Collusion Online

24th Jun 2000

I am not taking a position on the integrity of online poker games; I just want to discuss the logic governing online collusion.

Colluding online is trivial. You can make free long distance calls via the Internet and your computer.

"If" people were colluding online, what would we expect to see? They can't do too much second best hand mucking, as it would be too easy to catch by the site. However, they could ram-n-jam when one of them has the nuts or close to it. So, you should see a lot of jamming and then folding, and a fair number of bad beats with nonsensical holdings.

For example, when one of a pair of partners has AA, the other could raise under the gun with 62. Most people think this has negative EV, because they reason that the partners are paying two to one, but this is not the case in a loose game. The AA has about 50% equity in a loose game, so even if the 62 had 0% equity, they are making a ton of money capping it preflop against several opponents. Additionally, the guy who raised preflop with 62 may get some additional profit on later hands, when people misplay against him based on their memory of this event. The next time the guy raises under the gun, he may get extra action, and he may even have 62 again, and he'll just love it, because his partner has AA again.

Now, of course what sometimes happens here is that the flop comes A94, and so the 62 and AA cap it again, getting action because they seemed to be crazy on previous hands. The turn comes a 3 and they jam, the river is a 5, and the 62 wins with the second nuts. A lot of hands go down like that, apparently, so they don't raise any big red flags. The EV for this type of collusion is much bigger than what Ed thinks, I think.
Some of the victims might jump up and down after being bad beat repeatedly, accusing the online site of some sort of problem with their random number generator or site security. Others dismiss the complaints based on the apparent fishiness of the players. However, the apparently psychic plays need not require psychic powers, fishiness, or hacks of the data/servers. It is merely exactly what you would expect if there were collusion.

The level of collusion could be at an insignificant level, making healthy profits possible, but the level of collusion could quickly increase in a short time. If that were to occur, you might see several good players dump back their winnings all at once.

You might also see this, however, as a result of either variance or sloppiness. The standard deviation of two tables of $10-$20 online is about $350 per hour, compared to about $150 per hour for a tight real world game. (The $350 per hour s.d. comes from my back of the envelope extrapolations from real world loose game s.d., and also from a friend's spreadsheet of his results using Mason's formula.) So, you're going to have swings of $700 in an hour. I guarantee that a lot of the balking is due to this high standard deviation, which comes both from the increased hands per hour and the guys jamming, for whatever reason, with 62. As for sloppiness, I heard one guy saying how his girlfriend plays two games, watches TV, and talks on the phone all at the same time. The sloppiness encouraged by early success could creep in and turn a pro into a fish without his realizing it, conceivably.

So, again, I don't have any position here. I just wanted to point out that if people were colluding then we would expect to see this higher than normal bad beat incidence, and also that the profitability of the games could drastically change in a short time according to the number of colluders. However, just because there is this high level of observed bad beats does not mean that people are colluding, as there could be other reasons for it, with the most obvious alternate explanation being just plain fishiness.

By the way, I agree with Ed that if the online rooms are going to charge $3 per hand, the players should insist the online room have several experienced poker players hired full-time as game "police."

Abdul Jalib

This site uses benevolent cookies.