The Games We Teach Computers to Play
Computers are sophisticated enough to play a flawless game of checkers. They can beat the world’s greatest chess masters. But humans still put them to shame on the Go board.
Go is an ancient strategy board game, known for its simple rules and incredibly complex playing possibilities; it has more possible outcomes than there are particles in the known universe. But as Wired reports, computer scientists are starting to get a handle on the game, using what’s known as the Monte Carlo method — which, by the way, involves the same kinds of algorithms that drive computer simulations on Wall Street.
Go-playing computers are just now getting good enough to beat mid-range professional human players, and it may take a decade or more before they reliably beat the best players. Hopefully it won’t take them that long to beat our financial markets.