Poker Bots: Come With Me if You Want to Live

by , Feb 20, 2011 | 7:32 am

tim chilcote poker bots

Tim Chilcote


Machines have always been the enemy of man, at least in movies and on television, yet somehow we never see our A.I. overlords coming until it’s too late.

Case in point, in Terminator 2: Judgment Day, Doctor Miles Bennett Dyson spends a lifetime developing artificial intelligence as part of the Skynet project, only to discover that his work is more suitable for evil than for good – the price? His life. You remember the scene: Dyson detonates his own lab, and in doing so blows himself up, sacrificing himself to save us from an army of Schwarzeneggers.

With the computer Watson now a winning contestant on Jeopardy!, the man v. machine debate has been rekindled, and it would seem that we’re in danger again, if not for our lives (yet), then for pride. In a recent Slate article, “Jeopardy, Schmeopardy: Why IBM’s next target should be a machine that plays poker“, author Chris Wilson asks whether the next logical progression from Jeopardy!’s Watson is a poker playing robot, and suggests that robots have a lot to teach us about poker, and might even be — gasp! — better.

Bots sound dangerous, and it would be easy to infer that their skill is only going to grow and that their dominance of the poker world is a forgone conclusion.

Poker bots have been hot topic in online poker for years. The fervor usually stems from a fear of the unknown. Gambling robots reached a fever pitch in 2007 when Phil Laak and Ali Eslami played against the Polaris poker bot. At the time I asked myself the obvious question – is Polaris the next terminator, and if not, then what’s the point of this experiment and why should I care? I sat down that summer with computer-poker researcher Darse Billings of the University of Alberta Computer Poker Research Group to have him explain how Polaris actually worked and to see what he thought was the point of his poker robot experiment, just to put my mind at ease.

I first asked Billings if he could kindly describe for the layperson like myself how Polaris operates. “No,” he said, and that was that. I simply wouldn’t get it. Fair enough. Next I asked if I couldn’t outsmart Polaris by playing really loose for the first third of the match and then really tight, and finally sporadic. Again, the answer was a simple and convincing, “no.” Apparently I just didn’t understand robots. Polaris, Billings explained, cannot get stuck in one particular mode of thinking. The bot is equipped with “memory decay,” which means that after a set of 20 hands it completely forgets it was ever playing to begin with and restarts its thinking process. Polaris bluffs at random, switches playing styles, and is not even one robot, but an “umbrella” bot where control is rotated to several different bots over the course of the match.

One of Billings’ bots, Mr. Orange, is programmed to interpret pots as being larger than they actually are, and to consequently over-bet, creating ever-growing pots that the bot fights harder and harder to win, putting significant pressure on a human or robot opponent. Mr. Orange’s aggressive play was so successful that Billings’ team nicknamed it Agent Orange for the bot’s ability to kill its opponents. Interestingly, Billings pointed out that Mr. Orange’s success proved what good poker players already know — aggression is key, and the only mistake a player can make against an aggressive heads-up opponent like Mr. Orange is being too tight. “Folding too much is the only thing dangerous,” Billings said, and went on to say that any two cards are worth taking to the flop. “Two-handed, luck is bigger than anyone thinks.”

Bots sound dangerous, and it would be easy to infer that their skill is only going to grow and that their dominance of the poker world is a forgone conclusion. But for online players who fear losing their bankroll to a robot opponent, Billings wholeheartedly dismissed their concern as ludicrous. “The amount of time and knowledge that would go into making even a mediocre bot,” he said, “makes them nearly impossible to run online.”

Billings also offered that bots are completely beatable, even suggesting that players should welcome bots. Part of Billings’ initial research for the Polaris project was to build a rock-paper-scissors bot that would consistently break even. The Polaris poker bot was an extension of this rock-paper-scissors design, and the bot plays poker exclusively to break even, only profiting when human players succumb to their own weaknesses. If you’re still worried, it’s worth mentioning that, despite fears, bots are not out fishing in multi-table tournaments; they’re designed to play heads-up limit games. So a strong limit player could, in theory, profit by playing bots.

Where then does this irrational fear of bots come from? Well, perhaps it’s not fear at all, but a display of how social a game poker is, even online. People want to play other people. “The bot,” Billings said, “doesn’t even know it’s playing poker.” Now what fun is that? Ego is such a huge part of poker that, not only are players obsessed with winning, but if the victory doesn’t somehow boost a player’s ego, the thrill is lost. In fact, although Billings considers himself a fine poker player, and is even the test subject against his bots, he is not interested in poker, at least not in poker against bots. Billings’ interest is in writing mathematical equations to deal with the imperfect information available at the poker table. He hopes his research might someday translate to other arenas of artificial intelligence.

Bots, with their play-for-the-tie philosophy, sound to me like the poker equivalent of soccer players, not futuristic no-limit killing machines. Billings said in 2007 that a no-limit bot was in the works and would be shockingly good, though it was “far away.” I’m guessing four years isn’t “far,” but maybe I’m wrong. Maybe 2011 is the year of the robot uprising, and Watson’s success on Jeopardy! is the first sign of apocalypse. Billings did admit that eventually the computers would win because computers “maximize what computers do well.” Now that sounded like a poker lesson for the current moment – because computers have no ego and often don’t even realize they’re playing poker, they are able to “focus” on what they do well, which is exploiting the weakness of other players.

The lesson: Don’t worry too much about bots, at least for now, but don’t get overconfident either – someday they may take over the world, so stay on their good side. Billings says his bots have “no fear, no shame, and an unlimited imagination.” Best advice is to stay ahead of the curve and allow for variance in your game, otherwise you might get stuck in a rut – and if a shifting mercury avatar asks in the chat box, “Have you seen this boy?” – fold.

Tim Chilcote blogs at Great Lakes Guru and is Managing Editor of BULL: Men’s Fiction. Follow him on Twitter @TimChilcote.

Comments are closed.