Art of Attack in Chess

by Vladimir Vukovic

  5 ... Qf5 0-1

  For if 6 Qb3, then 6 ... Rac8 and 7 ... Qb1#.

  This position comes from Euwe-Colle, Amsterdam 1926. It is Black’s move and he hits at White’s castled position by means of a bishop sacrifice.

  1 ... Bxa3! 2 Bxf6 gxf6 3 Kb1

  If 3 bxa3, then 3 ... Ne5, while if 3 Qxd7+ Kf8 White is threatened by ... Qxc3+ and so has no time for Qxb7.

  3 ... Bxb2! 4 Na2

  If 4 Kxb2, then 4 ... Qb4+ 5 Kc2 Ne5 6 Nxe5 Be4.

  4 ... Ne5 5 Nxe5 Bd5 6 Qxd5 exd5 7 Nxd7 Ra8 0-1

  The following game serves as a demonstration both of how castling queenside forms part of a plan of development and of how an attack can be carried out against such a position.

  Lasker - Marshall

  St. Petersburg Final, 1914

  Petroff Defence

  1 e4 e5 2 Nf3 Nf6 3 Nxe5 d6 4 Nf3 Nxe4 5 Qe2

  Morphy’s old move; it is sound enough, though it gives White no more than a microscopic advantage.

  5 ... Qe7 6 d3 Nf6 7 Bg5 Be6?!

  Correct, of course, is 7 ... Qxe2+ 8 Bxe2 Be7 or 8 ... Nbd7. Avoiding the exchange of queens favours White, whose position is the better developed. One might say that Lasker knew which opponents it paid to play drawing variations against!

  8 Nc3 Nbd7 9 0-0-0

  Castling on the queenside is completely justified in this position, where the bishop is hemmed in at f1. Moreover, on the d-file the rook supports the d-pawn in moving to d4 and then threatening to advance to d5.

  9 ... h6 10 Bh4 g5 11 Bg3 Nh5 12 d4 Nxg3 13 hxg3 g4 14 Nh4!?

  A risky and adventurous move, typical of Lasker’s play in the middlegame. The knight clearly has no good way of getting away from the h4-square, a circumstance which could have helped Black to equalize. Better would have been 14 Nd2.

  14 ... d5?

  A decisive mistake. Black should have played 14 ... Nb6! and if 15 d5 then 15 ... Bd7. Against other attempts by White, the reply ... Qg5+ plays an important part. However, moves like 14 ... Nb6! were not to Marshall’s taste. What he really enjoyed was setting a trap for his opponent!

  15 Qb5! 0-0-0

  Castling is forced, since 15 ... Qb4 comes to grief after 16 Nxd5 Qxb5 17 Nxc7+. It was probably because of this trap that Marshall played 14 ... d5?, but Lasker naturally refuses to be diverted by the d-pawn and calmly plays for an attack on the castled position with 16 Qa5.

  JN: Vuković doesn’t actually say what the trap is – the point is that after 15 ... 0-0-0, 16 Nxd5 fails to 16 ... Bxd5 17 Qxd5 Qg5+ 18 Qxg5 hxg5 and the knight is lost.

  Perhaps Lasker even played 14 Nh4!? exactly because he knew Marshall would be tempted to set a trap and thereby get himself ensnared in the attack on the castled king which now follows.

  16 Qa5 a6

  Clearly, 16 ... Kb8 fails against 17 Nb5.

  17 Bxa6!

  A correct bishop sacrifice.

  17 ... bxa6 18 Qxa6+ Kb8 19 Nb5 Nb6

  Now that the black king has moved from c8 to b8, the distinctive character of the queenside castling position no longer exists and the situation is merely a replica of a type of attack used against the short castled position.

  An important advantage for the attacker in this position is his ability to bring a rook into action quickly (Rd3-b3), since without this Black could defend himself by ... Nc8 and then ... Bd7.

  20 Rd3 Qg5+

  This move simplifies White’s task somewhat. However, Black could not have saved himself even by 20 ... Nc4 which contains the transparent trap 21 Rb3? Qg5+ 22 Kb1 Nd2+. At the time Lasker had reckoned on defeating 20 ... Nc4 in the following manner (and this has found its way into the commentaries): 21 Re1? Rd6 (in order to meet 21 ... Qg5+ with 22 Ree3) 22 Nxd6 Qxd6 23 Rb3+ Nb6 24 a4 Bc8 25 Re8 Qd7 26 Qb5 Qxb5 27 Rxb5 Kb7? 28 a5 Bd7 29 Rxf8 Rxf8 30 Rb4 and White goes on to capture the knight, after which he is a pawn up in the ending. Let us disregard any doubts about this ending, for Black has a stronger continuation than 27 ... Kb7, namely 27 ... Bg7 28 Rxh8 Bxh8; then White must play 29 a5 (otherwise the knight remains alive on b6) whereupon the main line runs 29 ... Bxd4 30 axb6 c6 31 Rb3 Bxf2 32 b7 (if White allows ... Kb7, then the position is still more favourable for Black) 32 ... Be6 33 Kd2 c5! 34 Ke2 c4 35 Ra3 Bd4, and it is clearly White rather than Black who has to struggle for a draw. It is interesting to see how the original characteristic of the position (the inability of the knight on h4 to move) again emerges as the decisive factor as soon as the attack on the castled position peters out. Not even the best players are immune from such misfortunes!

  This game, which was the decisive one in Lasker’s victory at St. Petersburg, cannot be left to go on sleeping in the archives of chess history with an inaccurate commentary. Good games should not be allowed to go to their rest without good annotations as their lullabies! Therefore, we must show how in fact White does win after 20 ... Nc4.

  So that there will be no doubt about it, we shall supply two winning moves, both of them instructive as far as carrying out an attack in this kind of position is concerned.

  1) 21 a4! c6 (if 21 ... Rd6 22 Nxd6 Qxd6 23 Rb3+ Nb6, White wins easily by 24 a5) 22 Rb3! (not 22 Qxc6? because of 22 ... Qb7, forcing the exchange of queens and thus bringing the attack to a halt; nor does 22 Na7 work, e.g. 22 ... Qxa7 23 Rb3+ Nb6! 24 Rxb6+ Ka8, when White has to exchange queens, leaving Black once again with the better game; obviously the intermediate 23 ... Nb6! is the key element here) 22 ... Qg5+ 23 Kb1 Nd2+ 24 Ka2. Now Black is unable to take the rook on b3 with check, so White wins easily, e.g. 24 ... Qe7 25 Qb6+ Qb7 26 Qxd8+, etc.

  JN: Black’s best defence is 24 ... Rd7, when White must find 25 f4! Qf6 (25 ... gxf3 26 Nxf3 deflects the knight from the attack on b3) 26 Rc3! to keep an advantage.

  2) 21 Na7 Rd6 22 Rb3+ Nb6 (forced, since 22 ... Rb6 is defeated by 23 Nc6#) 23 a4 Bg7 (or 23 ... Qd7 24 Nb5! Rc6 25 Qa7+ Kc8 26 a5) 24 Nb5! and White wins.

  JN: Recent analysis by Kasparov reveals a third possibility: 21 Rb3 Qg5+ 22 Kb1 Nd2+ 23 Ka1 Nxb3+ 24 cxb3 Bd6 25 Qa7+ Kc8 26 Nxd6+ Rxd6 (26 ... cxd6 27 f4! gxf3 28 Nxf3 and the rook reaches c1) 27 Qa8+ Kd7 28 Qxh8 with two extra pawns.

  21 Kb1 Bd6 22 Rb3 Rhe8 23 a4 Bf5 24 Na7 Bd7 25 a5 Qd2 26 axb6 Re1+ 27 Ka2 c6 28 Nb5! cxb5 29 Qa7+ 1-0

  29 ... Kc8 30 Qa8+ Bb8 leads to the picturesque 31 Qa6#. In this ‘posthumous finale’ of the game Black’s king and rook are back once again in the original long castling position!

  10 Defending against the attack on the castled king

  In presenting, up to here, the various aspects of the attack on the castled king we have looked at the operation from the attacker’s point of view. However, it is also true that the defence against such an attack has its own principles and particular characteristics, and it is therefore fitting that we should also try to discuss this defence systematically.

  Chess literature tends for the most part to talk of active and passive defence as being the two main types of defence. Steinitz also introduced the term ‘self-defence’, and H. Kmoch (in his work The Art of Defence) speaks of ‘automatic defence’. Although it is now fairly well established, this division is not really the happiest one for teaching purposes; moreover, when it is a question of defence against an attack on the castled king, it is an imprecise one.

  I think it better to talk of direct and indirect defence of the castled position, in which case the direct defence is predominantly ‘passive’ and the indirect always ‘active’. The difference between these kinds of defence is usually clearly marked, since the former involves the defence of the actual castling area or the removal of one’s opponent’s aggressive pieces, while in the latter case it is a question of actions further away from the castled position, which the defender uses to keep the attacker so busy that he fails to pursue his attack to its conclusion.

  In actual play direct defensive moves usually occur in conjunction with those of an indirect kind, and the main rule is that direct defence should be economical and that indirect should be executed at the right moment. Economy in direct defence consists o
f not making superfluous moves and not employing an unnecessarily large number of pieces in defence. The right moment for indirect defence is fairly difficult to gauge, since it is normally determined by two interacting factors; on the one hand, when direct defence has reached the limits that necessary economy allows, and on the other, when the attacker has really made a move which commits him fairly strongly, i.e. he has staked his money on the success of the attack.

  First, we shall set out clearly the various elements which go to make up each type of defence.

  A. Direct Defence

  1) Protecting the squares of the castled position, including so-called overprotection.

  2) Transforming the castled position by moving the pawns in front of the king, a particular case being the setting-up of a blocked position.

  3) Altering the castled position by moving the king. Here it should be kept in mind that each move by the king changes the situation, changes the conditions necessary for an attack and, above all, changes the possible focal-points. King moves fall into the following graded categories:

  a) The consolidating move (e.g. Kb1 in the case of queenside castling and Kh1 in the case of kingside castling);

  b) A more radical alteration of the situation as a result of moving the king (e.g. to f2, thus producing a so-called ‘semi-castled’ position);

  c) A king’s flight. i.e. when it goes out in front of the castled position. Steinitz called this the king’s self-defence, an older and more correct term than that of ‘automatic defence’ applied by Kmoch.

  4) Defence by means of repulsing one’s opponent’s pieces, through their capture, exchange, or ejection from powerful positions.

  B. Indirect Defence

  Before considering the ways of using the indirect method we must first of all examine three aspects of such a defence. The first is purely spatial, i.e. whether the defender carries out his own counterattack in the centre or on the wings; the second is concerned with the degree of the threat; and the third with the question of whether the indirect action turns at some point into the direct method of defence. In the following review we shall combine these three aspects:

  1) The defender carries out a counterattack on the opposite side, where the attacker has castled. In such a case his action is usually itself an attack on the castled king, the position being one where the players have castled on opposite sides.

  2) An action on the opposite side first of all prepares the ground (files, ranks, etc), and then turns into an outflanking manoeuvre against the attacker’s castled position, i.e. while the attacker is attacking along the files, the defender gains control of the attacker’s first or second rank.

  In this case, as in ‘1’, the critical question is naturally either who will be the first to force mate or which of the two attacks will be brought to a halt through the diversion of pieces to provide a direct defence.

  3) An action on any part of the board with the aim of gaining material.

  Such an action obviously only has a point when a forced mate is not part of the attacker’s plan. The critical continuation then is that which costs the defender the least. For instance, if he can ‘pacify the attacker’s wrath’ at the price of a knight, then a counterattack which brings him a rook is worthwhile.

  In practice the most important case is where the attacker sacrifices; the defender’s repertoire is then greatly increased, since he is given various ways in which to return material in order to counter the attacker’s plan. This method of returning material is extremely important; the extent to which it is possible affects the type of sacrifice which can be made, and the sacrifice is also a signal to the defender to counter with a sacrifice which would previously have been out of the question.

  4) The counterattack results in material equality, but the question of position still remains open. Mate cannot be forced, both attacks are brought to a halt, and eventual simplification leads towards an ending. The resulting position, which is often extremely difficult to assess beforehand, is then decisive.

  5) The defender’s counterattack does not lead to a mating attack, nor to material or positional advantages, but instead consists of an eventual transition to the direct method.

  As a rule, this takes the form of either attacking the enemy pieces ‘from the rear’ or of gaining control of a file or diagonal so as to cover a certain square in one’s king position (e.g. when the white queen has captured the square c7 and from there covers the focal-point h2).

  6) A special case of the kind of counterattack mentioned in ‘3’ is provided by the advance of a queenside pawn majority. Its aim is to promote a pawn to a queen, and such an advantage justifies the most generous surrender of material in order to curb the attacker’s forces.

  7) Retaliation in the centre can have the features of sections ‘3’ to ‘5’ either separately or combined. The following possibilities are particularly worth noting here:

  a) A central thrust demolishes or weakens the attacker’s formation there and prepares the ground for further operations by the defender, or it enables him to transpose to direct defence ‘via a turning manoeuvre’;

  b) The destruction of the pawn centre eliminates its auxiliary function with regard to an attack on the castled king. For example, the disappearance of a pawn on e5 can also mean the end of an attack on the castled king, if the pawn is vital to a certain continuation;

  c) Because of the danger of a central thrust, many attackers at a suitable moment deliberately create a blocked position in the centre. As a rule such a central formation cannot be rapidly activated except by the sacrifice of a piece. If the attacker has already used a sacrifice in attack, the defender is presented with an opportunity of making a return sacrifice in the centre. For example, the defender’s knight on c6 is sacrificed for White’s pawns on e5 and d4, thereby creating the necessary conditions for various kinds of defence without loss of material;

  d) The destruction of the attacker’s centre usually gives the defender the prospect of better conditions for positional play, as mentioned in section ‘4’. For it is not only that the opponent’s centre pawns disappear, but the actual central squares also come under the defender’s control; consequently, if the attack is held, the considerable positional advantage enjoyed by the defender persists, even when the material is equally balanced, until well into the endgame.

  This outline of defensive methods brings out above all else the richness of their variety and the flexible way in which they can be combined. Because of this, defence as a whole is harder than attack, and its finesses are as a rule more concealed than those of the attacker, a point already noticed long ago by Lasker. The importance of the centre must be particularly emphasized as the critical area where the decisive step is often made. The great variety of possibilities afforded by a defence based on an action against the centre gives such an operation first place in the defender’s repertoire, and so every attacker must be able to calculate accurately what is likely to happen in the centre during an attack on the castled king.

  Defence by over-protection

  Passing now to some examples of direct defence, let us examine the following position, which can arise in the Closed Defence to the Ruy Lopez.

  It is Black’s move; he has to think how to supplement the direct defence of his king position, especially in view of White’s threat of Nh2 followed by 2 f4 and 3 f5. Black plays:

  1 ... Nd8 2 Nh2 f6

  Now 3 f4 can be met by 3 ... Nf7. Moreover, he can also answer most other moves by 3 ... Nf7, setting up a distinctive position with a ‘double knight fianchetto’.

  This position offers very strong resistance against all kinds of attack, and the reason for this lies in the over-protection of the squares f5, g5, and h5, at which points White might have been able to execute a breakthrough. These squares are now covered both by Black’s pawns and his knights, an example of ‘prophylactic over-protection’, to use Nimzowitsch’s phrase.

  The over-protection which Black achieves by transferring his knight
to f7 is in this case neither a passive nor an uneconomical means of defence. It is not difficult to establish that Black’s knight on b7 has no better position anywhere than on f7 (if ... c4, for example, White plays b4 and does not allow ... Nc5); it is better to use the knight for defence, freeing Black’s other pieces for later queenside activity. In some cases over-protection can mean economical direct defence; on the other hand, depending on the position, it may be superfluous.

  Defence by blockade

  A blocked pawn formation means a general reduction in the number of actions possible in any position, and as a result blockade is often the weapon employed by the weaker side. A means of defence which stands out as particularly important is the blockade of the pawns in front of the king.

  The continuation from the above position shows us first of all a typical case of defence by blockade, and then the way to overcome this method of defence. Although White has the superior position with all his pieces directed against Black’s kingside, he still can not achieve anything by ‘normal’ methods. If 1 g6, Black plays 1 ... h6, while if 1 h6, then 1 ... g6; in both cases blocking the pawns brings the attack to a halt. Attempts to break through by sacrificing a knight can simply be ignored by Black. Thus after 1 g6 h6 2 Nf3 Bd7 3 Ng5 Black does not take the knight but waits; while if 1 h6 g6 2 Nh5, Black again does not need to take, for by playing 2 ... Kh8 3 Nf6 Qf7 he can hold the draw.

  White’s only chance in this position is:

  1 Nhg6 Bd7

  Clearly, the knight cannot be taken. 1 ... Kf7 is met by 2 h6 and a breakthrough, since if 2 ... hxg6, then 3 h7 Be7 4 Qh4 followed by 5 Nxg6.

  2 Rgh1!

  White must play with great precision and thoroughly prepare for the breakthrough by h6 and hxg7. If he tries 2 h6 immediately, then Black replies 2 ... hxg6, when 3 hxg7 Bxg7 is not dangerous, while after 3 h7+ Kh8 4 Rh6 gxh6 5 gxh6 White’s attack will probably not succeed.