In last week’s Brains Vs. Artificial Intelligence poker tournament at Pittsburgh’s Rivers Casino, the human brains beat the computer. But according to some, the declared win may have been a draw.
Four of the top poker players played 80,000 hands of Heads-up No-limit Texas Hold’em–a two player game poker game–in total against the artificial intelligence program Claudico. They played over the course of two weeks, according to Information Week.
By the end of the two-week competition, three of the four players cashed in more money than Claudico. The poker pros beat the computer’s final total by $732,713 collectively.
Doug Polk, Dong Kim and Bjorn Li were ahead of the artificial intelligence while Jason Les lagged behind. Li earned the strongest lead; he led Claudico by $529,033. Polk came in next and led by $213,671. Kim led Claudico by $80,482.
However, the participants bet $170 million during the contest. The humans’ monetary lead added up to less than half a percentage point of that total.
That would make it a statistical tie. But Polk tweeted that although Claudico was a “strong opponent” he did not want “this misrepresented as a tie.”
Polk said that Claudico had great problem solving skills throughout the competition. Nevertheless, he made strange decisions that humans would be reluctant to make. He said that the AI would wager more money than humans would usually be willing to wager. “That’s a computer thing,” he said.
Toumas Sandholm, a Carnegie Mellon University computer science professor and Claudico development director, called the AI the “strongest computer poker program in the world.” Although its strength was obvious, Sandholm was not sure how the computer would perform against four Top 10 poker players.
He explained that this competition was a great way to test the AI. He said that poker is great for the machine because it is an incomplete information game. In other words, players could not mislead each other in this situation by bluffing or using other tactics.
Sandholm said that their goal was not to beat humans. “What we want to do is create an artificial intelligence that can help humans negotiate or make decisions in situations where they can’t know all of the facts,” he said.
