Internet sleuths from the 2+2 poker forums have identified a number of players (mostly registered in Russia) who they claim are using bots to profit unfairly. The players in question play Pot-Limit Omaha at the $0.50/$1 and $1/$2 6-Max cash tables and have collective profits of about $1.4 million dollars.
The accusation is based on analysis of the players’ statistics. Not only do all the accused players exhibit preflop behaviour that would be unusual for a human – squeezing far more often than the average regular at those stakes, for instance – but their statistics are uncannily similar to one another. Having created a metric for how similar or dissimilar any two players are from one another, 2+2 member Schwein found that any two human players will usually have a difference value of around 600 to 1200, while the presumed bots all differ from one another by only a few dozen points.
There is also some suspicion that the bot-users are colluding by sharing their hole cards with one another, although the evidence for this is more tenuous. The suspicion stems from the fact that most of them have been running well above expectation over the long run in all-in situations. The reasoning is that if the bots know one another’s folded hole cards, they have better information available to them than non-cheating players when it comes to assessing their likelihood of hitting a draw; if they are therefore tending to fold their draws for which several outs are in the muck, and play them when all the outs are in the deck, this would naturally lead to them hitting those draws more often standard probabilities would predict.
It’s natural to be skeptical of these sorts of claims, because online poker players have always been paranoid about getting cheated, ever since the first sites appeared. Usually, the supposed “proof” of rigged games or cheating opponents proves to be nonsense, but that appears not to be the case this time: Although PokerStars policy is not to disclose very much information about the activities of their Integrity Team, they have confirmed that at least some of the accused players have now been banned. All told, Schwein has identified 26 players he believes are using bots, only 10 of which are still active; thus, PokerStars may have already banned as many as 16 of the members of the ring.
Omaha is supposed to be hard
This is terrible news for online poker for a number of reasons. For one thing, the fact that the bots were found in Pot-Limit Omaha games is worrisome, in that multi-player Pot-Limit games in general and Omaha specifically are supposed to be among the most resistant to artificial intelligence. For instance, heads-up Limit Hold’em has already been solved, but poker AI researcher Eric Jackson told me in our interview a few months ago that even for heads-up play, the best No-Limit bots are still theoretically exploitable by over 300 big blinds per hundred hands, and that multi-way games are even harder.
Now, if the bots are sharing hole cards with one another, it’s possible that they could not beat the regulars at these stakes without that additional advantage, but the bots are by all accounts crushing the games, not just grinding out a small profit. Given how tough the mid-stakes online cash games are these days, if the bots are capable of cleaning up at $1/$2 PLO, then there is no longer any game or stake level which could be considered safely out of the reach of bots.
Doubts about PokerStars’s countermeasures
According to one poster in the thread, some of the banned users were actually first reported to PokerStars security a year ago, but those initial accusations were dismissed after cursory investigation. Meanwhile, given the amount of work that’s been done by Schwein and others and passed on to PokerStars, it’s reasonable to wonder whether many (or any) of these players ever would have been caught by PokerStars without outside assistance.
The lack of transparency at PokerStars is also worrisome, though understandable. Disclosing their methodology publicly would of course give an advantage to those trying to circumvent their countermeasures, so you can’t blame them for being tight-lipped. At the same time, though, it means that players have nothing to reassure them that they’re being looked out for, except for vague promises that “every” effort is being made to ensure the integrity of the game.
The scandal is also reminiscent of another recent controversy, surrounding the use of special software for heads-up Sit-and-Gos by the user “Skier_5” and several of his associates. Just as with the bots, the players using Skier_5’s software have remarkably similar statistics to one another, leading to speculation that the software amounts to the same thing as a bot, just with a human middleman clicking the buttons (which may be the case here too, given that the “bots” have been known to chat, and seem to keep human-like playing schedules). In Skier_5’s case, the software was explicitly declared legal for use by PokerStars, so although it’s reassuring that they’re banning accounts in this case, there are bound to be more questions asked about exactly where and how PokerStars draws the line.
Little compensation and no recourse
Worse, the eventual banning of the guilty parties is of little consolation to their victims. When a cheater is banned, his account balance is seized and the funds divvied up between the players determined by PokerStars to have been disadvantaged by him; the trouble with this is that there are many victims, and usually not much money to go around, as cheaters tend to withdraw frequently precisely because of the danger of being caught. Some posters in the 2+2 thread, for instance, claim to have lost several thousand dollars to one or more of the banned accounts, while only receiving a few tens of dollars in compensation.
Schwein calculates, meanwhile, that PokerStars collected nearly $2.5 million in rake from the probably-banned accounts during the time that they were active; a million more, in other words, than the profit earned by the cheaters themselves. This fact, combined with the insultingly small payouts people have received so far is likely going to amount to a PR disaster for PokerStars, particularly if the scandal continues to grow and bot rings prove to be more prolific than previously thought.
Since the takeover by Amaya, scarcely a month has passed without some new scandal or complaint emerging about PokerStars. At first, the contentious issues had to do with decisions in terms of games on offer, rake and so forth; these, it seemed to me at the time, could be dismissed as players simply not liking change, particularly changes which would force them to adjust to a new game or see a decline in profits. Those are the sorts of complaints you expect to go away as everyone eventually gets used to the new normal. The controversies of the last couple of months – first Skier_5, and now the bots – not so much. If stories like these continue to break on a regular basis, the future for online poker may start to look quite bleak quite quickly.
Alex Weldon (@benefactumgames) is a freelance writer, game designer and semipro poker player from Montreal, Quebec, Canada.