A Beautiful Waste
The ascendancy of machine over man
Tempus fugit. It is almost nine years since this column began, in the first issue of Standpoint. The founder-editor, Daniel Johnson, a great student of chess history, had the idea that I should write about the game in essay form: it is an idea I have delighted in exploiting, for every issue of the magazine’s existence.
But I have probably written enough such essays: I have begun to repeat myself. So this is the last of the series, with apologies to any readers who have an appetite for more of the same.
It coincides with the 20th anniversary of what we must now consider to be the most significant match in chess history. Not the Fischer-Spassky match: that was as long ago as 1972. No, it’s the match played at the beginning of May 1997 — between the world champion Gary Kasparov and the Deep Blue chess computer program.
IBM had sponsored the programmers, led by its employee Feng-Hsiung Hsu, in order to demonstrate what had hitherto been impossible: that computers could beat the world’s best carbon-based life form at a pursuit thought too amorphous to be captured by algorythms and digital calculations.
Kasparov had beaten Deep Blue in a match the previous year and was confident he could do so again. He underestimated just how fast IBM’s programmers had been improving their ability to make the machine “understand” chess — and the pure number- crunching power of computer chips was increasing dramatically: the theoretical maximum search speed of Deep Thought in 1997 was a billion positions per second.
Kasparov’s confidence was only increased when he won the first game of the rematch. But in the next game he was — as we learnt immediately afterwards — completely disconcerted by what he saw as strategically masterful moves by an opponent hitherto regarded as a mere calculating machine. So perplexed was Kasparov that he resigned the game in a drawn position: his state of mind was not improved when this was pointed out to him afterwards by one of his team.
This illustrated one of the biggest advantages that machine has over human. We become exhausted and upset — which affects both our concentration and our peace of mind. The computer program has no fear because it has no feelings. And it never runs out of energy, unless it is unplugged (which, of course, is what Dave the astronaut does to HAL in 2001: A Space Odyssey, when he realises that the computer — which also played chess in the film — is bent on destruction).
This helps explain Kasparov’s otherwise inexplicable performance in the final game of the match, when the scores were level and it was all to play for. The human champion, right in the opening, allowed a sacrifice already known to be dangerous: it had occurred in previous grandmaster games. In his book, Behind Deep Blue, Hsu suggests that Kasparov had bet that the machine would not play a move that involved giving up material for an attack of uncertain outcome. But Deep Blue did exactly that. Hsu records: “Gary acted a little surprised, but then . . .” Well, you can read what actually happened next in the game notes at the end of this column.
In the intervening 20 years, computer programs have developed at such a pace that the difference in chess ratings between those “things” and the world’s current top player, Magnus Carlsen, is about as wide as between me (a mere county-standard player) and a strong grandmaster. In other words, they are completely invincible.
Some see this as “the death of chess”. But while there have been sad consequences — something of the mystery of the game has been destroyed and with it the mystique of the greatest human players — more young people than ever are playing and enjoying the game. And while it is true that both the opening stages of the game and the endings have been increasingly solved by the programmers, the middle-game is still delightfully obscure: we remain enthralled by the creativity shown by humankind as they battle for intellectual supremacy over the 64 squares. And chess remains, as I wrote in my very first column, “beautiful enough to waste your life for”.
Here, then, is the final game from the 1997 match that ended the chessboard supremacy of carbon over silicon. 1.e4 c6 (Kasparov plays the Caro-Kann Defence; far from his usual choice, but apparently suitable against his non-human challenger) 2.d4 d5 3.Nc3 dxe4 4.Nxe4 Nd7 5.Ng5 (This looks strange, but there is a point, as we will see) Ngf6 6.Bd3 e6 7.N1f3 h6?! 8.Nxe6! (Precisely so! Deep Blue sacrifices for an obscure attack — not the sort of move then associated with computers, which prioritised material gain) Qe7? (This is probably Kasparov’s biggest mistake. Later it was shown that Black’s best defensive chances lie in capturing immediately with 7…fxe6, since after 8.Bg6+ Ke7 Black will be able to place his Queen on c7, challenging the crucial h2-b8 diagonal). 9.0-0 fxe6 (if 9…Qxe6 10.Re1 pins and wins the Queen) 10.Bg6+ Kd8 11.Bf4! (It was discovered afterwards that all the previous moves were in Deep Blue’s “Opening Book”. But this first move of its own calculation is excellent) b5? (Another mistake from Kasparov. He seeks to prevent White from playing c4, but this just allows more attacking lines to be opened) 12.a4 Bb7 13.Re1 Nd5 14.Bg3 Kc8 15.axb5 cxb5 16.Qd3 Bc6 17.Bf5! exf5 (Kasparov jettisons his Queen for Rook and Bishop, in a vain attempt to reduce White’s attacking potential) 18.Rxe7 Bxe7 19.c4!…
At this point, to the astonishment of the watching television audience, Kasparov could be seen to mutter something before rushing from the board waving his arms in a gesture of angry helplessness.
In fact, his resignation after just 19 moves was hardly premature, as can be seen by the plausible variation 19…bxc4 20.Qxc4 Nb4 21.Re1 Re8 22.Rxe7 Rxe7 23.Qxb4 Re6 24.Qc4 Rf6 25.Ne5 and Black’s entire position drops off. Sic transit . . .