General Poker Discussion Poker Forums

http://www.evplusplus.com/site_media/images/variance/wr_6.00_sd_96.00_nruns_100_nhands_100000_False_g7trcE.png


http://www.evplusplus.com/site_media/images/variance/wr_6.00_sd_96.00_nruns_100_nhands_100000_False_g7trcE_downswings.png

those represent a 3ptbb/100 winner with SD of 48ptbb/100. over 100 100k hand trials, the largest downswings were somewhere around 40-45 buyins. if you switch that to 100 1MM hand samples, you have somewhere around a 30% chance of having a 40BI downswing with that same winrate/sd. switch it to 100 5MM hand samples and every single player had at least one 30 buyin downswing.

variance is a bitch.

This is an interesting point. But it is not the same as AI EV.

This is an interesting point. But it is not the same as AI EV.

and all in ev doesn't take into consideration about a hundred other ways to run good/bad.

and all in ev doesn't take into consideration about a hundred other ways to run good/bad.

I am not talking about running good or bad. I am talking about the randomness of the cards. It is an impossible task to analyze it any other way because you never know what everyone was holding on every hand and there are different skill levels to take into account. In an all in situation, you have complete knowledge and there is no skill, only luck or randomness.

I am not talking about running good or bad. I am talking about the randomness of the cards. It is an impossible task to analyze it any other way because you never know what everyone was holding on every hand and there are different skill levels to take into account. In an all in situation, you have complete knowledge and there is no skill, only luck or randomness.

ok sure, i just find it hard to believe that there is any other explanation other than run bad. ftp has dealt out billions of hands of poker, probably trillions, maybe even more; if there is some sort of anomaly someone would have found it by now, esp with the advent of sites like PTR.

besides, stars at the very least has independent commisions that verify the randomness of the software.

http://www.pokerstars.com/poker/rng/

what do you know, so does FTP:

http://www.fulltiltpoker.com/random-number-generator

so unless you believe they are in conspiracy with each other, cigital, and the other auditors, AND that over trillions of hands no one has yet done any of this sort of analysis, there's pretty much no way that the RNG's are rigged.

I am particularly interested in two hands on FullTilt. AKo(45%) vs. and underpair (55%) all in preflop. Everytime I get AKo, I shove preflop, if I get called by an underpair, then I have a data point. Either a win or a loss. I use PokerTracker to keep them. You analze the 'variance of the variance' by the binomial distribution http://www.stat.tamu.edu/~west/applets/binomialdemo.html

I only have 13 samples so far, 2-11, but that is 3 standard deviations away from the norm. I really need more samples than 13 but if it gets to 5 standard deviations away from the norm, then we can feel confident that the winrate would be lower than 45% in this case.


The other hand I will tell you about at another time.

in fact here is cigital's analysis of 100M pokerstars hands, see for yourself:

http://www.scribd.com/doc/13710664/Cigital-100M-Hand-Analysis-Report

in fact here is cigital's analysis of 100M pokerstars hands, see for yourself:

http://www.scribd.com/doc/13710664/Cigital-100M-Hand-Analysis-Report

I play on Poker Stars and I love it. I trust them. It has been in Microstakes FullTilt.com where things haven't seemed right. That is why I wanted to talk to that guy who made this post, he said that he played microstakes, fulltilt and that he was a rakeback member.

I play on Poker Stars and I love it. I trust them. It has been in Microstakes FullTilt.com where things haven't seemed right. That is why I wanted to talk to that guy who made this post, he said that he played microstakes, fulltilt and that he was a rakeback member.

well the same company independently audited both sites and found nothing wrong, and furthermore i have provided ample evidence that its really easy to run really bad. if you still think the sites might be rigged, have fun investigating further. if you find any anomalies over a statistically significant sample i'll be pretty shocked.

I only have 13 samples so far, 2-11, but that is 3 standard deviations away from the norm.

I hope you realize that 13 samples is minuscule. You should really go read the 2P2 post I linked to above - there are hundreds of pages of people who have made similar claims, and tons of good responses as to why they are statistically insignificant. I believe there are even a few analyses of AK vs underpairs in there. In short, no credible data has ever been presented that the sites are rigged. None. Despite hundreds, if not thousands, of unsubstantiated claims. There was an audit done by a group called the ACPW that is laughable, biased and easily debunked.

If you are able to gather a statistically significant sample of data that shows problems with the RNG, people will be interested. But no one has done that to date despite many many many unsubstantiated claims.

I hope you realize that 13 samples is minuscule. You should really go read the 2P2 post I linked to above - there are hundreds of pages of people who have made similar claims, and tons of good responses as to why they are statistically insignificant. I believe there are even a few analyses of AK vs underpairs in there. In short, no credible data has ever been presented that the sites are rigged. None. Despite hundreds, if not thousands, of unsubstantiated claims. There was an audit done by a group called the ACPW that is laughable, biased and easily debunked.

If you are able to gather a significantly significant sample of data that shows problems with the RNG, people will be interested. But no one has done that to date despite many many many unsubstantiated claims.

I wouldn't say none. Potripper was proven to be cheating.

well the same company independently audited both sites and found nothing wrong, and furthermore i have provided ample evidence that its really easy to run really bad. if you still think the sites might be rigged, have fun investigating further. if you find any anomalies over a statistically significant sample i'll be pretty shocked.

I read the paper. That paper is a piece to show that poker is a game of skill, not luck. It doesn't say anything about the randomness of the cards. It says that in 74% of the hands, they don't even go to showdown, so it doesn't matter who has the better cards most of the time.

I hope you realize that 13 samples is minuscule. You should really go read the 2P2 post I linked to above - there are hundreds of pages of people who have made similar claims, and tons of good responses as to why they are statistically insignificant. I believe there are even a few analyses of AK vs underpairs in there. In short, no credible data has ever been presented that the sites are rigged. None. Despite hundreds, if not thousands, of unsubstantiated claims. There was an audit done by a group called the ACPW that is laughable, biased and easily debunked.

If you are able to gather a significantly significant sample of data that shows problems with the RNG, people will be interested. But no one has done that to date despite many many many unsubstantiated claims.

I agree that the sample size is too small to prove anything. But three SD's is 3 SD's. It is commonly accepted that 5 SD's does not happen in a normal distribution. I will let you know after I get some more data points.

I wouldn't say none. Potripper was proven to be cheating.

Completely different. You are talking about the RNG and the way cards fall in all-in situations. AP/UB was a superuser account. Totally unrelated.

EDIT: You're just going to say that this proves that poker sites are capable of shadiness. I won't deny this. I am done with this thread and will again recommend you review the 2p2 thread where this has been discussed approximately 2 million times. Good luck - last response from me.

I read the paper. That paper is a piece to show that poker is a game of skill, not luck. It doesn't say anything about the randomness of the cards. It says that in 74% of the hands, they don't even go to showdown, so it doesn't matter who has the better cards most of the time.

yeah fair enough, there's still 0.0001% chance in my mind that the rngs aren't random. regarding potripper, he was a superuser, it doesn't matter if the rng is perfect or totally broken if you can see the opponent's hand.

it doesn't matter if the rng is perfect or totally broken if you can see the opponent's hand.

That's why this issue doesn't really bother me too much. With my amazing hand-reading skills, I'm way above the need for randomness.