TWO YEARS AGO, a poker-playing robot called Libratus whooped professional card sharks at their own game, winning $1.7m in the process. What a robot would spend $1.7m on was never answered, as the chips were imaginary, but the message was clear: no game is safe, humans. If you were looking for any form of consolation,
TWO YEARS AGO, a poker-playing robot called Libratus whooped professional card sharks at their own game, winning $1.7m in the process. What a robot would spend $1.7m on was never answered, as the chips were imaginary, but the message was clear: no game is safe, humans.
If you were looking for any form of consolation, it was that Libratus could really only deal with one player at a time, and one-on-one poker is very different from taking on a whole table. Well, now Libratus’s follow up – Pluribus – has taken away that crumb of comfort, something the researchers call a “recognised AI milestone.”
The research is a joint project between Facebook and Carnegie Mellon, and the bot was upgraded for full multiplayer poker. For a start, Pluribus has an online search algorithm to look for ahead for options. It also has “faster self-play algorithms for games with hidden information.”
This is important, because poker is all about the bluffs, which presents a unique challenge for bots: do it too often and your opponents will be wise to you, even if your poker face is gratifyingly blank.
And it worked.
“If each chip was worth a dollar, Pluribus would have won an average of about $5 per hand and would have made about $1,000/hour playing against five human players,” Facebook wrote. “These results are considered a decisive margin of victory by poker professionals.”
Its human opponents were grudgingly impressed, too. “It was incredibly fascinating getting to play against the poker bot and seeing some of the strategies it chose,” Michael Gagliano said. “There were several plays that humans simply are not making at all, especially relating to its bet sizing.”
Chris Fergoson was equally impressed. “Pluribus is a very hard opponent to play against. It’s really hard to pin him down on any kind of hand. He’s also very good at making thin value bets on the river. He’s very good at extracting value out of his good hands.”
“He,” indeed. Well done Pluribus: you are clearly now just one of the guys.
While all of this is very impressive, bluffing aside, it does largely involve probability and the ability to remember cards – something artificial intelligence has a natural advantage in. If you’re ever in a Bill and Ted’s Bogus Journey-style situation where you need to beat a bot at a game to stay alive, we’d recommend football:
Just try to keep the score to single-figures. µ
Image: Morgan, used under Creative Commons