Kasparov's ruin

Game 4

D.Shapland – L.Vallans
Sheffield League
Sheffield
4th February, 1997
Caro-Kann Defence: Karpov-Smyslov Variation

This game was played during a brief stint of league chess that I played in Sheffield which took place after I’d left University but before I’d started in full-time employment. I played for the University of Sheffield for a season, sometimes for their ‘A’ team but more often for their ‘B’ team. They had a strong ‘A’ team which won the Davy Trophy in the 1996-97 season. I can’t remember which team or league this game was played in, but I like to tell myself that it contributed in some small way to helping the ‘A’ team win the title.

Aside from the fact that I very much enjoyed playing this game – who doesn’t enjoy delivering check mate after a king hunt in 23 moves! – I think it’s of interest because the exact same opening idea was seen exactly 25 years ago today in the final game of the match between the World Chess Champion of that time, Garry Kasparov, and IBM’s Deep Blue computer in New York just three months after I played this game.

The story of the Man vs. Machine match is interesting in its own right and Kasparov has had much to say about it in various interviews since then. It’s often forgotten that he had already won a match against a previous incarnation of the IBM super-computer the year before the New York match. He won the first game of the 1997 match too but then, in game 2, resigned in a position it now appears he could have held and was psychologically rattled by a decision made by the Deep Blue in that game. The World Champion felt that the computer was benefitting from ‘human assistance’ to alleviate some of its short comings.

Games 3, 4 and 5 were drawn, but Kasparov had been so distracted by what had happened in game 2, and in subsequently looking for more signs of outside intervention to help build his case, that he made a terrible decision in the final game of the match.

Playing with the Black pieces the World Champion decided to play an opening line of the Caro-Kann Defence that was highly topical and known by elite Grand Masters of the day to be extremely dubious. In the diagram shown in the title image above Black has just played 7...h6?! attacking the White knight on g5 which now must retreat to e4 or h3. Or does it? What would you play here with White? Find out the best move in the game commentary that follows.

The refutation to Kasparov’s move required Deep Blue to make a long-term piece sacrifice that it couldn’t possibly calculate through to a clear conclusion. On that basis Kasparov seems to have assumed that the silicon monster would not play the sacrifice but instead opt for another continuation that would give him a perfectly acceptable game.

It was a huge gamble and it back-fired abjectly. Subsequently it transpired that Deep Blue’s programmers had added this exact variation into the computer’s opening book the day before and so, when the critical moment was reached, it didn’t even consider playing any other move than the correct one it had been programmed to make. To Kasparov’s horror he found himself defending a diabolical position that ironically only a super-computer would have any chance of holding. He collapsed to a swift and devastating defeat and the match was over.

I remember picking up a newspaper with the moves of this fateful game in it and being dumb founded by Kasparov’s choice. The variation was one that I’d recently spent time looking at myself from the White perspective and had won with only three months before.

Here then, is that precursor game to Kasparov’s ruin.