The First Man-Machine
Poker Championship

Laak and Eslami: Meet Polaris

By Christopher Hunt

First, there was "Chinook" - a software program that shocked onlookers with a world title over human checkers savant Dr. Marion Tinsley in 1994.

Then came the infamous IBM supercomputer "Deep Blue," relentlessly battling world chess champ Garry Kasparov to his breaking point in 1997.

Tomorrow, the world meets "Polaris" - the world's most advanced poker-playing computer program. And we'll soon see if the scientists who put it together have cracked the last code standing in its way to poker dominance - human adaptability.

Put together by an award-winning team of researchers from the University of Alberta, the program will take on two of the best poker players in the world - Phil "The Unabomber" Laak and Ali Eslami (hand picked by Laak himself to be his partner) - in the first-ever real-money match between man and machine.

The stakes: $50,000. The location: Vancouver, Canada, as part of the annual conference of the Association for the Advancement of Artificial Intelligence (AAAI).

If the scenario sounds familiar, that's because it is. Two years ago at the 2005 WSOP Laak took on and defeated - barely - an earlier version of the software called Poki-X.

The difference this time: They've improved the program, incorporating more and more human abilities - to bluff, and to read an opponents style of play and adjust accordingly. They've also put their own money on the line to prove it.

"This match is extremely important," University of Alberta computing science professor Jonathan Schaeffer told the Associated Press last week, "because it's the first time there's going to be a man-machine event where there's going to be a scientific component."

The set-up will be four 500-hand duplicate matches, with Eslami in one room and Laak in another. In each match the same series of cards will be dealt, with teammates playing the opposite hands in each game. So whatever cards Eslami gets in one hand will be the same the computer gets against Laak, and vice-versa. Community cards will be the same for both, and no communication will be allowed.

At the end of each session, the combined bankroll of the human team will be compared to the combined bankroll of the bot team to determine the winner.

If the human team wins by more than 25 small bets in a session, they'll take $5,000. If it's 25 or less (a statistical tie), the players will get $2,500 per session.

The four separate sessions will be played over two days, allowing both teams to learn more about their opponent and adjust their strategy accordingly.

The question on everyone's minds: Does the computer stand a chance against two of the top pros in the world?

Says the research team:

"We think it might. At the very least, we don't think it will embarrass itself terribly. Can a computer program bluff? Yes… probably better than any human. Bluff, trap, check-raise bluff, big lay-down… name your poison.

Does the computer really stand a chance? Yes, this one does. It learns, adapts, and exploits the weaknesses of any opponent. Win or lose, it will put up one hell of a fight."

Want to know how it turns out? Anyone who is registered to attend AAAI 2007 in Vancouver can join the audience to watch, and a live blog will be updated during the event. Match logs will also be available after the event.

To tune in or to find out more, visit the University of Alberta Research Group's Web site here.