I was impressed by Black's play in this game. He unleashed a new move in the opening, seizing the initiative and making White the one to look for equality. A few moves later he offered a sharp exchange sacrifice. White unwisely accepted the offer, and his king was soon stranded in the middle of the board, perilously short of safe squares. The battle was decided long before White could get his extra material into action.
Khouseinov,R (2330) - Magomedov,M (2600) Dushanbe, 1999 [D36]
1.d4 d5 2.c4 e6 3.Nf3 Nf6 4.Nc3 Be7 5.cxd5
The Exchange Variation is recommended by many authors as a simple and safe way for White to handle the Queen's Gambit Declined. The reality is that it does not promise White a simple and safe advantage -- IF Black knows what he is doing. Many years of practice have shown that the lines with an early Nf3 are not the most incisive. To achieve an edge, White's priority moves are Bg5, Qc2, e3, and Bd3, taking control of an important diagonal and denying Black a safe equalising manoeuvre with ...Bf5 (sometimes assisted by ...g6). Depending on Black's specific reaction White can then try a setup with Nge2, or revert to lines with Nf3. But in this game White has already committed himself to the move Nf3.
5...exd5 6.Bg5 c6 7.Qc2 Na6!?
Interfering with White's intended setup in the most direct way. If White had played e3 instead of Nf3 he could simply capture this knight, at the same time spoiling Black's queenside pawn structure.
Some players have tried 8.e3 Nb4 9.Qd1 Bf5 10.Rc1 intending 11.a3, but Black can keep his queenside demonstration going with 10...Qa5!?
There are many positions in the Exchange Variation where this central break gives White a space advantage and pressure on Black's weak points. This is not one of those positions. White's play has been slower than usual, so the advance has fewer chances of success.
The game Portisch-Jussupow, Rotterdam 1989 saw the more restrained 9.e3 Bf5 10.Bd3 Bxd3 11.Qxd3 Nc7 and Black had few problems, the game eventually ending in a draw.
9...Nxe4! 10.Nxe4 (first diagram) 10...Nc7!!
A new move which had been suggested by Lubomir Ftacnik but not actually played until this game. Because of a later inaccuracy by Magomedov I suspect that he was not following home analysis and may even have been unaware of Ftacnik's suggestion.
The best reply. Black cannot play 11...Bxc5 because 12.Bxd8 wins a piece, so White manages to get his queen to an active position. Ftacnik's analysis ended after 11.h4 dxe4 12.Qxe4 Bf5, but after the further moves 13.Qe5 Kd7!, it is clear that the upcoming exchanges will be favourable for Black.
11...dxe4 12.Qe5 Rf8 13.Qxe4 Bf5 14.Qe3
White avoids 14.Qxe7+ Qxe7+ 15.Bxe7 Kxe7, when his isolated d-pawn is the central feature of a much-simplified position.
14...Nd5 15.Qd2 Be4!
White is being driven back on all fronts and this move adds to his difficulties. The upcoming exchange on f3 will seriously compromise White's defences, and for specific tactical reasons it cannot be easily prevented. For example, 16.Be2 Bxf3 17.Bxe7 Qxe7 18.gxf3 and Black has the advantage after 18...Qf6. White finds the only sensible reply to 15...Bd4 – a counterattack on Black's rook – but there is another surprise waiting for him.
16.Bh6 (second diagram) 16...Bxf3!
Far more energetic than 16...Rg8. Recapturing on f8 will clear the e-file for Black's remaining rook, adding to the pressure on White's king. It might be said that the exchange sacrifice is based less on exact calculation and more on what Kasparov famously described as “chess culture.”
White should not accept the offer, although he is still in difficulties after 17.gxf3 Rg8 18.Bd3 Qd6! 19.h4 Bf8 20.Be3 f5 (Magomedov).
A strange decision because the simple 17...Kxf8 18.gxf3 Bg5 19.Qc2 would transpose to the game. This might be described as Magomedov's only slip in an otherwise-impressive attacking display.
Much stronger was 18.Qd3. In his analysis Magomedov gave 18... Kxf8! 19.gxf3 Qa5+ 20.Kd1 Re8! 21.Kc2 without further comment, but there are some blank spots here. To start with, in the position after 21.Kc2 Black wins easily with 21...Nb4+ 22.axb4 Qxa1. More critically, what happens if White preserves his pawn structure with 19.Qxf3? The best line for both sides seems to be 19...Qb6 20.Bc4 Bd2+! 21.Kf1! Qxb2 22.Rd1 Qxd4 23.Qd3 Qxd3 24.Bxd3 Bc3 25.g3 when Black has two pawns for the exchange but one can hardly speak of an advantage for him.
Instead of 21.Kc2 White can try 21.Rb1 but after 21...Re1+ 22.Kc2 Qa4+! 23.b3 Qxa3 24.Rxe1 Black has a mating attack: 24...Qa2+ 25.Kd1 Qa1+! 26.Ke2 ( 26.Kc2 Nb4#) 26...Qb2+ 27.Kd1 Qc1+ 28.Ke2 Nf4#) These lines were given by Magomedov.
18...Kxf8! 19.gxf3 Qa5+ 20.Kd1 Rd8!
In Magomedov's line given above, this rook went to e8 in order to get at White's king. The switch to d8 is a pleasing variant on this idea. White's king is just as vulnerable on the d-file, the isolated pawn being a minor obstruction that will soon be swept away.
Black mates after 21.Qc5+ Qxc5 22.dxc5 Nb4+! 23.Ke2 Rd2+ 24.Ke1 Nc2#.
The editors of Informant 76 pointed out that 21...Nb4 wins more quickly, but the text does not spoil anything.
22.Qc3 Qe5! 23.Qe3
Or 23.Re1 Rxd4+ 24.Kc2 Nxe2, etc.
23...Rxd4+ 24.Kc2 Qf5+ 25.Kb3 Qd5+ 26.Kc3 Nxe2+ 27.Qxe2 Rd2 28.Qe4 Bf6+ 0–1