Latest cliffsnotes on Absolute soulreading.

[Redgrape]

How about we ask the major Poker websites, lp.net, ftr.com, ith.com, to educate their readers that they should think twice before depositing money in AP? Yes, they prob have affiliate money with AP, but its not like someone who wants to deposit money in a poker site isnt going to deposit at all if one site is bad, they'll just find another one, so the affiliate money is just going to come from elsewhere.

[knappis]

[ QUOTE ]
thats nice, but i bet you blow at poker, because you don't seem to understand the fundamental reasons for our suspicion

[/ QUOTE ]

I'm not sure about you but I have read several others that state their fundamental reason for suspiscion is based on stats and probabilities.

[ikestoys]

[ QUOTE ]
[ QUOTE ]
thats nice, but i bet you blow at poker, because you don't seem to understand the fundamental reasons for our suspicion

[/ QUOTE ]

I'm not sure about you but I have read several others that state their fundamental reason for suspiscion is based on stats and probabilities.

[/ QUOTE ]

yeah, it is, but you haven't calculated the probability of our fundamental suspicion, which isn't that he folded everytime someone else had a big pair (he didn't) but that he started wit 93% of his hands but never called a river beat.

[McSeafield]

Most of the poker players (my estimate is approx. 95% of all online players) out there have not the time to read or cannot understand the various aspects in these threads about this subject. For most of the remainig ones it is much to academic to unterstand correlation coefficients, testing null hypothesis with significance tests, whether they are based on normal distribution, t-distributon or priori distribution, does not matter. Experts however know also the various examples where correlated and at the same time uncorrelated random figures exist and what the co-variance is such examples really mean. Since nobody knows what the purpose of such academic discussion really is, I think, we should go back to a more fact orientated discussion.

The facts according of my view are as easy as follows:

We have overwhelming evidence that cheating happend at Absolute Poker.

A lot of other assumptions are still speculation, mainly:

- did the cheater use some kind of super-admin account or another technic?
- are the five player-accounts used by the same person?
- is Mark Seif or another well-known person involved in this affair?
- why has been this case not detected by Absolute?
- is also Absolute management or only IT-personal involved?
- are the players in question Absolute shills?

Only Absolute or an external auditor can check the software and clear up all facts of this case. As long as Absolute does not contribute and support the clearing up I have the following question. Assume that cheating happened to the extent we already know, how high is according to your life experience (not priori distribution) the likelyhood that Absolute is cheating or allows and protects cheating? Consider this as a case where stats experts cannot reject a certain null hypothesis with a security likelyhood of 99%, where they want to make another test, but where the burden of proof has already Absolute and nobody else.

And as long as Absolute is not willing to proof what really happened, we should stay together within a global cooperation.

[ QUOTE ]
4. Make a website, and target it to clearly and logically lay out the proof against absolute poker.
.
Try and target it so that it is visible to people interested in Absolute Poker, but is clear that other sites are not under suspicion.

[/ QUOTE ]
5. To inform more poker players arround the world about Josem`s webside.

I think about links in various forums and an identical flyer which can be distributed during live tournaments or other poker related events. One example is the Full Tilt Poker One Million Challenge in Germany. There are still four major events (in Hamburg, Stuttgart, Berlin and Munich) with more than 1000 poker players during each event as an example. I noted yesterday in Cologne that more or less nobody was informed about the Absolute cheating case. I think the same applies to other countries or events. We should make any effort to change this. I would suggest that Josem should post a uniform printable text which can be used as a flyer and as a post, which he wants to be distributed as far as possible in all countries arround the whole world.

[knappis]

[ QUOTE ]
yeah, it is, but you haven't calculated the probability of our fundamental suspicion, which isn't that he folded everytime someone else had a big pair (he didn't) but that he started wit 93% of his hands but never called a river beat.

[/ QUOTE ]

Yeah, thats a tough one. Modelling post flop play is far more difficult than preflop play. If we had a very large sample we could assume an even distribution of cards for a random board and hole cards. I'm not so sure this is appropriate for the small number of hands we have here and even if it was, the confidence intervals would be extreme.

The easisest to model is what I did above, and as you can see, even those estimates aren´t exact and rely on assumptions that we don´t know are true.

The second most easy thing to model is the probability of having a high winrate. It has been reported that this guy had a SD of 270ptbb/100 which too me is insane and I am not to sure we can trust these numbers. However, If these numbers are correct and we assume a losing player with -50ptbb/100 it would mean that 5% of all 100 hand sequenses would be outside -50+-2*270PTBB/100. That is, 2.5% of the 100 hand sequenses would show +490PTBB/100 or more and another 2.5% would show a loss of 590PTBB/100. Both would occur in average every 4000 hands. I imagine that the extreme winning sessions would look like he could see hole cards and the extreme losing sessions would look like chip dumping.

Anyway, a SD of 270PTBB/100 sounds really extreme too me and I'm not too sure that figure is beleivable even for a maniac like this.

[ibluffoldladies]

Any new developments besides this thread being destroyed by the three guys from x-files?

[You're No Daisy]

I plan on getting Josem my HH tonight after I get home from work and put the kids to bed. Anyway...are there any formally schooled statisticians on here that think a chi-square goodness of fit test would be a good starting point to determine whether or not we can really dive deeper into whether there was cheating or not? Believe me, I do think these fools cheated. However, I think a chi-square goodness of fit test would validate our observed values of the cheaters (as seen in PT) vs. what our expected values are (for winning players). Our sample size for winning players is certainly large enough, and I imagine we have a similar sample of the losing players.

I wasn't a stats major, but I remember enjoying my stats class when I was in my masters program. In fact, one of the things we talked about was a casino using the chi-square goodness of fit test to determine whether the dice a gambler was using was fair. I have a link to this exact scenario Chi-square

I'm not sure if this helps, but I'd really like to catch these fools if possible.

AC

[Marnixvdb]

[ QUOTE ]


Anyway, a SD of 270PTBB/100 sounds really extreme too me and I'm not too sure that figure is beleivable even for a maniac like this.

[/ QUOTE ]

that number seems to have fallen from the air. I have no clue where the SD of 270 ptnn/100 comes from, and i doubt it shows from the samples.

[darkcore]

yeah. it pains me to see that this thread is only about math now. and to me as a mathnoob some of the assumptions used for modelling the case seem strange/incorrect. but as most posters here i don't have a f'ing clue about statistics. so i suggest for those people who want to talk statistics

1. go to http://forumserver.twoplustwo.com/po...rd=probability
2. discuss all the statistical aspects in a seperate thread there with people who are skilled (and interessted) in math
3. find a solution that is a consenus
4. then come back here.
5. ????
6. profit

the rest should focus on other -less academic- things that can/should be discussed and done about this issue here...

[knappis]

[ QUOTE ]
yeah. it pains me to see that this thread is only about math now. and to me as a mathnoob some of the assumptions used for modelling the case seem strange/incorrect. but as most posters here i don't have a f'ing clue about statistics.

[/ QUOTE ]

Well, I think it is about time we discuss math since it is the basis of the whole cheat theory in the first place. I can understand that everybody is so sure there was cheating going on if the odds for these things to happen by chance was in the 1:billion range. But closer examination of the evidence suggests some of the most compelling evidence of a cheat is much much more likely to be due to variance than has been suggested before.

I have skimmed through the threads to find solid evidence for some of the more popular claims made about odds. And I think I found the source of the 10^20:1 estimate:

[ QUOTE ]
I have a masters in statistics and an off the cuff estimate I would give would be something like 10^20:1

[/ QUOTE ]

That evidence is pretty thin with my standards.