Computers Are Great at Chess, But That Doesn’t Mean the Game Is ‘Solved’
On this day in 1996, the computer Deep Blue made history when it beat Garry Kasparov
On this day 21 years ago, the world changed forever when a computer beat the then-chess champion of the world at his own game.
On February 10, 1996, Deep Blue beat Garry Kasparov in the first game of a six-game match—the first time a computer had ever beat a human in a formal chess game. Two other games in that match were draws. The next year, Kasparov and Blue faced off again and Kasparov lost the match. It was a new frontier in computing.
“The result was met with astonishment and grief by those who took it as a symbol of mankind’s submission before the almighty computer,” Kasparov wrote in 2010. It was a pivotal moment in computing, one that changed both computers and chess forever.
Two decades later, computers now regularly beat humans at chess, writes Klint Finley for Wired. The great contest of man-versus-computer chess is over. “Today, for $50, you can buy a home PC program that will crush most grandmasters,” Kasparov wrote. The search for a computer that can beat even the best at chess was only really interesting between 1994, when computers were too weak, and 2004, when they got too strong.
Although that contest is over, he wrote, there is still a wealth of complexity to plunder. Having a computer opponent can help chess players train, writes Finley, but Kasparov also said the original draw of teaching computers to play chess wasn’t just about teaching them to win.
“There were other goals as well,” he wrote: “to develop a program that played chess by thinking like a human, perhaps even learning the game as a human does.” That leads to the next computing frontier for chess: solving the game altogether—playing an objectively perfect game.
Computers have been able to beat humans in ever-more complicated games, like Go. Not even poker, which Kasparov declared to be the next frontier in computer-versus-man games in 2010, solely belongs to humans anymore: a computer just defeated a human opponent in poker for the first time.
Despite these advances, the slow quest to solve chess is still ongoing, writes Michael Byrne for Motherboard. “Solving chess means computing every possible move and every possible update until there are no more possibilities,” he writes. “Nothing less.” According to current estimates, he writes, there are more possibilities in chess than there are atoms in the observable universe. That means it’s a quantum problem—and we don’t have a large-scale quantum computer. Yet.