I showed you the links of that other riggie, not because of the specifics of his theories, but more because of his approach, whereas he chose to identify rigs from what he believe he saw or believed and then tried very hard to prove the existence of those rigs. That has innate bias within it, just like your approach has as well.

You have said that the site's motivations are not important, which of course they would be for any programmed rig to exist. The concept that an "unintentional" rig that somehow rewards players that lost chips a couple hands later, as thousands of tables are going and people are bouncing between tables makes no sense to exist at all.

However, reality is not really important to you - trying to hunt your personalized rig belief (still not quite sure what it is since you have yet to be specific about it) like Ishmael is what matters, and anyone with a passing familiarity of that story will know how useful it is to talk sense to a person like that.

This guy's theory is in a sense different given how obscure and pointless it is (vs say new player boomswitches), but he is kind of a mixture of riggies before him (one was obsessed about flopped bottom sets vs top 2 pairs for instance). Perhaps he will find this list helpful with his research.

https://forumserver.twoplustwo.com/s...ostcount=84039

Whether your theory is valid or not, having three times the data you have now is meaningless. You have been told how large a sample you need. If you are set on following this through you need to figure out how to get that data.

Ah, the classic "It's really subtle so statistics won't show it, BUT I CAN SEE THE PATTERN EASILY".

A slightly different perspective: So what you're saying is that you have found evidence of non-random dealing, that GamingLabs, an independent third party whose job is to test random number generators - by using actual statistics and inspecting the source code - couldn't find after reviewing millions of hands? If anything, you should be contacting GL for a job interview; I expect they'd want to bring you on board after you found such a glaring flaw in their methodology.

Edit to add: Is it possible that GLI and PokerStars are involved in some secret collusion to hide a rigged RNG? Yes. Is it even remotely likely? No. I appreciate your concern for the fairness of the game, but this path you're on right now is not a productive one, and I fear you may be wasting time that could be better used on something (anything!) else.

Kelvis, I see now that this whole time your point was not to help me at all, but rather to nit pick at my methodology and reasoning behind why the "rig" (if that's what you'd like to call it) would exist in the first place.

Monteroy, likewise, you have chosen to go the way of personal attacks and to label me as a "riggie" (this guy is like that other guy) instead of helping me figure out how to prove or disprove what I am seeing. The only useful contribution so far was the suggestion about trying a home game.

Curiously, none of you are able to answer the questions that really matter but you're spending a lot of your time ridiculing my attempts to figure out the answers. Your minds were made up from the beginning: the answer is that there is no "rig" and there is no reason to study any of this.

Instead of spending your time on unnecessary responses, why not go and earn some money on online poker? I am looking for very specific information here and recently you are just cluttering the thread.

I know I don't have enough data. I know how vague this looks because of the lack of data. But why repeatedly dwell on this? The whole point of this topic is to help me set up data collection. If you can't do that, then why bother wasting your time on useless posts?


For anybody that is actually willing to help, please refer to the first page of this thread, posts #5, #8, and #24. This is the best that I have right now in trying to explain what I am trying to study. Again, I know it looks vague, but if it turns out to be something that allows a player to bet less or more chips and to avoid traps, then it is significant.

Either:

1. Assume that my hypothesis is true. Assume there is no poker and that the board's outcome is pre-determined (i.e. everyone goes all in, the chips either go to one player or certain players tie). Players who lose chips in the hands immediately previous to the current hand are more likely to be dealt winning hole cards that will beat every other player's cards at the table in the current hand AND players who win chips in the hands immediately previous to the current hand are more likely to be dealt losing hole cards in the current hand.

2. Assume that my hypothesis is false and that players are awarded hole cards completely randomly, so anybody has a roughly random result if every single player at the table went all-in and the board played out to the river.

What is the most efficient way to collect data to prove the hypothesis or to disprove it?

Your problem is that your conspiracy involves bad cards that suck out, in tournaments.

So, your theory is that when a player loses a lot of his/her stack in a tournament, then they intentionally get dealt bad cards that will suckout and win, replenishing their chips at the expense of the players that have not recently lost a lot of chips and consequently are dealt good cards that are destined to lose by the river.

Did I get that right, I read the specific posts ITT that you suggested.

Since the thread wound up in probability subforum, I say that you have not established a baseline of enough hands, and have not applied statistical analysis neither classical nor modern Bayesian, to illustrate any statistical bias or conspiracy.

Imo, thread belongs in psychology subforum because OP is weaving quite a web and getting lost in same.

Yes. That is one possible scenario for this pattern. However, the more practical scenario is that larger stacks can also lose the most chips in a hand (doubling up a small stack, for example) and get strong hole cards next that can be used to increase their stack size. This can involve bad hole cards or good hole cards for any of the players. It is not just limited to bad cards.

Again.. instead of taking shots at the idea, help me establish a way to test it?

No, I asked you what you were going to do with a bigger sample and how you would do this mathematically. You came up short.

But be welcome to turn it around as you please so you won't feel bad.

So the web gets more tangled, and can involve both good and bad cards, and large and small stacks.

You can not test it. Therefore I can not help you test it. I do not need to take shots at such a fantasy, and I already gave rudimentary mathematical reasons, which you responded to with further conspiracy.

I can tell you are smart, and motivated to solve a problem that you feel has been overlooked by an entire industry for possibly two decades.

I help you by saying there is no conspiracy, you have gotten too attached to this fantasy.

It is not healthy.

First, you would need many millions of hands with access to hole cards of every player at the table, along with player chip stacks before/after every hand. As far as what to do with the data, I'm not an expert on the current statistical theory for this type of question. My biggest concern is that you will be unable to collect a sufficiently large data set on which to perform your analysis. Perhaps come up with a solution to the data collection problem first.

Iceball, thanks for probably the simplest and nicest way of describing my current impasse.

Otherwise this has been a redundant cycle of asking for help with planning data collection, only to hear that I have not collected enough data to make my claims (which I already know).

I can spend hours analyzing replays, only to net maybe 40 semi-significant results per replay. My time would expire long before I reach a collection of hands that would be statistically valid.

I can arrange "home games" to maybe acquire a few thousand hands.

I can ask Pokerstars to arrange a simulation where the board is dealt to the river card regardless of whether everybody folds or not, and to collect millions of hands that way showing the board's cards and everybody's hole cards.

I forgot to mention this in my last post, but you're correct that you would also need the board cards that would have run out for every one of those hands (since a player may be dealt better hole cards but lose if the board played out to the river). The only way to source this data completely (without spending several lifetimes watching/playing tables) would be through PS like you said, but the likelihood of them approving your request is near zero.

Not to mention that, even if you were to obtain the data, it's highly likely that you wouldn't find anything of value. This is why I think there are better uses of your time than taking this investigation further. Certainly you are free to spend your time as you see fit, but I just wouldn't recommend it.

If I get some time, I can try, but for it to make any sense I have to use the full community board. There are very few hands in there that qualify. Also, your statements seem to me to make it even less likely that there would be this type of rig. Why would they rig it so that a particular hand would win, when they know that the hand probably won't even see a flop?

Also, I'm going to try to answer your question regarding sample size. You absolutely do not need millions of hands. Anyone suggesting such a thing simply hasn't done any inferential statistics. Unfortunately, there is no one answer for how many you need, because we don't know exactly what kind of result you would need to confirm your hypothesis. For instance, if you think that something should happen 10% of the time and it has been rigged to occur 11% of the time - this would need a very large sample. If, however, you would only consider it rigged if it happened 25% of the time, a much smaller sample would be needed. In any event, you need a whole lot larger sample than you have. And I believe that you have to have the final board from all hands - and you simply aren't going to get that, because I don't believe it actually exists.

You are a riggie, that is not so much a personal attack as a correct statement of who you are, but that is not really important. You will not have anyone else be interested in helping you figure out how to prove whatever it is you are trying to do (it is still pretty unclear), other than offer some gentle suggestions that you are not following any proper form of proticol for doing this type of research. You can flat out ask people yes/no if they think there is anything to your idea, and whether they believe you can research it properly. Who knows, perhaps I will be surprised with the result (ie: if 1 or more yes is said).

I suggested you provide essentially a strategic plan for whatever your rig is be clearly defining it and explaining as part of that why a site would do it. You dismissed that as not being important to your needs for now, which was exactly the answer I expected (since the only direct answer would be that there is no reason for the site to rig it as you suggest), and that showed that you only want "help" that will agree with whatever it is your quest is about. The Ishmael syndrome.

Anyway, while others are correctly telling you that this endeavor is only a waste of time for you, I will go about it in a way that might be more acceptable to you.

Don't worry about being clear on your concerns. Do not worry about obtaining a proper sample or doing a proper statistical analysis. You can decide to interpret whatever hands you look at in whatever way you want after the fact, and the good thing is that in any sample of 1,000 hands or more you will certainly find outliers that are not expected. Sure, outliers are part of the math of poker and are unremarkable, but what you can do is see what the outliers are and then construct a theory to fit them and convince yourself that you are uncovering a truth that nobody else in the industry for decades was able to identify.

Sure, nobody else will believe you or your results due to the obvious cherry picking, but do not let that upset you since this activity is all about you getting the emotional satisfaction of researching the concerns that are firing in your gut.

I would suggest that you set a max timeline, perhaps 6 months of full time research, so that you have a fixed limit of your life you choose to consume on this, and then you can move onto the next project, whatever that is. You can start this one by emailing Stars to ask them to set up special games for you to do research. Sure, their answer to that request will be a polite no, but you should spend the time doing that as part of the early stages of this special quest. Hope these suggestions helped, as at least some of them will be things that you want to hear.

All the best.

I get that I will need a very large sample size if the difference from the expected occurrence rate is small, like what you said with 11% vs. 10%. Fortunately what I am observing is happening upward of 50-70% of the time vs. 11%.

I'll answer your question for why this type of rig would exist:

1. to allow people who know the rig to exploit the rig - ex. you know that your Ac7c will hit a flush against KdKh, which will lose, so you will call a large raise OR you know your aces will be cracked so you will slow play and fold

2. to create more action, since losing players are more likely to tilt whereas winning players are more likely to be over confident in their hands

3. It can be a completely random glitch. I do not think this is the case, but who knows?

In any case, another data set of strong anectodal evidence is on its way soon.

This is the problem - Your evidence isn't strong. This has been pointed out to you.

Cool, then as I said do not worry about whether you are following proper statistical procedures or why you are the first one in the history of this industry to discover this. Just take pride in that you are working towards a mission to prove to yourself about this, and in doing so you will achieve some sort of self actualization.

I would point out that if this was the goal then it could be accomplished in MUCH easier ways than the multiple steps needed for your theory (whetever it actually is), so if this was some form of superusing backdoor - they would just do something a bit easier - you know like let people see the opponents cards or something.

Still, I will suggest to you what I have with every other riggie who believes they have cracked the code - use that secret knowledge to exploit the system for riches. Perhaps you can be the first one to do that.

Action hands decrease rake in cash games as clumps of money move instead of small amounts between players. Poker rooms would love nothing more than a game where small amounts move back and forth between players at a relatively equal rate, all the while raking that churn. In tournaments they are pointless as the rake is already paid and why draw attention to a rig in that case (yeah, I know - why a room would do it does not matter to you much).

This makes zero sense as a glitch.

Perfect! Remember to spend as much time as you like researching this (don't forget to email Stars asking them to set up special tables for this experiment), and do not let anyone else frustrate you with questions about sample sizes, proper statistical methods or the logic behind your beliefs. You need to do this for you, so continue on this special journey!

All the best.

Monteroy, my greatest moment of self actualization may truly lie in leading you to spend so much of your time on responding with reasons as to why this is such a fruitless pursuit. This is time that you will never get back.

While I have you engaged, why do you think a simple rig is preferable? Like the one with a superuser who can and did get caught in the past?

Why must a rig favour an entire poker room for rake? Action may benefit certain players, who can bring money back to the company from depositors directly through rigged hands.

I agree that it probably is not a glitch. Thanks for your time.

Don't worry - the 10 -15ish minutes I have spent total replying to you amused me enough, but that being said - I do limit the amount of back and forth with people like you, because after a very short while it has nowhere to go, as your last post illustrates.

Anyway, you have work to do that nobody else will place value on nor care about, so stop dilly dallying and move forward to your extensive testing phase. Coffee's for closers, so get to it!

All the best.

I really don't want this thread to turn into another endless discuss of various riggie theories and why or why not they may be viable or why or why not an online poker site would rig their RNG in such a fashion, etc. If the thread continues down that path, I will move the thread to the "Online Poker is Rigged" thread in the Internet Poker forum.

To the extent that a fruitful discussion can and does take place on how to develop a specific hypothesis regarding the fairness of an online poker site's RNG, how to collect a non-selective and unbiased sample to test the hypothesis, how much data would be required to adequately test the hypothesis, how to set up and perform statistical tests of the hypothesis, how to interpret the results of the statistical tests, etc., then I think the Probability Forum is a good place for this thread.

Thank you very much.

Got it, whosnext.

By the way, did you have an answer to my question regarding why limiting my selection of data points to action hands is a bad thing, if I have the whole board of cards and everybody's hole cards for each data point, and if I am determining the hole cards that won (including those that folded, but should have won)? If I get hundreds of these and previous chip winners (in last two hands) end up with lower odds of winning in the following hand than what would be expected, why would that not be a start for this study?

Fair enough whosnext, and I certainly agree that at this point the thread will be much more interesting from my perspective with the discussion being composed of puzzlefish asking many questions (like he just did) and others trying their best to guide him and answer him as he continues his statistical journey into the unknown.

Do some good and extensive work puzzlefish and update this thread often with your results as you see them, I will definitely not say anything further to hinder that.

All the best.

You can't "prove" a negative. I already gave you the answer: before you even spend one second thinking of how to collect data first ask why would a site rig tourneys in such a pointless way that literally benefits nobody, and then apply Occam's Razor to any theories your mind produces.

Im not a statistician by far. However, for the data we gathered at university for sociology lab, the first rule was:

“Absolutely do not run everything against everything”

The professor was trying to help us understand that quirky data happens over any sample that a human can reasonably gather by hand (questionnaires). The software would legitimate causal relationships that “technically” passed muster but were in fact worthless bunny trails that would waste our time and eventually prove to be non-causal and barely spurious.

We were instructed to have a hypothesis first, gather data specifically for that theory, and see if data supported the theory. Any other weirdness in the data was to be ignored.

You can make data say anything you want by simply creating a hypothesis after the fact, supported by spurious relationships within a certain size of sample.