Silicon beats carbon in chess battle

Ouch

By Lucy Sherriff, 4th July 2005

Customer Success Testimonial: Recovery is Everything

In the battle of man vs. machine, man has been utterly trounced. In fact, supercomputer Hydra's five triumphs out of six over human grandmaster Michael Adams dwarfs even the All Blacks' recent demolition of the British Lions. Adams managed to squeeze a draw in the second bout, salvaging some pride.

Hydra was well equipped for the battle, which concluded last week. With its many, many pieces of well-deployed silicon, it can analyse up to 200 million chess moves per second, and plan its game 18 to 40 moves ahead. As we reported earlier, that is six more moves than IBM's Deep Blue was capable of.

The result is being touted as a demonstration of just how far supercomputers have come in the last decade. In 1996 Garry Kasparov actually beat Deep Blue, although the supercomputer got its own back a year later.

Computers might have the edge over people when it comes to chess, but that doesn't mean they actually understand the game. After each move in chess, there are approximately 30 legal moves that could be made next. This is known as having a branching factor of 30. A sufficiently fast computer can, with the help of some decent software that prunes the dead ends, merely brute force its way through enough moves to calculate the one most likely to win it the match.

Computers have yet to master the game of Go, a much trickier proposition with its branching factor of 250. Brute force is no good here, because the number of moves gets really big, really fast. Researchers at Microsoft's labs in Cambridge have made some improvements to computer Go, using Bayesian ranking to calculate probable good moves. But the researchers concede that even the best programs are pretty easily dispatched by good human players. ®