Chess: What Did The Deep Blue See?

Chess Sets     Material Rewards     Sissa's game     Deep Blue

In an early "Star Trek" episode, familiar to series junkies, Mr. Spock engaged the Enterprise’s on-board computer in tri-level chess, and won all five games he played. Because Spock programmed the computer to play chess himself, he knew that such an outcome should have been impossible, and that at best he would only have been able to achieve a draw. This insight led him to realize the computer has been tampered with; he used the knowledge to save Captain Kirk from a court martial. The computer connection was a handy plot device, but chess enthusiasts might find a more interesting computer-human chess confrontation in the real world. An I.B.M chess-programmed computer dubbed "Deep Blue" took on, in 1997, the World Chess Champion, Garry Kasparov. Streamed live on the IBM website, the six-game match, it was viewed by millions of computer experts and chess fans. Exactly what was Kasparov facing? A computer’s talents--namely, computation speed, and memory capacity. Pit those against a Grand Master’s abilities to recognize the ever-changing patterns of a chess match, and to plot corresponding strategies. Deep Blue could, in response to a Kasparov move, calculate in a single second almost two hundred million possible responses. But did Deep Blue actually understand the patterns of each of those positions, and the consequences of each as a countermove? Chess masters, psychological tests have determined, become chess masters because of their uncanny ability to recall real-game board patterns, their underlying strategies,and tactical positioning. But if a board contained pieces placed at random, with no specific strategy implemented, the chess masters’ recall was no different from that of non-chess players. Deep Blue instead utilized its speed in analyzing the positional changes Kasparov’s moves created, and searched "ahead" in its database for possible moves Kasparov might make in response to the moves it made. The more moves, or "plies", Deep Blue anticipated, the more successfully it played. Deep Blue, however, displayed an inability to adapt; it always gave the same response to a specific situation. Human chess players can sense changes in their opponent’s demeanor, detecting a more or less aggressive style, and factor those changes into their decisions. Deep Blue won the 1997 match, 3.5-2.5, becoming the first computer to defeat a Grandmaster. It then retired from the Chess wars, but one of its racks resides in the "Mastering The Game: A History of Computer Chess" display at Boston’s Computer History Museum. Mr. Spock, one assumes, would have known all about it.