14 Ng3 b4 15 axb4 Nxb4 16 g5
Black has made a start to his preparations for the thrust ... d5, but White’s attack is faster. White does not play 16 e5 here, as Black can then reply 16 ... Nfd5.
16 ... Ng8
17 f4
A pawn attack on the kingside is a valuable weapon for White in the Scheveningen Variation, while Black’s best response is a counter-stroke in the centre by ... d5. Black is somewhat behind in preparing for this thrust, and consequently White is able to set in motion his pawn attack. 17 Be3 was, by the way, also solid and good.
17 ... Bc6?
A weak move, which considerably helps White in regrouping his forces, because 18 Nce2! immediately threatens c3. Correct was 17 ... Rfd8 preparing ... d5.
18 Nce2! Bb5 19 Bd2 d5
19 ... Nxc2? would not have worked because of 20 Rc1 Rac8 21 Bc3! Ne3 22 Qd4.
20 Bc3!
Black has in fact managed to play ... d5, but White has made good use of the time and now, besides his bayonet (the pawn on g5), he also has a ‘long sabre’ in the shape of his dark-squared bishop on the long diagonal!
20 ... dxe4 21 Nh5 f6 22 Nxg7!
Grandmaster Tolush, himself an officer, uses his weapons well. After the ‘bayonet’ and the ‘sabre’, here comes a suicide ‘bomb’. The sacrifice is correct and characteristic of an attack based on a pawn avalanche.
22 ... Bxe2
The correctness of the sacrifice has its basis in the powerful threat of Nd4, as can be seen in the following variation: 22 ... Kxg7 23 Nd4! Bc4 (or 23 ... Qd7 24 Bxe4 and if 24 ... Bc6, in order to prevent 25 Qh5, then 25 Nxc6 Qxd1 26 Rfxd1 Nxc6 27 Bxc6 Ra7 28 Rd7, when White picks up another pawn and makes his task easy) 24 Bxe4 Rfd8 25 Qh5 Kf8 26 Qxh7 with a decisive attack.
Also interesting is 22 ... Rad8 23 Nd4 Rxd4, when 24 Qxd4 is not very effective because of 24 ... Kxg7, and White does no more than win back the sacrificed material. However, if 23 ... Rxd4, White has the fine reply 24 Qg4!, which prevents 24 ... Kxg7 and creates the threat of 25 Nxe6 followed by g6!. No answer to this has been found for Black.
Aware, therefore, of the force of Nd4 Black first of all takes the knight on e2 and then accepts the sacrifice.
23 Qxe2 Kxg7 24 Bxe4 Nd5 25 Qh5 Rfd8
If 25 ... Rf7, then 26 g6 hxg6 27 Qxg6+ Kf8 28 Rg1 and White wins.
26 Rg1 Bc5 27 gxf6+ Kf8
28 Rxg8+! Kxg8 29 Bxh7+ Kf8
If 29 ... Qxh7 then 30 f7+! is decisive.
30 Qg6 Nxf6 31 Qxf6+ Qf7 32 Qh6+ Ke7 33 Qg5+ Kd7 34 Qxc5 Qxh7 35 Rd1+ Ke8 36 Qc6+ Kf8 37 Bb4+ Kf7 38 Qc7+ Kg6 39 f5+! Kh6 40 Qf4+ Kg7 41 Qg5+ 1-0
For 41 ... Kf7 is answered by 42 Qe7+ and mate in three more moves.
This game won the First Brilliancy Prize at the tournament.
Concluding remarks
Before we conclude this discussion of the role of the pawns in the attack on the castled king, we must warn the reader of the immense diversity of practical cases which resist all attempts to embrace them within a set of rules. Our acquaintance with certain roles should not lead us to handle the game in a routine fashion; the particular characteristics of every position must always be thoroughly examined from one move to the next.
In connection with pawn attacks, there are two cases worth mentioning which often arise in practice. In one case the position requires a rigid, positional use of the pawns in attack, and in the other the position is not at all suitable for the cooperation of the pawns and consequently the attack is better carried out by the pieces alone. For these two cases some suitable examples will now be given.
This position is from the game Steinitz-Marco, Nuremberg 1896. White is the more aggressively placed but Black’s position is fairly solid, especially in the centre, where White lacks a pawn base. The diagonal a2-g8 may become dangerous for Black, but it is not so yet, for if Bc4+, Black can play ... Be6. Nor are the conditions favourable for an attack with the ‘bayonet’ or ‘avalanche’ here; therefore, another way has to be found to strengthen White’s position, one which is essentially of a purely positional nature. Steinitz finds it: he advances his f-pawn to f5, where it restricts Black; the fact that this leads to a blocked position is a good thing, since Black’s f-pawn can then be fixed at f6 and subsequently ‘rolled up’ by g4 and g5. This is a clear positional plan, which leads to the destruction of Black’s pawn chain and thence to a real attack on his castled king. There is no bayonet or avalanche here, though pawn moves are still the decisive means of attack.
1 f4! Bd7
Black wants to get his bishop to c6, since it cannot settle on the diagonal a2-g8 because of 2 f5. If 1 ... Bg4, then 2 Rd2 followed by h3 is strong, while 1 ... g6 would be a voluntary weakening of the castled position.
2 f5 Bc6 3 Bc4+ Kh8 4 g4 b6
Black has no satisfactory answer to g5. If 4 ... h6 5 h4 Nh7, we get 6 g5 fxg5 7 hxg5 and then either 7 ... Nxg5 8 Qxg5 and 9 Rh1 or 7 ... hxg5 8 Qh3, threatening 9 Qxh7+.
5 g5 fxg5 6 Qxg5 h6
Otherwise 7 f6 follows.
7 Qh5 bxc5 8 Nf7+ Kh7 9 Ng5+ Kh8 10 f6 gxf6 11 Qxh6+ Nh7 12 Nxh7 Rxh7 13 Qxf6+ Rg7 14 Rxe5 1-0
As our second concluding remark we cited the case where the player has a choice between attacking with his pawns and attacking with his pieces. Every practical player will have experienced the difficulty of such a choice, and an immense amount of time has been consumed in master chess on such decisions. The trouble is that it is difficult to set up any general rule, since everything depends on the particular characteristics of the situation concerned. Before discussing this question any further, let us take an example, and then, by varying the pawn structure, try to penetrate to the heart of the problem.
It will be easy for the reader to see that this position has arisen as the direct result of a classic bishop sacrifice. It is clear that White must aim to get a rook on to the h-file, and the only question is how to do it. One way would be f4 and then Rf3-h3, and the other Re1-e3-h3. The first is faulty and allows Black to draw, while the second is correct and leads to a win for White. The position is such that only the pieces can cope with the task; the pawns should stay at home to avoid getting in the way. The following variations prove this assertion.
1) 1 f4? Qd7
Of course, Black must on no account play 1 ... Re7? since 2 Rf3 would result in the loss of his rook.
2 Qh5+ Kg8 3 Rf3
White has nothing better. After 3 g4 g6 4 Qh3 Nh7 5 Nf3 Nc6 6 Nh4 Ne7 White is still unable to play f5 as planned, while Black obtains counterplay by ... c5, etc. Also good for Black is 3 f5 exf5 4 Rxf5 Qxf5 5 Qxe8 Nc6.
3 ... g6 4 Qh4
Or 4 Qh6 Qg7.
4 ... Nh7
This forces a draw. It is of no importance to our thesis whether Black can now play for a win, e.g. by 4 ... Qg7 followed by ... Nc6 and ... Ne7 or ... Nd8, etc.
5 Rh3
White cannot avoid the exchange of knights; if 5 Nh3, Black gets time to bring out his queen’s knight and set up a complete defence.
5 ... Nxg5 6 Qh8+ Kf7 7 Rh7+
JN: Here Vuković overlooks that White can mate in two by 7 Qf6+. However, his thesis is basically correct, because the stronger move 4 ... Qg7 casts doubt on White’s attack.
7 ... Nxh7 8 Qxh7+ Kf8
White now has to keep up perpetual check by 9 Qh8+ Kf7! 10 Qh7+, but if the unfortunate pawn on f4 were not there, he would be able to mate by Bh6+, etc.
2) 1 Re1!
Not 1 Rd1 because of 1 ... Ba6! 2 Re1 Bd3 followed by ... Bh7, etc.
1 ... Qd7 2 Qh5+ Kg8 3 Re3 g6 4 Qh4 Qg7
If here 4 ... Nh7, then 5 Rh3 Nxg5 6 Qh8+ would win. In this line 5 ... Re7 is met by 6 Qh6 Rg7 7 Nxh7 Rxh7 8 Qxg6+ Rg7 9 Qh5 Rh7 10 Bh6 Nc6 11 Qg6+ Kh8 12 Bg5 and 13 Bf6+.
5 Rh3 Nbd7
Against anything else White’s bishop decides the game by reaching f6, e.g. 5 ... Nc6 6 Nf3 Nh7 7 Bh6 Qd7 8 Bg5, etc.
6 Bd2
Opening the way for the other rook, for which the f-file, for instance, has attractions.
6 ... c5 7 Re1 cxd4 8 cxd4
a5
8 ... Rac8 does not work at once because of 9 Rf3 Re7 10 Bb4.
9 b4
A simple device to hinder ... Rac8; 9 ... a4 is answered by 10 b5, and the white bishop gains control of b4. Black now tries to put up a stand with ... Re7 and ... Nh7.
9 ... Re7 10 Qg4 Bc6 11 Qe2
Black is tied up on the kingside, and White’s best plan consists of trebling on the h-file. White carries out the necessary regrouping in such a way that not only is Black’s bishop prevented from penetrating but also his rook’s invasion along the c-file is deferred; at the same time, he over-protects the e5-square as a precaution against any desperate sacrifices.
11 ... Bb7 12 Rh4 Ba6 13 Qe3 Bb5 14 Rc1 a4 15 Rc3 Ree8 16 Qe1! Rac8 17 Rch3 Rc2 18 f4
Only now, when his rooks are ready for the lethal blow on the h-file, can White touch his pawns.
18 ... Rec8 19 Rh6 Bd3 20 Qh4 Rc1+ 21 Kf2 Rf1+ 22 Kg3 and Black cannot prevent mate by Rh8+.
In comparing these two variations, the reader cannot but recognize the essential feature of the position that condemns 1 f4? to failure but brings victory to the attack carried out by the pieces. The move f4 obstructs both the diagonal for the bishop on c1 and the f-file for the rooks, and so limits the possibilities of action along the very two lines on which Black is weak. His weakness on the f-file can be seen at a glance, and it is not difficult to deduce (on examining the diagrammed position) that he will also be weak on the dark squares, since he will be unable to postpone ... g6 for long. After f4 White is left with only his domination of the h-file, whereas if he does not play it, he also has the f-file and a line of action for his bishop, and it is the combination of all these factors that gives him victory.
We shall now vary the position of the pawns in front of White’s king in the above diagram, so as to train our eyes to be keener in spotting the various finesses contained in this kind of position.
First let us suppose that White’s f-pawn is already on f4 (instead of f2) and that it is White to move.
Now of course, 1 Rf3 will win at once, but if this possibility is excluded, e.g. by White’s having his h-pawn on h3, what then? Even so, White wins – by 1 f5! exf5 (otherwise 2 f6 is decisive) 2 Rxf5, and Black has no answer to the two threats of 3 Rf4 and 3 Qh5+ Kg8 followed by either 4 Nf7 or 4 Rxf8+ Kxf8 5 Qh8+ Ke7 and the queen mates on 6 Qxg7#. If, for example, 2 ... Qd7, then 3 Qxf8+, while if 2 ... g6, then 3 Rf4 Qxg5 4 Qxe8.
The pawn on f4 wins in this case precisely because it disappears and in doing so immediately opens up the f-file with disastrous consequences for Black.
Let us now try a variation with the h-pawn at h4 (rather than h2).
This position is fairly promising; White has his f-file and the bishop’s diagonal is free as in variation ‘2’, but the h-file has to be opened up by advancing the h-pawn. All White has to do is induce ... g6, after which the task of line opening is made easier and a blocked position averted.
White’s best is 1 Re1 Qd7 2 Qh5+ Kg8 3 Re3 (forcing ... g6, since it threatens 4 Rf3 and Rxf8+; thus if 3 ... Nc6 4 Rf3 Nd8, then 5 Rxf8+ Kxf8 6 Qh8+ Ke7 7 Qxg7+ Nf7 8 Nxf7 Rg8 9 Qh7 and White wins) 3 ... g6 4 Qg4 Nh7 5 Nh3, when he has good prospects.
If White’s h-pawn is on h3, 1 h4 comes into consideration, with the idea of reaching a similar system to the one above, but with a tempo less.
With White’s g-pawn on g3 (rather than g2) the problem is more difficult, since the pawn is only in the way there. Perhaps White ought to give perpetual check at once. If 1 Re1, Black plays 1 ... Nc6, after which 2 Re3 does not work because of 2 ... Qxg5, while after 2 g4 Qd7 3 Qh5+ Kg8 4 Re3 g6 5 Qh4 Qg7 Black’s defences are already in quite good shape.
Things are different with the g-pawn at g4 (which is, admittedly, impossible after the classic bishop sacrifice, but that does not matter here).
In this case the third rank is clear for the rook, and in addition the pawn at g4 also lends force to the thrust f5. As well as the method starting with 1 Re1 (on the pattern of variation ‘2’), still more successful here is 1 f4 (threatening 2 Rf3) 1 ... Qd7 2 Qh5+ Kg8 3 f5 and White must win.
Here, finally, we have a case where an attack using the pawns is correct. One tiny change in the position and already the choice between attacking with pieces or with pawns has swung in favour of the pawns. Such is the sensitivity of the attack on the castled king, and it is now clear how difficult it is to create any more definite rule on this point! Still, I shall attempt to conclude the chapter with a series of tips – which will naturally apply only generally.
Ten Practical Tips
1) It is an essential condition for any attack on the castled king that the opponent should not be able to counterattack in the centre or on the other side of the board, or rather that his counterattack should not be dangerous and not develop fully before the attack on the castled king. An attack using pawns usually progresses more slowly than one with pieces, since it often contains moves without any immediate threat. So, before embarking on a pawn attack, one should beware particularly of the possibility of a counterattack by one’s opponent.
2) A central pawn blockade prevents or limits counterattacks and so makes a pawn attack on the castled king easier.
3) One should always examine whether a promising attack can be made by the pieces alone, and if it can, the pawns should be left ‘at home’.
4) As a rule, it is difficult for pawn assaults to succeed against unweakened castled positions. since the possibilities of blockade are very great. A preliminary action by the pieces is necessary to produce a weakening of the castled position or to remove the possibility of a later blockade when the pawns advance.
5) Every blockade in the castling area radically changes the character of the position, for it affects the activity of individual pieces and the whole nature of the play which follows.
6) One’s own pawns usually constitute a great obstacle to one’s rooks; if a position is characterized by the opponent’s weakness on the files, the rooks are important pieces and pawns have no business on these tiles, except when their advance actually helps to clear a file. Rooks can be particularly usefully deployed on the third rank, if it is clear of pawns (an example of this was provided by the variations arising from the diagram here.
7) A pawn storm generally increases the risk of an inferior endgame. If other elements in the position are also weighted against an endgame, the added risk entailed in an attack on the castled king is acceptable. However, if the player has the necessary conditions for a good endgame, he must make certain that his assault with the pawns will be successful, for otherwise it represents a real risk.
8) A player who already has an advanced pawn facing his opponent’s castled position but which somehow gets in his way (e.g. a pawn on f5) must contemplate advancing the adjacent pawn, so that the lone pawn can be made an active component of an avalanche. Every advanced pawn presents a pretext for an avalanche and draws its neighbour towards it. This is least so in the case of an advanced h-pawn and most so if the pawn is on f5.
9) Attacks on the castled king in general, and those using pawns in particular, should always be judged on the basis of an assessment of the position as a whole. The placing of the pieces, the situation in the centre and on the other side of the board, positional strengths and weaknesses, how far the position is blocked, the prospects for the endgame – all these come into the reckoning which an experienced player must make before deciding to use his pawns in an attack on the castled king.
10) Cautious before opening his attack, the player must be incisive, consistent, and merciless once he has set out on his attacking course. For an insufficiently justified attack is not helped by belated doubts, but only by resourcefulness in the struggle to provide one’s opponent with the most difficult task possible.
9 The attack on the fianchettoed and queenside castling positions
Up till now we have, on the whole, based our examination of the various aspects of attacking play on the ordinary kingside castling position. The fianchettoed and queen
side castling positions present certain special characteristics, which affect the particular forms of attack used against these types of position.
The attack on the fianchettoed king position
The term ‘fianchetto’ (from the Italian, meaning ‘little flank’) denotes a bishop being developed on the flank (on g2, b2, g7 or b7), and if a player castles behind such a fianchetto formation, we talk of a fianchettoed king position.
Castled positions of this kind have their own particular good and bad sides; these will be examined on the basis of Black’s kingside fianchetto position, which occurs in the Dragon Variation of the Sicilian, and the King’s Indian and Grünfeld defences.
The good aspect of the position is, above all, its considerable capacity to resist attack along the diagonal b1-h7, thanks to the pawn formation, f7, g6, and h7; nor are the squares f6 and h6 too weak, for they are covered by the bishop on g7. The bishop is also well concealed against attacks and possibilities of exchange, while in many cases the defence is assisted by its ability to cover not only the squares f6 and h6 but also the possible mating square h8. Finally, the fianchettoed position is also suited for those operations in which the g-pawn disappears, as in the following well-known manoeuvre: Black plays ... f5, and then, on White’s reply exf5 he coolly continues ... gxf5, with the aim of strengthening his position in the centre. Without the bishop on g7 such play would be risky because of the opening of the g-file, but it often succeeds with the fianchettoed bishop, since the bishop defends Black’s king on the g-file about as well as a pawn does.
The weak aspect of the fianchettoed position lies in the fact that the position of the g-pawn at g6 simplifies the attacker’s task in opening up the h-file by the advance of his h-pawn; there is also a danger that the fianchettoed bishop may be exchanged either for its opposite number or some other enemy unit. Without the bishop the castled position becomes weak and vulnerable and the squares previously covered by the bishop form a typically weak network.
Art of Attack in Chess Page 19
