Online Poker Rigged? A Thorough, Informed Discussion
Here at ProfessionalRakeback, we are not just superb poker players and enthusiastic fans of the game. We also like to think of ourselves as educators, helping to dispel myths about online poker, its legality, and the mathematics involved in the game. It is on this latter area that we wish to turn our focus today.
We've recently received a communication from a customer claiming that there's something funny going on with the random number generator at his poker site. These kinds of comments are par for the course in the online poker world, and we generally react to them with an exasperated sigh and click over to something else although we sometimes cannot resist the urge to poke fun at rigged theories.
However, this particular message was unusually well-written, for a post of its nature, and it contains many of the themes frequently found in these types of allegations. Therefore, we've decided to go through the text, point by point, and show where and how the reasoning given is flawed. Much of our analysis is also applicable to other similar accusations, so you can read it and arm yourself against these kinds of erroneous complaints going forward.
The Message + Our Thoughts
We've reproduced the text we were sent below, divided up into sections. After each paragraph or two, we give our thoughts about the arguments raised in that portion of the message.
I am not here to give a bad beat story nor claim that Phil Nagy has created bots that know the run outs in order to steal my money. In blitz I have been about a +1-2bb/hr at .25/.50, and in tournaments I’ve broken even across my two usernames on ACR and BlackChip.
Well, this sounds pretty reasonable. The author is a small winner in cash games and about break-even in tourneys on Americas Cardroom and BlackChip Poker, which are part of the Winning Poker Network. But wait, there's more:
However, I have now played at least 150-200,000 hands on ACR, and about the same on ignition, and I do strongly believe at this point that while I don’t think ACR is “rigged” (as in, once again, I don’t think Phil Nagy has intentionally created a site that is designed to fuck me) I do believe that there is something significantly wrong with their RNG.
Ah, now we get to the main point of the piece. He believes that there's something wrong with the RNG in use at the network. Let's continue on and see why this player feels there's something not legit about it:
I have the same belief of the app PokerBros — my sample size there is smaller, probably closer to around 50k hands, however I find it interesting that both ACR and PokerBros use the same lab to verify their RNG. To use PokerBros as an example, it’s an app specifically designed for people like me to say, let’s get 20 buddies together and use PokerBros as a platform for our home game cash games/tourneys, and they make money off of the VIP purchases and other benefits. Why would they want to rig some of my friends and not others? Makes no sense, of course. There is of course the elaborate “it’s rigged for action so big clubs can get more rake through more KK’s vs AA’s, but that’s so far fetched that I certainly don’t believe that.
The author now contends that both the mobile app PokerBros as well as ACR are utilizing suspect random number generator software. He incidentally notes that they both use the same testing agency to certify their RNG algorithms.
The company he is referring to is iTech Labs, an Australian firm. However, the suggestion that iTech's testing methodologies are not up to snuff is a big claim. They're plenty good enough to maintain business relationships with high-profile clients, including GVC (parent company of partypoker), Connective Games (software developer for the Chico Poker Network), and NetEnt (a renowned slots developer). If iTech is really cutting corners or otherwise doing subpar work, then they have somehow managed to pull the wool over the eyes of many organizations, some of which are publicly traded enterprises operating strictly in licensed markets.
Curiously, even while forwarding his unlikely theories, this individual seems to be backtracking at the same time, conceding that “[it] makes no sense, of course,” and “…that's so far fetched.” This type of “No, there is not anything funny happening, of course, but THERE REALLY IS SOMETHING FUNNY HAPPENING” logic is not exactly unheard-of with these kinds of speculations.
However, the string of issues with their RNG, as well as ACR’s, in my experience, has not come close to the trueness of Ignition or Pokerstars’ RNGs, for example. Historically, I have been very open minded to the possibility that I am simply watching variance show it’s powers in front of my eyes.
At the same time, while I have not done any mathematical study or analysis of calculating the unlikely run outs on any of these sites, I do believe that 150-200k hands on two sites can begin to give me a clear indication as to whether something is fishy or not.
Here, the author of the message contrasts the RNG at ACR to those of Ignition and PokerStars, which he deems to be legitimate. We do indeed concur with his assessment that Ignition Casino Poker is a trustworthy room and one of our highest-rated online poker sites. Yet, Ignition too has its share of detractors who concoct elaborate tales of unfair card distribution. Riggage, it would seem, is – like beauty – in the eye of the beholder.
The complainant does admit that he has “not done any mathematical study or analysis,” which is very honest of him. However, this revelation doesn't tend to strengthen his case.
As far as the contention that 150,000 to 200,000 hands played on a site can indicate possible rigging, this is extremely unlikely. People often underestimate the swings involved in online poker and the amount of time it takes for variance to smooth out.
There are online simulators that can run a specified number of trials of a particular volume of hands and plot the results. We ran one such simulation for 200,000 hands, given a winrate of 2 bb/100 (about what the commentor said he had achieved in fast-fold Blitz Poker) and a standard deviation of 90bb/100 (typical for six-handed Hold'em). The tool generated 20 random samples and then graphed them. This is the output:
Though most of these simulations generated winning results, a not-insignificant fraction of them saw the hypothetical player book a loss after 200,000 hands played. In the statistics accompanying the simulation, we were informed that the probability of a loss after 200K hands is 16%. The minimum bankroll needed to ensure against going broke (with 95% confidence) is 6,066 big blinds (more than 60 buyins assuming 100bb stacks). Within the 200,000 hand simulation, the chance of a downswing lasting 50,000+ hands is 57%.
With the size of the swings that are possible and probable, 200,000 is clearly nowhere near enough hands to be able to say anything definitive about a random number generator just based on casual observations and gut feelings. We really would need a much larger sample size and detailed statistics to be able to even begin to analyze if something was amiss.
Our correspondent didn't go to the trouble of performing any kind of probability analysis on his results. This illustrates a common misjudgment that plagues human psychology: the tendency to use heuristics, or “practical” guidelines, to arrive at a conclusion rather than a more rigorous (also more difficult) mathematical approach. Vice Chairman of Berkshire Hathaway Charlie Munger briefly discussed this type of mental error in a speech he gave at Harvard University in 1995.
Once again, I am basically a break even player on ACR, and that is likely due mostly to the fact that the field is very strong. So I’m not one to send this message saying “I break even on ACR and kill it everywhere else therefore it’s the RNG”. I’m very well aware I could break even or lose money to the players I’m up against on blitz. They’re strong players. But the strings of 1-4 outers on rivers, Aces on flops after 4betting with QQ, KK, horrific runner runners, has no longer become a variance related result, in my opinion, because it happens an unrealistic amount of times.
In case this helps my argument, these runouts have been on my side many times as well. I’d definitely be lying if I said it helps me just as many times as it hurts me, but I do think it’s close, especially because we tend to forget the ones that help us more often than we forget the ones that hurt us.
This passage again just contains anecdotal evidence about bad beats and coolers, of the sort that happen regularly at any online poker site. The poster implies that he has been hurt by these shenanigans more than he has been helped by them although he does admit that “it's close.” He's displaying the classic symptoms of negativity bias whereby individuals tend to overestimate the frequency of undesirable outcomes as compared to favorable ones. The sender of this missive even recognizes this bias in the last sentence quoted above.
To finish, I trust ignition with everything in me. I simply never distrust their RNG by any means — I like that they show your percentage during an all in, and more often than not, it sticks true — you tend to win more often when you’re ahead, and lose more often when you’re behind. I have seen sick runouts on Ignition, but my gut always says: wow that was sick. But it’s not the RNG, it’s variance. I had one day where I lost 7-8 buy ins on Ignition Zone in 2 hours (rare for me) with sets over sets 3 times, flushes over flushes with 3 of same suit on board, etc etc and I closed my computer and said, my goodness that was what you call running bad! But on ACR it’s often gotten to the point where I can smell the fishy nature of that RNG. It just cannot be...
Anyways, thanks for reading this. I’m sorry I didn’t provide any evidence or math behind my argument, and yet again I will say I’m still open to the possibility of variance, and I also strongly do not believe that the (what I believe to be) unrealistic RNG on ACR/PokerBros is done intentionally to hurt any of us. I am, however, making the difficult decision to leave ACR for all cash related activity, and only play maybe a couple of the bigger tournaments per day, and switch solely to Ignition because as much as I love everything about ACR/BlackChip, I have gotten too skeptical (and frankly sick of) the RNG. While I would be happy to provide you with many, many stories, I know you have heard them all and variance is bound to produce these bad beats with the amount of hands played by all players every day. It’s the amount it happens, however, that has made me leave and send you guys this message to see if we can have a conversation about it, poker player to poker player. Thanks again, hope we can have a brief discussion about it.
Our correspondent closes by concluding that the RNG at Ignition “sticks true,” but the one in use at Americas Cardroom and BCP is untrustworthy. What he apparently fails to realize is that virtually every poker room on the internet has detractors who feel that the card dealing at that particular place is suspicious while simultaneously believing that there are other sites that are fairer. Indeed, the frequency of searches for “Is Bovada rigged?" compelled us to compose a dedicated piece answering the question. (Cliffs: It's not).
Check out these real-life comments left by actual players:
Moving Right Along…
In conclusion, we find the propositions contained in the message we received to be without merit. The author even half-realizes this by making acknowledgments like “variance is bound to produce these bad beats,” “I didn’t provide any evidence or math behind my argument,” and many others.
We've certainly never encountered any compelling evidence that any of the mainstream poker sites that serve Americans has a faulty RNG. Even at rooms that we do not recommend, it's other aspects of the gaming experience that we're unhappy with, not the RNG.
If you would like to learn more about the leading poker destinations for Americans, then head over to this rundown of the best U.S.A. offshore poker sites. They all have fair and honest RNGs along with many other advantages.