Seeing as the World Championship is almost upon us, it's a good moment to recall this lovely move, knight sacrificing knight, from nearly twenty years ago. I'm sure I saw it first in CHESS, since we were still then in a world, one which may be incomprehensible to younger readers, where games from the leading tournaments were not live online. Or indeed online at all.
It's been a while since the Queen's Gambit Accepted was in fashion, but it was back then. In the previous year's Candidates' match Short had tried it twice against Karpov, drawing both game three and game five and having the better of both of them (in contrast to the Black game before, which he lost with the Budapest Gambit, and the Black game after, which he lost with the Queen's Gambit Declined). I was impressed: by the end of the year I was playing it myself.
The Karpov-Short match took place in Linares, where, at the then-annual tournament, today's favourite move was also played. The position looks at first glance like one of those perverse ...b5 lines that Benko players play if you don't play 2.c4, e.g. something like the Benko Deflected 1.d4 Nf6 2.Nf3 c5 3.d5 b5 4. Bg5 Ne4 that you can see here. Black's apparently played a line with ...Qa5+ and is hoping to make something of the jam along that diagonal - and maybe get at the d5 pawn - before it all goes pear-shaped.
But it's not any kind of Benko. Whoever could have known the QGA could be so interesting? (Doubly interesting to me since I was playing it now.) But enough about the opening - what a move! And what a massacre!
It's arguable that the best move isn't actually my 9...Nb4, but the follow-up 11...g6
which definitively sacrifices a whole knight in return for turning White's position into marmalade. Maybe. In some ways I'm reminded of one of the greatest of all chess games, from Fischer's 1963/4 US Championship whitewash, and particularly this position
of which Wade and O'Connell in The Games Of Robert J Fischer (Batsford, 1972) comment:
The intriguing question is to know when Fischer switched his thinking from an assessment of the position based on 18...Nxd1 19.Rxd1 [I have translated descriptive into algebraic - ejh] when White has two knights for rook and one (two?!) pawns(s) - possibly a tenable position for Black, to that of trying to mate the exposed White king.
And of course Byrne is quoted in My Sixty Memorable Games as saying
I sat wondering why Fischer would choose such a line, because it was so obviously lost for Black.Fischer by contrast says no more than "White is all right again", which gives us three different opinions of the same position, but the reason the position comes to mind, apart from some very slight similarities with the position in Gelfand-Anand, is the idea that Black might have considered an obvious manoeuvre to win back the exchange for a piece already sacrificed, but then changed his mind to make it a real piece sacrifice after all.
Or, even more so, that White might have been expecting Black to do this, and then been rocked back when he did not.
Be that as it may, it's hard to believe that Gelfand was really expecting 9...Nb4 any more than he had seen 11...g6: in fact, one wonders whether he had already lost his way by the time he played 9.f3. Was he expecting 9...g5, even though to my mind (or more accurately, my computer's) it also looks promising for Black?
I don't know. It's funny, in retrospect it looks like White rather than Black who was playing recklessly. Maybe he was. I don't know what Gelfand was thinking and I don't know what happened to the theory of that line, though presumably something did since I don't recall seeing it since.
But I do recall seeing the game in CHESS. And I do recall seeing 9...Nb4. And I do recall thinking that it was absolutely fantastic.
[My favourite moves index]