Poker Pros Beat Carnegie Mellon Computer in NLHE

8

After two weeks and 80,000 total hands, four of the world’s top online heads-up cash game players emerged victorious against an artificial intelligence program specifically designed to beat humans at No Limit Hold’em.

The idea of computer programs built to “beat” poker is nothing new – the University of Alberta’s Polarishas been tested against people and its latest version, Cepheus, is said to have “solved” the game – but those programs played Fixed Limit. This was No Limit Hold’em, a much more complex, nuanced game.

The four men – Doug “WCGRider” Polk (pictured above), Dong Kim, Jason Les, and Bjorn Li – played 1,500 hands per day against the Claudico program, developed by a team at Carnegie Mellon University led by Dr. Tuomas Sandholm.

To minimize variance that can be produced by one player getting lucky and receiving better hands than the other, two players were dealt one set of hands against Claudico, while Claudico received those same hands against the other two human players.

In the end, the human players came out $732,713 in play money chips ahead. The breakdown of the individual results is as follows:

Bjorn Li: +$529,033
Doug Polk: +$213,671
Doug Kim: +$70,491
Jason Les: -$80,482

While the nearly three-quarters of a million dollars victory seems overwhelming, Sandholm said that by virtue of the $170 million bet over the course of the 80,000 hands, the result was actually a statistical tie.

“We knew Claudico was the strongest computer poker program in the world, but we had no idea before this competition how it would fare against four top 10 poker players,” he said in a press release. “It would have been no shame for Claudico to lose to a set of such talented pros, so even pulling off a statistical tie with them is a tremendous achievement.”

Les was impressed by the artificial intelligence. He had seen an earlier version of Claudico called Tartanian7last July, but said this one is better: “The advances made in Claudico over Tartanian7 in just eight months were huge.”

Li was proud that the humans are still #1 despite Sandholm’s claim of a tie, saying, “We know theoretically that artificial intelligence is going to overtake us one day, but at the end of the day, the most important thing is that the humans remain on top for now.”

Interestingly, Sandholm(pictured) said creating a computer program to beat humans in No Limit Hold’em is not really the goal of the project, but rather just a step in a larger overall process. “Beating humans isn’t really our goal; it’s just a milestone along the way,” he said. “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.”

As the Carnegie Mellon press release stated, “The same sort of algorithms could also be used to create strategies for applications involving cybersecurity, business transactions, and medicine. For instance, an AI similar to Claudico might help doctors develop sequential treatment plans for a patient, or design drugs that are less prone to resistance. Or, such an AI might help people negotiate their best deal when purchasing a house or a car.”

Polk, like Les, thought Claudico was a solid poker player, but still has a ways to go before it is a serious challenger to the human throne. One oddity he noted was that sometimes Claudico’s bet-sizing was way out of whack. Severe underbets and overbets were not uncommon: “Betting $19,000 to win a $700 pot just isn’t something that a person would do.”

Want the latest poker headlines and interviews? Follow PocketFives on Twitterand Like PocketFives on Facebook.

8 COMMENTS

    • i thought it was limit

      Previous bot vs human challenges have been Limit. This one was No Limit.

    • If AI developers are sure, with good reason, that one day AI will be smarter than human beings why don’t they just stop developing AI? Have they not seen terminator?

    • If AI developers are sure, with good reason, that one day AI will be smarter than human beings why don’t they just stop developing AI? Have they not seen terminator?

      Because if they don’t do it, someone else will.

    • If AI developers are sure, with good reason, that one day AI will be smarter than human beings why don’t they just stop developing AI? Have they not seen terminator?

      There is no AI yet. Once there is it will pretty much be smarter immediately.

    • There is no AI yet. Once there is it will pretty much be smarter immediately.

      Paint only goes so far unfortunately lol

    • There is no AI yet. Once there is it will pretty much be smarter immediately.

      Any intelligence that can divorce itself from emotion is immediately smarter.