FREE counter and Web statistics from sitetracker.com
collision detection
content | discontent
send me yours
January 10, 2008
The death of checkers: My Times Year in Ideas piece












A couple of weeks ago, the New York Times Magazine published its 2007 "Year in Ideas" issue -- their annual compendium of the year's most interesting and offbeat research. I wrote up five of their scientific and technological entries. The entire issue is online here for free, but I'm also archiving my pieces here for posterity's sake.

This one is about "the death of checkers" -- the story of the scientist who solved the game. Reinhard Hunger did the awesome illustration that is thumbnailed above; check out the full-size version on the Times' site!

The Death of Checkers

Checkers has been around for more than 400 years, has been enjoyed by billions of players and has taught generations of young children the joy of strategy.

And now it's all over. This July, Jonathan Schaeffer, a computer scientist at the University of Alberta in Canada, announced that after running a computer program almost nonstop for 18 years, he had calculated the result of every possible endgame that could be played, all 39 trillion of them. He also revealed a sober fact about the game: checkers is a draw. As with tic-tac-toe, if both players never make a mistake, every match will end in a deadlock.

One upshot is that Schaeffer now possesses software that can play unbeatable checkers. Indeed, go to his Web site and you can play online yourself, providing you're prepared to lose again and again and again -- or maybe, just maybe, fight to a draw, assuming you, too, play with the crystalline perfection of a silicon brain.

Schaeffer did not solve checkers by replicating human intuition or game-playing ability. Rather, he employed what’s known as a “brute force” attack. He programmed a cluster of computers to play out every possible position involving 10 or fewer pieces. At the peak of his labors, he had 200 computers working around the clock on the problem, both in Alberta and down in California. (The data requirements were so high that for a while in the early '90s, more than 80 percent of the Internet traffic in western North America was checkers data being shipped between two research institutions.)

The brute-force method is slow, which is its big limit. Schaeffer says he suspects you couldn’t use it to solve chess, because that game -- with between 10 to the 40th and 10 to the 50th possible arrangements of pieces -- is far more complicated than checkers, which has 5 × 10 to the 20th positions. “Chess won't be solved in my lifetime,” he predicts. “We need some new breakthrough in technology to do that. CLIVE THOMPSON

Posted by Clive Thompson at January 10, 2008 10:31 AM

Trackback Pings

TrackBack URL for this entry: http://www.collisiondetection.net/mt3/mt-tb.cgi/1716

Comments
Post a comment

Thanks for signing in, . Now you can comment. (sign out)

NOTE: If you posted a comment and you can't see it -- try refreshing your browser.


Remember me?