The PJ Tatler

Busted Flush: Computer Cracks Texas Hold 'Em

Looks like you’re plumb out of luck, pardner. But you just got cleaned out in a head-to-head showdown for a good cause:

Scientists have now created a program that has essentially solved the two-player limit version of Texas hold ’em poker… This doesn’t mean a machine always knows when you’re bluffing and will beat a human every time, especially since so much luck and guile is at play in poker. What it means, statistically speaking, is that there is no significant chance to see if the computer’s strategy is not the best within a human lifetime of play — 200 games per hour for 12 hours a day for 70 years. The team published a study about the program this week in Science.

Solving this kind of poker is more than just fun and games, researchers say. “Our algorithm can attack a number of problems that have similar aspects to this game — adversarial settings where two players have antithetical goals against each other,” says lead study author Michael Bowling, a computer scientist at the University of Alberta in Edmonton, Canada. “One application is in the realm of security. In the post-9/11 world, we want to set up security regimes that can protect strategic infrastructure like our airports and our port systems.”

Give it a try for yourself here, although at the moment the service seems to be unavailable. Meanwhile, here’s the link to the study in Science.