Jul 19, 2007

A crowning achievement?

Playing chess is like looking out over a limitless ocean; playing checkers is like looking into a bottomless well.
—Marion Tinsley, world checkers champion 1955-58, 1975-91

Not so bottomless, apparently. A research team at the Yoo of Eh has solved checkers.

Like the far simpler tic-tac-toe, the game is a theoretical draw: unless someone makes a blunder, it will always end in a tie. The lead researcher's checkers program, Chinook, is now literally unbeatable.

Bobby Fischer once said he believed that chess was also a theoretical draw. In terms of complexity, chess is to checkers as checkers is to tic-tac-toe, but Fischer's assertion seems likely. The better the players, the more likely the game is to end up drawn. At the grandmaster level, entire matches are decided by the few games that don't end that way.

Having committed my share of thudding blunders over the board, I find something reassuring about this yin-yang balance, the knowledge that no matter what White pulls out of its bag of tactical tricks, Black can always restore harmonious equilibrium with the right counter. It reminds me of music, the way tension is created, then finally resolved in the return to the keynote. From chaos, order.

But there's something terribly boring about this orderly perfection, isn't there? Even if these programs eventually reduce the unfathomable possibility of chess to a foregone conclusion, it will only prove that old technophobe maxim: to err is human, but to really fuck things up takes a computer.

No comments: