Chess world champion Gary Kasparov has been defeated by Deep Blue, the world's highest ranking chess computer. The match ended when Kasparov, who described himself has having "no strength left to fight", made an uncharacteristic error in the deciding game leaving Deep Blue the winner of the six game match: 3.5 to 2.5.

Deep Blue's main strength is its speed of computation. It can analyse millions of possible moves a second. In contrast, human players typically look no more than two or three moves ahead. However, human players also plan their moves with the whole game in mind. The number of possible moves in chess is huge and despite their speed, chess computers cannot normally analyse the game far enough ahead to avoid being tricked by skilful humans into weak positions.

In February last year Kasparov exploited Deep Blue's weakness to beat it 4-2. Deep Blue's computer programmers worked hard over the Summer and in the Autumn of last year computer giant IBM offered 1.1 million US dollars for the rematch.

The mathematics of game-playing is highlighted elsewhere in this issue (see "The mathematics of games") and is collectively described as "Game Theory". However, studying the way people and machines play games can have some surprising applications - for example, game theory is currently being used to analyse the actions of criminals to help formulate policing policies in the North-East of England.

Kasparov vs Deep Blue - the rematch

IBM's web-site dedicated to chess.

Club Kasparov

Gary Kasparov's Home Page.

Game theory backs crackdown on petty crime

By Robert Matthews, New Scientist, 19th April 1997.