Computers are now better at poker than any humans

As Kenny Rogers once sang, you've got to know when to fold 'em - and the time might be now.

Computers are now better at Texas hold'em poker than the world's top players.

An artificial intelligence (AI) program developed by researchers at Carnegie Mellon University and Facebook has emerged victorious against Darren Elias, the record-holder for most World Poker Tour titles, and Chris 'Jesus' Ferguson who's won the World Series of Poker six times.

Pluribus, as they called it, was also able to routinely beat five pro players at a time. 

"Pluribus achieved superhuman performance at multi-player poker, which is a recognised milestone in artificial intelligence and in game theory that has been open for decades," said Facebook computer scientist Tuomas Sandholm.

"Thus far, superhuman AI milestones in strategic reasoning have been limited to two-party competition. The ability to beat five other players in such a complicated game opens up new opportunities to use AI to solve a wide variety of real-world problems."

He said different strategies are needed to win two- and six-player games, and the moves Pluribus came up with might even inspire real-life players to try different things.

"For instance, most human players avoid 'donk betting' - that is, ending one round with a call but then starting the next round with a bet," the researchers said in a statement.

"It's seen as a weak move that usually doesn't make strategic sense. But Pluribus placed donk bets far more often than the professionals it defeated."

Michael Gagliano has won more than US$2 million playing professional poker. He couldn't believe some of the moves the computer pulled. 

"It was incredibly fascinating getting to play against the poker bot and seeing some of the strategies it chose," he said. "There were several plays that humans simply are not making at all, especially relating to its bet sizing."

The researchers said the computer's secret is its ability to avoid mistakes.

"Although the AI's strategy guarantees only a result no worse than a tie, the AI emerges victorious if its opponent makes miscalculations and can't maintain the equilibrium."

But that only works against one opponent - to defeat five at a time, Pluribus had to develop new ways of playing - then mix up those strategies in order not to appear predictable.

Perhaps it's time for humans to walk away - or with AI approaching Terminator levels of ability, perhaps even run.

The research was published this week in journal Science.



Contact Newshub with your story tips: