Twenty years ago today, in 1997, the chess-playing computer Deep Blue beat human chess champion Garry Kasparov. The two adversaries had faced off in a six-game match the year before. The computer won the first game, but Kasparov won the contest. So IBM went back to work and upgraded Deep Blue.
When the time came for the rematch, Kasparov won the first game easily. And in the second game, he laid what he considered to be a foolproof trap for the computer — but the computer didn’t go for it. It made a completely unexpected move. That rattled Kasparov’s confidence, and he was confused even more when the computer’s next move was a really bad one. Kasparov was visibly frustrated, and eventually got up and left the stage, forfeiting the game. “I lost my fighting spirit,” he later said.
It turns out that that unexpected move by Deep Blue was probably due to a glitch in the software. It was faced with so many choices that it couldn’t decide which move to make, so it just picked a move at random. A later analysis shows that Kasparov could have played that game to a draw, but he had psyched himself out, convinced that the random move was a sign that Deep Blue had a long-term strategy that he, Kasparov, was unable to visualize. And in game six, with the match tied at two and a half games each, Kasparov misplayed his opening. Deep Blue took advantage and defeated him in 12 moves.
IBM’s work on Deep Blue led to the development of Watson, a computer that played against humans on the game show Jeopardy, and won. Deep Blue has retired and now lives in the Smithsonian’s National Museum of American History in Washington, D.C.