 e5 Nd5 6 Rb3 Re8 7 Qe2 Nf8, since Black would then have managed to centralize his knight at d5 and invalidate the attack based on Qe4. The move e5 must clearly be reserved for a later moment when everything else is ready.

  5 ... Re8 6 Rb3!

  Primarily, this frees the square b1 for the bishop so that the queen may be placed in front of it on the diagonal b1-h7. It also makes it possible to transfer the rook to g3 or h3 as part of the attack later on, but it does not otherwise commit White to an attack on the castled king, since the rook can also be used for play, for instance, on the c-file.

  6 ... Qc7 7 Bb1 Nh7

  Black is too weak to fight back in the centre by 7 ... e5. After 8 dxe5 Nxe5 9 Nxe5 Qxe5 10 Bb2 White obtains a strong attacking position. If 7 ... Nf8 (in order to meet 8 e5 with 8 ... Nd5) then 8 Ne5 followed by Rg3, etc.

  8 e5

  This central pawn thrust commits White more than any other move hitherto. Clearly, he is now no longer thinking of the endgame but is intent on forcing home an attack on the castled king. It is of decisive importance that Black does not have ... Nd5 available; if he plays, for example, 8 ... Nb6, White’s threats take effect even move quickly, e.g. 9 Qe4 Nf8 10 Qg4 Kh8 11 Qh5, etc. So Black must resort to defence and not play for ... Nd5.

  8 ... f5

  To rid himself of some of the pressure. It also sets a trap, i.e. 9 Qc4 Nhf8 10 Bxf5? Nb6 11 Bxe6+ Kh8, when White loses his bishop.

  9 exf6 Bxf6 10 Qe4 Nhf8 11 Qg4 Kh8 12 Qh5 Nh7

  Otherwise 13 Bxh6 wins.

  13 Re1 Rad8 14 g4! and now we have reached the position which was dealt with here. White’s heaviest positional commitment in the game, the move g4, occurs exactly at the moment when the attack on the castled king has come to a head.

  The reader should observe the time sequence of each operation in this game, particularly with reference to the degree of commitment. The mild commitment incurred by Rb3!, the greater one by e5, and the heaviest one by g4, correspond exactly to the phases of the attack and to the strength of the threats which White has at his disposal.

  After this example, in which we have seen the temporal relationship between the various operations designed to create the preconditions for an attack, let us turn to a separate examination of each of these operations.

  First, the inducement of weaknesses in the castled position. This is an important operation, although it usually forms part of another operation and tends to be limited to one or two moves. It can take the form either of the unprovoked or faulty weakening of the castled position or of the correct advance of a pawn or pawns in front of the king which, because of faulty play later on, results in an appreciable weakness (e.g. when a flight square for the king, originally advantageous, becomes a weakness); then there is the case of the weakness which is enforced early on without any loss of tempo, while on the other hand, there are weak points which arise in the sacrificial whirlpool of the attack’s final phase. All these cases show our theme in its wider sense. In the narrower sense, the inducement of weaknesses implies an operation on which the attacker spends one or more tempi with the sole aim of weakening his opponent’s castled position in a definite way. An operation of this kind which entails an expenditure of several tempi represents a very real commitment, and so it usually belongs to a stage when most of the conditions necessary for an attack have already been fulfilled.

  There have already been many examples of the inducement of weaknesses, so two will be sufficient here.

  This positio
n is from the game Lilienthal-Bondarevsky, USSR Absolute Championship, Leningrad/Moscow 1941. Black had come to grief through playing an incorrect version of the Marshall Attack in the Ruy Lopez; a pawn up, White does not have a hard task ahead.

  1 Bg5

  Provoking Black to weaken his kingside by ... f6, which is difficult to avoid.

  1 ... f6 2 Bd2

  White has devoted a tempo to inducing the weakness ... f6; we shall soon see that it is well worth it.

  2 ... Kh8 3 Re1 Qf8 4 d4 Bd6 5 Nxd6 Qxd6 6 Qh5 g6

  The threat of 7 Re8+ provokes a new weakness.

  7 Qh4 Bc6 8 f3 Kg7

  White threatened 9 Bxd5 Bxd5 10 Bf4 Qc6 11 Re7.

  9 Bh6+ Kg8 10 Re4 g5

  Black is lost anyway, with or without this; for example, he faced a threat of 11 Bf4 Qf8 12 Bxc7.

  11 Qh5 Qd7 12 Bxg5! Qf5

  The main variation of the combination runs 12 ... fxg5 13 Qxg5+ Kf8 14 Rf4+! Nxf4 15 Qg8+ Ke7 16 Qf7+ Kd6 17 Qxf4+ Ke7 18 Qf7+ Kd6 19 Qf6+ Qe6 20 Qxe6#.

  13 Qh6 Kh8

  White wins after both 13 ... Qxg5 14 Rg4 and 13 ... fxg5 14 Qxc6.

  1-0

  Black did not wait for the unanswerable 14 Bxd5.

  The continuation from this diagram shows the correct enforcing of a weakness on g6 followed up by consistent play against the focal-point g7; an alternative line illustrates how provocation can be exaggerated with the result that it turns against the attacker.