@op have you also reviewed the hands that the suspected cheaters play against each other? are they simulating normal poker and bluff / bluffcatch each other? If they are smart and your theory is correct they should play regular poker and only take advantage of the known hole cards for All ins (since they are probably just sharing winnings afterwards).
Also interesting if they are collude in other ways like trapping players in sandwich-position and all that stuff that you could do when involved with others in a pot.

That's probably not a bad idea as it shouldn't be that hard to look into. My suggestion for the soft playing would be to look at vs player stats for suspicious accounts and then have something like river hand value = nuts (im just making these filters up) and then river action = call bet. It would be a really obvious way of soft playing, but makes sense to start there. Maybe even just go wider and look at every hand where a suspicious player had the nuts on the river vs a suspicious account.

It'd also be interesting to see if the people you suspect are cold calling 3bets an inordinate amount of the time. That was one of the things I found peculiar was how many 3 and 4 way 3bet pots I was getting into, then never seeing a showdown.

Lots of sns. I try to change every month so I'm sure you hate a few of them.

I looked at the evidence and it looks pretty damning imo that these 5 accounts at 50-100 plo are cheating. I won't be playing in 6max games on ipoker until they take a more proactive stance on collusion.

Additionally Willy was kind enough to offer analyzing 200 and 400 plo hands if someone else buys the hands. I figured some ppl wouldn't mind chipping in $5-10. If you're still playing in those games then you'll want to know who the cheaters are. Any trusted regs want to volunteer to buy hands, collect $, and send to Willy?

I'd ship some $ for 2/4 to 5/10 hands

I would just like to comment on this so that people aren't mislead. The main argument based on the data (in my opinion) is for automated play (i.e. bots or maybe a poker snowie like program that is operated by a human).

Once you have reached this conclusion, then you can make arguments for collusion based on the connectedness of the accounts in question.
However straight up collusion without the assumption of connectedness of certain accounts is very hard to prove due to all of the noise in the data. So unless you think you are playing against accounts that are using automated play, then I wouldn't assume anything has changed in your games. And keep in mind that a certain amount of cheating is always to be expected when dealing with online poker. It's just a cost of playing the game online.

Maybe a stupid question, but what is a snowie program?

really sick thread. i got beat up pretty good in ipoker ev too

COLLUDER!!!!

This is all true and it's pretty much my only possible concern with OP's analysis. The fact that he plays in these games and obviously has "run bad" against certain accounts makes him and his selection of accounts to analyze potentially biased.

But still, the evidence seems to be pretty overwhelming. OP sent me the hands to look at, though I've yet to start looking at them and am still considering what the best approach might be.

I was referring to Poker Snowie, the artifical intelligence NL bot/learning software that gives you a recommendation for every game state. If somebody developed a program like that for PLO but chose not to publicize it, then that might explain certain styles as well.

I did run bad on these games for sure but I didn't feel that I ran bad against these certain accounts.

What made me suspicious was several accounts with almost identical pre-flop stats and their simultaneous&extensive adaptation to my game.

iPoker is probably one of the most ridiculous networks to play at. I have 5 different screen names on there that I can think of (from various different sports betting accounts). Over the years, I may have even more than that. I could probably start a new screen name right now just by signing up somewhere I've never been before and exploiting their sign up bonus. Tbh, I'm surprised a thread like this hasn't started previously.

cts, you might be better off showing this in a stakes break down? It seems the problem is biggest at certain, higher stakes.

I probably worded that poorly - I don't think you've selected the accounts in a particularly biased way, but there's potential for some small bias, which is almost certainly small enough to be irrelevant with a large enough sample.

That said, we should try to do this analysis in a way that it can't be criticized for being potentially biased.

We reported accounts we suspected the most to ipoker with the relevant data so I think there's not much more we can do right now. The thread on small stakes PLO is more detailed.

woops, didn't realize there's two separate threads. gonna go read the other one now

http://forumserver.twoplustwo.com/38...-ring-1481094/

Filtered for all-in on the flop in ring games:



How unlikely is a run like that?

CTS, for filtered all in spots I'm not sure what the standard deviation is but if we take:
-500bb/100 the probability to run at or bellow is around 2%
-1000bb/100 the probability to run at or bellow is less than 10%
-2000bb/100 the probability to run at or bellow is less around 25%

std deviation for those hands is 780bb/100

Then it's around 4.5% with a 780bb/100 std deviation.

ty guimz, so a bad but not horribly improbable run

Now the colluding PLO bots are playing HSPLO vs cts too. Impressive programming.

Definitely still within the "safe zone" statistically. I had a 3.5% bad run last month on ipoker. Once you drop below around approximately 2.25% is when things start to become interesting/worrisome.