Schachmaschinen: "a horse, a horse, my kingdom--"

An item on the list of Things Everyone Knows: that, without qualification, computers are better than humans at chess. Ever since Kasparov lost the first game of the original 1996 match against Deep Blue it's been downhill for chesspeoples looking to protect their mystique. It seems the Royal Game is fit just for calculators. Whereas in the 19th century it was held that a master's gift was for inspiration and fantasy, and optimism so reigned it was felt that a player could in the most seeming-rotten position discover some extraordinary resource to save the day --

-- or so the chess mystifiers and hagiographers tell us. I would doubt it, except that not too long before Everyone Knew that all functions are continuous and with one or two silly exceptions differentiable and so forth, which is a right bit of fantasy itself -- albeit abetted by ambiguity in the meaning of the operative terms.

Anyway, Kasparov lost the game, then stormed back to win the match 4-2 (+3=2-1), winning the sixth game in a particularly gruesome manner; he was in such high spirits as to spontaneously agree to a rematch during the press conference; and then next year came and crumbled in front of the world, losing the 1997 rematch 3.5-2.5 (+1=3-2) after another particularly ignoble sixth game, dual to the previous match; and that was it, chess is something you can calculate, and, incidently, all the hopes of the early AI pioneers that it would take real intellect to make a fine mechanical chess machine -- just vapour, like so much else.

But one didn't have to be a chess mystifier to know it just wasn't so. The computers were still weak relative to the best, demonstrably so, to those wanting to take enough time to demonstrate it, even if it was long past the time a humble patzer could hope to score a few points off his laptop any more than off a flesh-and-blood master, and Kasparov's loss was far from convincing.

So despite Deep Blue's "retirement" the experiment continued. The computers got better. Man still had some triumphs; for example, Ilia Smirin's 2002 match against four leading programs, which he won, 5-3 (+2=6-0). The game against "Gambit Tiger" was particularly curious and prompted one commentator to remark ironically that although computers were superior in strategy to humans, we could still overpower the machines tactically, (as Smirin did in a dubious position).

Fast-forward to 2004, with two high-profile matches between leading players and leading machines: Kramnik v. Fritz and Kasparov v. Junior. (You can always tell which side is computer by the goofy name, and which side human by the Russian name.) In the former Kramnik looked to be coasting to an easy victory before dropping two points in the second half to end 4-4 (+2=4-2), once simply hanging a piece and once playing a slightly speculative combination with a very deep refutation which the machine very impressively found. In the latter things looked up when Kasparov cooked the machine in a knife fight in the first round, then slightly down when in a later game he overlooked a mating attack from the machine, then slightly askew when in the final game he agreed to a draw in an unbalanced position -- in front of millions (maybe), since the game was televised on ESPN (!). Ending up at 3-3 (+1=4-1), one could be forgiven for having lost almost all interest in Man-Machine matches.

Then, of course, there was Hydra, which pretty much made it mandatory to lose all interest. ("Lasciate ogni interesse, voi ch'entrate"?) Finally the knowledge that the tail wags the dog came true with a painful demolition in Adams v. Hydra. (Finally not a Russian, but still, to be sure, a world-class talent. For the record, Smirin is Israeli.) This travesty took place just a few months ago and saw the bon vivant Englishman lose 5.5-0.5 (+0=1-5). Ugh. Even worse, the prize fund was structured so that money only came for scoring points (unusual: usually the purse is split in some specified way to the winner and loser of the match, irrespective of the score), so for enduring this drubbing poor Michael Adams picked up only 10 thousand. (For comparison, Kasparov and Kramnik pocketed mid-six-figure sums for their labours -- minus pay to their seconds, and so forth.)

Finally, today started the second annual (so far) Man v. Machine tournament in Bilbao, Spain. Three men, all former "world champions" (so to speak; they comprise an Uzbek, a Ukrainian, and a -- Russian), against three leading programs, two microprocessors you too could own (Fritz and Junior) and one supercomputer (Hydra, again). Care to guess the day's score? Yes, that's right.

Certain chess fans might note encouragingly that all is not yet completely lost: correspondence chess players, who play games over duration of months instead of hours, sending moves by post, still know how to take the machines' measure. But this wagging wolf will just give a desultory woof!. It's all true. If a chess titan spends months preparing for a match, then we'll watch it, for sports' sake. But these sideshows are dispiriting.