Kasparov Beats Deep Blue: Human Chess Triumphs

The IBM computer Deep Blue defeated world chess champion Garry Kasparov in a six-game match on February 17, 1996, or rather, it didn't. Deep Blue won one game, Kasparov won three, and two were drawn. Kasparov won the match. But the single game Deep Blue took was the first time a reigning world champion had lost to a computer under standard tournament conditions, and it sent a shudder through the chess world. The rematch came in May 1997. This time, an upgraded Deep Blue won the match 3.5 to 2.5. Kasparov accused IBM of cheating, claiming the machine's play in the second game showed signs of human intervention. IBM denied it but refused to provide the computer's logs and dismantled Deep Blue shortly after the match, denying Kasparov a third encounter. The company's stock rose significantly after the victory. Deep Blue evaluated approximately 200 million chess positions per second. It used brute-force calculation combined with an evaluation function tuned by a team of grandmasters and computer scientists led by Feng-hsiung Hsu. The machine had no understanding of chess in the human sense; it searched a vast tree of possibilities and selected the statistically optimal move. Kasparov, by contrast, evaluated roughly three positions per second but understood patterns, strategy, and psychology at a level no computer of that era could match. The defeat raised questions that extended far beyond chess: if a machine could beat the best human mind at the most intellectual of games, what activities remained uniquely human? The answer, as the next three decades would demonstrate, was fewer than most people expected. But in February 1996, the story was still about one game in a match that Kasparov ultimately won, a warning shot from a future that wasn't quite ready to arrive.