Beaten by a microchip

This has been a bleak week for mankind. On Monday, Britain’s top chess player Michael Adams, who is rated seventh in the world, concluded his best-of-six-games match with the Hydra supercomputer. He had known that taking on this monster — a 64-way cluster computer that analyses 200-million moves per second — was going to be tough. But he surely can’t have envisaged quite how voracious the monster would be: Adams managed a solitary draw in the six games, losing heavily in the other five. As the Hydra team crowed at the end, ”Man is dead — long live the machine.”

There are a variety of ways of viewing this result. First, you have to feel sorry for Adams, a 33-year-old chess professional. When Garry Kasparov lost to Deep (or, more accurately, Deeper) Blue in 1997, at least the result was close — the machine squeaked it by a single point. But he was still psychologically wrecked and spent the next five years telling anyone who would listen that the computer was getting human assistance during games. It is hugely upsetting for a chess grandmaster to be beaten by an assemblage of microchips.

So, bad news for Adams. But even worse news for chess as a whole. It is already in a trough, thanks to the charismatic Kasparov’s recent retirement and political shenanigans within the sport that have left it without a generally agreed world champion. Now that a wretched machine is demonstrably better than any human, will anyone care who the best non-silicon player is? If a machine could be constructed that could beat Roger Federer, would we be so glued to Wimbledon?

The third point — and, admittedly, this is science-fiction — is the question of what machines can ultimately achieve. If they can outwit man over the chessboard and, in the space of 20 years, develop from basic chess calculators to sophisticated super-grandmaster-level chess players, who knows what they can ultimately do.

Remember, Neanderthal man once thought he was in control. If machines can master the Sicilian Defence (Scheveningen Variation), none of us is safe.

The aptly named Hydra is financed by the Abu Dhabi-based PAL Group, and the computer is housed in a secure server room in the emirate. All Adams could see during his match at Wembley Conference Centre in north London was a man dressed in black with a shiny laptop, connected to the monster by an internet link.

I met the man in black, German computer expert Ulf Lorenz, before the fourth game in the Adams match. In fact, Adams’s only chance may have been our photographer, who — just before Adams arrived and much to the Hydra project manager’s annoyance — started manoeuvring the laptop to get a better picture, in the process almost dislodging a wire. Oh, that he had succeeded — mankind might have been saved!

Lorenz let me play a game against his Frankenstein’s monster. I am a fair-to-middling player and lasted a respectable 38 moves, but I can’t claim Hydra was ever sweating. He — why do I assume Hydra is male? — grabbed a pawn, I launched a speculative attack and eventually got my comeuppance. It was a little disconcerting when, at about move 32, Lorenz told me Hydra was predicting ”mate in eight”. The thing isn’t just a cold-blooded monster; it likes to gloat, too.

Lorenz claims Hydra is now the strongest chess-playing entity in the world. ”We think we have crossed the 3 000 ELO line,” he says. ELO measures the strength of chess players; Kasparov, at his strongest, would have been just above 2 800, so to reach 3 000 is like cutting 20 seconds from the world record for the 1 500 metres. No human player could compete with that.

”Once it was thought that humans played almost perfect chess,” says Lorenz.

”Grandmasters were seen as perfect and it was said that no one could ever play ELO 3 000 because there’s a limit. Now we see this is wrong. It was also said a grandmaster would always be able to get at least a draw by keeping control of the game. That has been proved wrong, too.”

Twenty years ago, no one would have believed a computer would beat a grandmaster. Now, with Adams’ defeat, the game is up — grandmasters are not just being beaten, they are being humbled. ”Surely it is the end of the line for us human chess players,” I suggest to German grandmaster Christopher Lutz, another key member of Hydra’s programming team.

He tells me not to be so melodramatic. ”There are certain things a computer can simply do better than a human,” he argues. ”I am not really worried about it. It is still worth humans competing against each other.”

I hope Lutz is right, but fear for the future of chess. The grandmaster — eccentric, all-seeing, unbeatable — was an iconic figure in the public mind. Now he has been rumbled. The grandmaster sees seven or eight moves ahead; Hydra calculates twice that number on average, more if it exerts itself. Hydra’s programmers say that human players rarely play 10 optimum moves in succession; the computer plays optimum moves all the time. It has never lost to a human; perhaps it never will.

”Chess at a crossroads” proclaimed the posters at Wembley Conference Centre. Now the road has been chosen — and, speaking as a human, I don’t care for the destination.