Unstoppable Poker Bots?

Internet security guru Ed Felten takes note of a scary story on MSNBC that warns us that <fnord!> sophisticated card-playing robots – known as “bots” in the nomenclature of the Web – are being used on commercial gambling sites to fleece newcomers, the strategy-impaired and maybe even above-average players. </fnord!>

Felten himself looks askance at efforts of online poker sites to deter bot playing as well as collusion:

The online casinos are kidding themselves if they think they can enforce a no-bots rule. How can they tell what a player is doing in the privacy of his own home? Even if they can tell that a human's hands are on the keyboard, how can they tell whether that human is getting advice from a bot?

The article discusses yet another unenforceable rule of online poker: the ban on collusion between players. If two or more players simply show each other their cards, they gain an advantage over the others at the table. There's no way for an online casino to prevent players from conducting back-channel communications, so a ban on collusion is impossible to enforce.

By reiterating their anti-bot and anti-collusion rules, and by claiming to have mysterious enforcement mechanisms, online casinos may be able to stem the tide of cheating for a while. But eventually, bots and collusion will become the norm, and lone human players will be driven out of all but the lowest stakes games.

But there is another strategy. An online casino could encourage bots, and even set up bots-only games. The game would then become not a human vs. human card game but a human vs. human battle between bot designers for geekly mastery. I'll bet there are plenty of programmers out there who would like to give it a try.

The MSNBC article, once you get past the menacing shadows and ominous organ chords, presents a pretty good view of the state of the art of poker bots: the cutting edge of poker bots can play really well against a single player, but are much more at sea in full games with a constantly changing cast of characters. A reasonably skilled programmer can write a bot that could beat a soft game (I have fond memories of playing against r00lbot on the old irc.poker.net). It's a major challenge to produce a bot that can hold its own against skilled players, as Darse Billings and others at the University of Alberta can attest.

Collusion, on the other hand, is detectable over time because it has a definite signature. Colluders with knowledge about each others' cards can exploit edges that would appear to be marginal or negative-EV to someone without that knowledge. (Think of folding a flush draw because of knowing that two of your nine outs are gone and the pot is offering the wrong price, when an ignorant player would call thinking the odds were right). Colluders can also exploit occasional large-edge situations, e.g. when a hand that might not be the absolute nuts is known to be the nuts because of cards that are out of play.

Detection of colluders becomes a problem similar in kind to traffic analysis in signals intelligence. Effective colluders have to take advantage of their edge in such a way as to not reveal it, which means not taking advantage of it some (presumably randomized) part of the time. But that reduces the edge from collusion. If collusion becomes widespread, then (if the online poker sites are on the ball) collusion is widely detected and widely quashed. If collusion is low-key, then it doesn't get quashed ... and it doesn't take too much money out of the game. Sounds like there's some kind of Nash equilibrium to be reached, given a particular level of online poker server collusion vigilance.

But this doesn't alter the fact that at this point in time the fish are biting, and there are more of them every day. There's enough money for all of us to win, bots, colluders, and the merely skilled and experienced players, right now.

(via boingboing)