JN: White has no trouble forcing a draw at the end of this line, for example by 23 Rc1 or 23 b3, but he cannot gain the advantage.
Correct would have been at once 19 Rad1 with the following variations:
1) 19 ... Qb6 is simply met by 20 exf5.
2) 19 ... Nd3 20 f4 Qd4+ (if 20 ... Rh8, then 21 e5+ wins) 21 Kh1 Nf2+ (or 21 ... fxe4 22 Ncxe4+ Ke7 23 Qxg6) 22 Rxf2 Qxf2 23 e5+ Ke7 24 Qxg7+ Ke8 25 Qc7 and mates.
3) 19 ... Qe8 20 f4 fxe4 21 Ncxe4+ Nxe4 22 Nxe4+ Ke7 (or 22 ... Kf5 23 Ng3+ Kf6 24 Nh5+, etc) 23 Nd6 and White wins.
JN: After 22 ... Kf5, 23 Qh3+ Kxe4 24 Qd3# is quicker.
4) 19 ... Qxd1 20 Rxd1 Rh8 21 e5+ Kxg5 22 h4+ and White wins.
19 ... exf5 20 Rad1 Nd3!
Against other moves 21 Nd5+ would be decisive; but White is now really threatened with ... Rh8 and has to withdraw his queen, which means that the strongest weapon in his attack has been beaten back.
21 Qh3 Ndf4?
Black throws away his chance. With 21 ... Ngf4 (22 Qh4 Rh8) he could have given White a lot of trouble and perhaps even won the game.
22 Qg3 Qc7 23 Rfe1
23 ... Ne2+?
This loses a piece, but Black would probably lose whatever he played. Thus if 23 ... Be6, then 24 Rxe6+ Nxe6 25 Nd5#, while if 23 ... Bd7, White simplifies the position by 24 Nd5+! Nxd5 25 Nh7+ Kf7 26 Qxc7 Nxc7 27 Rxd7+ Kg8 28 Nxf8 Nxf8 29 Rxc7.
JN: The strongest defence is 23 ... Rd8, which prevents 24 Nb5 owing to 24 ... Ne2+. In this case White would have a draw at best.
24 Rxe2 Qxg3 25 Nh7+
The famous intermediate move!
25 ... Kf7 26 hxg3 Rh8 27 Ng5+ Kf6 28 f4 1-0
This game, here given a proper critical commentary for the first time, provides yet another blow against the legend of ‘the infallible Cuban’.
Another interesting point is that Capablanca often undertook attacks against the castled king in his earlier years, but extremely rarely as he became older. This was to a certain extent the result of the progressively greater ability of his opponents and a general rise in standards, but it was perhaps even more because of a gradual waning of enthusiasm under the weight of increasing self-criticism. Unlike Capablanca, Alekhine played in his own style from his youth right till the end, accepting the risk of difficult attacks against the castled king even against strong opponents.
Colle - O’Hanlon
Nice, 1930
Colle System
1 d4 d5 2 Nf3 Nf6 3 e3 c5 4 c3 e6 5 Bd3 Bd6 6 Nbd2 Nbd7 7 0-0 0-0 8 Re1 Re8
This does little harm, but 8 ... Qb6 is better, according to theory.
9 e4 dxe4 10 Nxe4 Nxe4 11 Bxe4 cxd4 12 Bxh7+?
A great lover of the classic sacrifice, Colle here too is unable to control his temperamental penchant for sacrifice, though even at a quick glance it is apparent that the presence of Black’s knight on d7 is a danger in both the main variations, while a valuable tempo is saved in the ... Kg8 variation by the rook’s being on e8. Correct would have been either 12 Nxd4 or 12 cxd4.
JN: As we shall see, Vuković is far too pessimistic about this sacrifice.
12 ... Kxh7 13 Ng5+ Kg6
So far all is well, since in the event of 13 ... Kg8 the sacrifice is half correct, that is, White can obtain a draw. After 14 Qh5 Ne5 15 Rxe5 Bxe5 16 Qxf7+ White can get perpetual check, but no more than that. Black is no better off if he tries a different line in this variation, e.g. 14 ... Qf6 (instead of 14 ... Ne5) 15 Qh7+ Kf8 16 Ne4 Qe5 17 cxd4, whereupon 17 ... Qxh2+ is forced, for 17 ... Qd5 is not feasible because of 18 Qh8+ Ke7 19 Bg5+ f6 20 Qxg7+ with a massacre of Black’s pawns.
JN: First of all, Vuković’s line 13 ... Kg8 14 Qh5 Ne5 15 Rxe5 Bxe5 16 Qxf7+ is not a draw, but a forced win for White after 16 ... Kh8 17 Qh5+ Kg8 18 b3, and the arrival of the bishop on a3 completes the mating net. His other suggestion, 14 ... Qf6 15 Qh7+ Kf8 16 Ne4 Qe5 is in fact Black’s best defence, but even here White need not take the draw, but can very well play for a win by 17 f4! Qd5 18 c4 Qc6 19 f5!. After 19 ... Nf6 20 Nxf6 gxf6 21 Bg5!, for example, White has an extremely dangerous attack. Thus Colle’s sacrifice gives White a very promising attack even against perfect defence, so Vuković’s acerbic comments regarding Colle’s play seem rather unjustified.
14 ... Nf6? is very weak indeed on account of 15 Qxf7+ Kh8 16 Re4.
14 h4
Obviously, this is the only continuation which can present Black with any difficulties; 14 Qg4 holds out no prospects here because of the reply 14 ... f5 15 Qh4 Nf6.
14 ... Rh8?
The decisive mistake. Black was probably only reckoning on 15 Qd3+ f5 16 Rxe6+ Nf6, and did not see that White also had 15 Rxe6+!.
Correct was 14 ... f5! 15 h5+ (or 15 Nxe6 Qxh4 16 Nf4+ Bxf4 17 Rxe8 Qh2+ 18 Kf1 Qh1+ 19 Ke2 Qxd1+ 20 Kxd1 Be5! and Black stands better) 15 ... Kf6 16 Qxd4+ Be5 17 Rxe5 Nxe5 18 Nh7+ Kf7 19 Ng5+ Kg8 20 Qxe5 Qd1+ when Black is the exchange ahead and should win.
JN: 17 Qh4! is a huge improvement which leaves Black defenceless, e.g. 17 ... Qa5 (17 ... Ke7 18 Qb4+ Kf6 19 f4 Bc7 20 Qd4+ e5 21 Qc4 mates) 18 f4 Bc7 19 Nxe6+ Kf7 20 Ng5+ Kf8 21 Rxe8+ Kxe8 22 Qe1+ Kf8 23 Qe6 Qc5+ 24 Be3 Qe7 25 h6 and wins.
15 Rxe6+! Nf6
If 15 ... fxe6, then 16 Qd3+ wins, e.g. 16 ... Kh5 17 g4+ followed by a quick mate, or 16 ... Kf6 17 Qf3+ Bf4 18 Qxf4+ with a slightly delayed, though still relentless, pursuit ending in mate.
16 h5+ Kh6
Or 16 ... Rxh5 17 Qd3+ Kh6 and 18 Nxf7#.
17 Rxd6 Qa5 18 Nxf7+ Kh7 19 Ng5+ Kg8 20 Qb3+ 1-0
We have used this game to illustrate a case where the classic bishop sacrifice was incorrect. The tournament jury, however, awarded it the First Brilliancy Prize. This was immediately after Alekhine’s reign of terror at San Remo and before the Carnival.
7 Ranks, files and diagonals in the attack on the castled king
An action in the vicinity of the castled king can be carried out by any of the pieces, but the long-range ones, the queen, rook, and bishop, are the most effective. They can attack from a distance along the ranks, files, and diagonals. The line concerned has to be ‘captured,’ i.e. one’s own long-range piece has to be placed on it, confrontation by the opponent’s corresponding piece has to be overcome, and the line has to be ‘cleared’ of all enemy influence. Outposts lie on such lines; batteries and pins take place along them. In a mating attack the line may finally witness the execution, i.e. checkmate. In this chapter we shall consider in turn the part played in the attack on the castled king by the ranks, files, and diagonals and reveal some typical operations involving them.
The weakness of the back rank
Of operations on the ranks, the first class to be considered is that aimed at a weakness on the back rank. A few practical examples will help.
This position is from
Bernstein-Capablanca, Exhibition Game, Moscow 1914 . It is White’s move, and he ought to have realized that because of his weakness on the back rank Black’s pawn on c3 was ‘taboo’. He should therefore have played 1 Nd4, but he decided not to, and the result was a disaster on his back rank.
1 Nxc3? Nxc3 2 Rxc3 Rxc3 3 Rxc3 Qb2! 0-1
But not 3 ... Qb1+? 4 Qf1, in which case 4 ... Rd1 is impossible because of Black’s weakness on his own back rank. However, the move 3 ... Qb2, typical of Capablanca, solves the problem completely: White resigned, for he is unable to take the queen and his own rook and queen are threatened; if 4 Qd3 or 4 Rd3, then 4 ... Qa1+ wins.
This position is from Mikenas-Maróczy, Folkestone 1933. White, to move, advanced his e-pawn to exploit Black’s weakness on the back rank.
1 e6 Nb6 2 e7 Re8 3 Rd8! Rxd8 4 e8Q+ and mate next move.
In this position, which occurred in the game Najdorf-Julio Bolbochan, Mar del Plata 1948, White exploited Black’s back-row weakness by an operation in which his passed pawn played a typically prominent part.
1 Rxd8! Rxd8
If 1 ... Qxc6, then 2 Rxf8+ Kxf8 3 d7 and White wins.
2 Qb6 Rf8
This loses at once, but Black could not have lasted long in
any case, e.g. 2 ... Qc8 3 d7 Qa8 (or 3 ... Rxd7 4 Qe6+) 4 Qe6+ Kh8 5 Re1 h6 (or 5 ... g6 6 Qf6+ Kg8 7 Qxd8+ Qxd8 8 Re8+) 6 Qg6 Rg8 7 h3 (preparing an escape hole in order to threaten Re8) 7 ... Qd8 8 Qf7 Kh7 9 Re8 and White wins.
3 d7 Rf1+ 4 Qg1! 1-0
Since after 4 ... Rxg1+ 5 Kxg1 Black cannot halt the d-pawn.
This position is from the game
Alekhine-E. Cohn, Stockholm 1912 . White, to move, must try to retain his extra pawn. A ‘realistic’ tournament player would quietly play 1 Rfe1 c5 2 Ndf3 and if 2 ... Bxc2, then 3 Rc1, or if 2 ... f6, then 3 exf6 Rxe1+ 4 Nxe1, etc. However, this seemed too limp a solution to the young Alekhine, who wanted to keep Black’s knight in its comer at h6 a little longer; in a typically ‘incisive’ way he defended the pawn on e5 indirectly, or rather ‘sold it for the back rank’, and this is in fact as good a method as that given above.
1 c4 Rxe5 2 c5!
A little more spice; if 2 ... Rxc5, then 3 Nde6 wins a rook.
2 ... Ra6?
He could have put up more resistance by 2 ... Rf6, for 3 Nf5 does not achieve much because of 3 ... Kf8. Correct would be 3 Ngf3!, against which Black must play 3 ... Re8, but not 3 ... Rxc5? on account of 4 Nf5!, when White wins the exchange at the very least.
JN: An interesting moment. After 2 ... Rf6 3 Ngf3 Re8, as Vuković gives, it is hard to see what White has gained in return for surrendering his extra pawn. Of course, White retains a positional advantage, but this is far from conclusive. Curiously, in the English edition of his best games collection, Alekhine does not mention 2 ... Rf6, but in his Deux Cents Parties d’Echecs he gives “On 2 ... Rf6 White wins by 3 Nf5!, etc.” Probably both 3 Ngf3 and 3 Nf5 are wrong; the strongest line is 3 Nc6! Re8 (3 ... Rxf1+ 4 Kxf1 Re8 5 Rd7 is also very bad) 4 Ne7+ Kf8 5 Rfe1!, threatening 6 Nxg6+, and Black’s position is hopeless (5 ... Bf5 6 g4 or 5 ... Bc2 6 Rd2).
3 Nde6?
This only leads to a small material advantage, whereas he could have won convincingly by 3 Nb5!, which is equally effective in stopping Black’s rook from returning to the back rank (with the help of Nxc7) but at the same time does not leave the knight en prise at e6. This point is demonstrated by the following variations:
1) 3 ... Kh8 4 Rd8+ Ng8 5 Nxf7+ Bxf7 6 Rxf7, and Black loses his knight while White’s on b5 remains intact.
2) 3 ... f6 4 Nxc7 Rxa2 5 Rd8+ Be8 6 Nxe8 Kf8 7 Nxc7+ Ke7 8 Re8#.
JN: Vuković is absolutely correct about 3 Nb5!, a move which goes unnoticed by Alekhine.
3 ... Kh8
The only reply.
4 Rd8+ Ng8
5 Nxc7
White cannot play 5 Rxf7 because of 5 ... Raxe6; while if 5 Nxf7+ Bxf7 6 Rxf7 Raxe6 7 Rff8 White merely regains the piece he has lost – that is the difference between the knight’s position on b5 and e6. Even more instructive is the fact that 5 Rfd1? will not work because of 5 ... Raxe6! 6 Nxe6 fxe6, and there is no mate (by Rxg8+ and Rd8+) since Black has freed the square f7.
5 ... Rxa2 6 Rfd1
Now that Black already has a rook at a2, 6 Nxf7+ Bxf7 7 Rxf7 is answered by 7 ... Ree2, while if 6 Rxf7, then 6 ... Rxg2+ 7 Kf1 Rgxg5 8 Rff8 Rgf5+.
6 ... f6 7 Rxg8+ Kxg8 8 Rd8+ Be8 9 Nxe8 Kf8 10 Nd6+ Ke7 11 Re8+ Kd7 12 Rxe5 fxe5 13 Nc4 Kc6 14 Ne4
One might now say ‘and White won after another twenty moves’, for this position no longer has anything to do with the difficulties on the back rank, but it is worth remarking that the task is a difficult one, so difficult indeed that Alekhine went astray and after 14 ... Ra1+ 15 Kf2 Kd5 16 Kf3 a5 played 17 Ke2? (17 Ncd2 was correct). This was answered by 17 ... a4! (not 17 ... Kxe4 on account of 18 c6!), after which White, objectively speaking, should not have got more than a draw. Alekhine, however, did not submit to 18 Nc3+ Kxc5 19 Nxa4+ Kd4 but went on to risk 18 Ncd2? a3 19 b4 Rc1! and so reached a losing position. Play continued 20 Kd3 a2 21 Nb3 Rd1+ 22 Kc2 a1Q 23 Nxa1 Rxa1 24 Nc3+, at which stage Black could have won by 24 ... Kc4! 25 c6 Ra7! 26 b5 Rc7. Instead, he spoilt his position by playing 24 ... Kc6?. The finish ran: 25 Kd3 Rf1 26 g3! h5 27 Kc4 h4 28 b5+ Kd7 29 gxh4 Rf4+ 30 Kd5 Rxh4 31 c6+ Kc7 32 Kc5! Rxh3 33 b6+ Kb8 34 Nb5 1-0
Combinations based on the weakness of the back rank are usually of short duration and matters come quickly to a head. This has been shown by the examples given above, and others, equally short-lived, could easily be added to them. However, just as success may come quickly, so too failure can suddenly occur as a result of a lightning counter-stroke which removes the weakness from the opponent’s back rank. As an example of such a turning-point let us examine the next position.
Flohr, when he was White in this position from a simultaneous display played in Leningrad during 1933, continued:
1 Rxe5
and left his opponent to think it over.
1 ... Qf2+
Both Flohr and the commentators thought that Black, in making this move, had fallen into a trap, but in fact he has nothing better. For example, if 1 ... Be6, then 2 Qe1 Rae8 (or 2 ... Bf5 3 Re7) 3 Bb5 Bd7 4 h6! (not 4 Rf3 because of 4 ... Qxh5) 4 ... Bxb5 (if 4 ... g6, then White wins with 5 Rf3) 5 Rxg7+ Qxg7 6 hxg7 Rf1+ 7 Qxf1 Bxf1 8 Rxe8+ and White wins.
JN: In this line 4 ... Qf6! favours Black after 5 Rxg7+ Kh8 or 5 hxg7 Rxe5, so White should prefer 3 h6 g6 4 Rge3 d4 (4 ... Bd7 5 Rf3) 5 Rf3 Qd7 6 Rxf8+ Kxf8 7 Bb5 and wins.
2 Kh2 Qxf1?!
This is a comparative mistake only, since Black’s game is now lost in any case. 2 ... Qf4 would have offered more resistance.
3 Qxd5+ Kh8
If 3 ... Qf7, White wins by 4 Re7!.
4 Rf3?
This was intended to be the decisive stroke, but in fact it is a mistake. White would easily have won by 4 h6, and if 4 ... g6 then 5 Rf3; alternatively, if 4 ... gxh6 then 5 Re8, after which mate is unavoidable. This variation shows the correct way to exploit the weakness on the back rank, while 4 Rf3? was inadequate as Black had the reply 4 ... Be6!, when by returning the extra piece he could have got rid of the weakness on his back rank with tempo and reached a drawn position. As it was, however, he played 4 ... Rxf3?, and everything was immediately put right: White replied 5 Qxf3, and Black with good reason resigned.
This example shows what vigilant care needs to be taken with combinations aimed at a weakness on the back rank.
In connection with this back-row weakness, something should also be said about the so-called ‘escape hole’, i.e. the gap made by moving a pawn with the sole purpose of eliminating the danger of mate on the back rank or at least of freeing the rook guarding the rank. On the other band, such a move, like any pawn move in front of the castled king, represents a weakening of the position, and it is a question here of correctly assessing whether the advantages involved outweigh the weaknesses or vice versa. This is one of the numerous double-edged features of chess.
As an example of the sensitivity of the connection between the weakness of the back rank and that of the ‘escape hole’ the above position is instructive.
It is Black’s move, and if it were not for his weakness on the back rank, the problem would be an extremely straightforward one: White’s pawns on the queenside would fall one by one. Thus it is a question of first averting the danger on the back rank and only then making the time to ‘pick off’ the pawns. Let us create an ‘escape hole’ for the king.
1 ... h6?
A mistake? Yes, and a bad one, for Black is now faced with a new weakness on g6. The other ‘hole’, 1 ... g6, is no good either, since White can get an easy draw by 2 Ng4 (with the threat of Nf6 and mate on h7) 2 ... Rf8 (or 2 ... Bf8 3 h6) 3 h6 Rf5 (the only move; (for if 3 ... b5, 4 Rd5 is a strong reply, while 3 ... Kg8? is not good on account of 4 Rg7+ Kh8 and 5 Ne5) 4 Rd8+ Rf8 (4 ... Bf8? 5 Ne3 and 6 Nxc4) 5 Rd7 followed by a repetition of moves.
JN: White would even stand to win after 4 Ne3!, winning the c4-pawn for nothing.
Also weak is 1 ... Bf8? 2 Ng4 Rxa2 3 h6! gxh6 4 Nf6 Bg7 5 Rd8+ Bf8 6 Rd7!, again with a draw by repetition of moves.
2 Ng4 Rxa2 3 Ne5 Kg8 4 Rd8+ Bf8
Or 4 ... Kh7 5 Ng6, mating.
5 Nxc4 b5
Or 5 ... Kf7 6 Rb8.
6 Ne5 Rxb2 7 Ng6 Rxc2+ 8 Kf3 Rc3+ 9 Kg4 Rc4+ 10 Kh3 Kf7 11 Rxf8+ and Black must lose.
As we have seen, the ‘escape holes’ did not work; in this position they represent too great a weakness, so the only correct continuation is 1 ... Kg8!, which removes the danger on the back rank without weakening the pawn position. What is more, Black must also avoid moving his pawns if White continues with his sharpest reply, 2 h6! (since if 2 ... gxh6 White can draw by 3 Ng4). He should instead play the solid 2 ... Bf8, after which he can reap his pawn harvest with ... Rxa2, etc. Another typical variation is 2 Ng4 (instead of 2 h6) 2 ... Rxa2 3 Ne5 Rxb2 4 Rd8+ Bf8 5 Nd7 Rxc2+ 6 Kf3 Kf7. Now one can appreciate the significance of White’s lack of an outpost for his knight – he wins the bishop by 7 Nxf8 but Black, with 7 ... b3 wins the game.
JN: This is all rather confusing. After what Vuković considers to be Black’s best continuation, 1 ... Kg8! 2 Ng4 Rxa2, White can play 3 h6 forcing 3 ... Bf8 (or else perpetual check), but after 1 ... Bf8 2 Ng4 Rxa2, which Vuković condemns on account of 3 h6, Black can reach exactly the same position by playing 3 ... Kg8!. The truth of the matter is that 1 ... Kg8 and 1 ... Bf8 work equally well, because neither weakens the kingside pawn structure.
The sensitivity of ‘escape holes’ of this kind can best be seen if White’s pawn on h5 is moved to h4. Then, as well as 1 ... Kg8, Black may also play 1 ... h6, a fact of which the reader may convince himself by his own analysis.
Art of Attack in Chess Page 12