Ehm, a player doesn't win 1/n of the hands. You win very few hands in early positions and many in late position. Also what tends to happen is that a player loses a hand in the BB, gets a few good button/CO/HJ shove opportunities and is more likely to win chips that way.

Ok, but that is assuming that we are actually playing poker and not just shoving for a pre-determined outcome. I want to simplify my analysis to completely ignore the influence of position on winning and losing. You either have a board that is won by one player, or there is a tie between certain players, or there is a tie for all players. There are other subsets of results that are beyond the scope of my audit, but I am sure I could get there later.

The point is, I am looking for the incidence of any particular player getting hole cards that will win if they go all-in and anybody else at the table calls them. I find these events by looking at only the data points where I can say for sure whose hole cards have won - which would be hands where the boards go all the way to the river card and sometimes by the turn card. Does that make sense?

In those instances, would you not agree that a player should have 1/n chance to win in any given deal? If we completely ignore any poker that may take place around the table and just focus on the cards. The best analogy to use is 888's Blast after the timer expires.

Ok so what you want to do is compare two situations.

Player moves all in, regardless of position and/or action, and gets called by one or more players.

1) He lost chips the past hand
2) He didn't lose chips past hand

If situation 1 wins more often than 2 then the hypothesis is confirmed.

Is that correct?

Get a bunch of buddies (you probably will have to buy them dinner) and create a play money homegame with low and slow blinds, and then simply create the scenarios you want for one person for a bunch of hands. Have them lose a lot of chips (whatever you define as a lot) even if they call an appropriate size raise preflop and open fold the flop and then so whatever tests you want to do.

Do a few thousand hands this way and present your work. Of course you will still have the issues like what if Stars knows you are doing this and changes things or they change it for play money etc. but you or others can theorize those changes of the goalposts as you need.

Almost. I want to include all instances where the play goes all the way to the river, whether it gets there via all-ins or via normal play. Also, the chips that were lost in the previous one or two hands have to be at least the size of the small blind.

Also, each positive result is only confirmed if there are no players at the table that held better cards that could have won had they not folded

This is a good idea. I doubt that Stars can change anything. If there is something actually there and it is not just randomness manifesting itself as order in small sample sizes, this would show itself across all games on Stars.

Most people who are paranoid about this stuff assume the site can change things as they like. After all, the whole thing about a rig is the site is doing it to cheat players etc. However, with your study, that will likely not be the real problem.


What does this mean? If I have JdTd and 4 other players have K2o on a 5 handed table I have nearly 60% equity if everyone goes all-in. Heads up vs K2o with random other cards folded by the 3 other players I am probably a tiny underdog. Which are the "better cards." What about if a player folds JcTs and would have won with a 4 flush?

Obviously the chances that whatever you are trying to do goes anywhere significant are pretty much 0 give or take 0, but you need to be very clear in the conditions you select and the things you are testing. "Better cards" is not clear at all. Still, you can certainly take comfort in the knowledge that that you can only exceed the expectations of your endeavor.

This thread is quite interesting.

You're not understanding what I am measuring, which explains your sentiment.

I am proposing that the outcome of a board for any given player at a table is predetermined before the flop. The 5 cards on the board may be random, but the player(s) holding the winning or losing hole cards are not.

In your example, (JdTd vs. K2o x 4) you think you have equity and you would in a normal game of poker. However, if your JdTd is predetermined to be a losing hole card combination, the board will form to benefit whoever is supposed to win. Example: the board will come 3,4,6,7,9. If a particular K2o is to win, you may see a running flush. It may be a diamond flush to induce you.

In my study the hypothesis is that there is nothing actually left to chance. There is the appearance of chance. The only way to study this is to look at complete boards and to know all the hole cards present at the table for a given hand. I am designing a study to test this hypothesis. I hope this clarifies it for you.

"Better cards" means somebody at a table holding hole cards that would have won on the board against all other hole cards if they stayed in the hand instead of folding. Ex. 3s8s on 4s7sAd5s6s vs. AhAc, 7d7h, QsJs, 8d9c... Get it?

My apologies if I come across in any offending manner in response to any of you. I know this whole concept sounds ridiculous and all I want to do, for the sake of the game, is to prove that it really is not true. However, as an investigator, I have to follow the evidence until I can reach that conclusion for sure. If that takes millions of hands, then so be it. My next work will be to look at another replay and see if I get different results. I will do this two more times and report back.

If doing this study is +EV to your mental health then do it, but realistically as presented it will have zero value to anyone else. Still, your time to use, so choose whatever is best for you.

If I am understanding correctly, in order to test your hypothesis it sounds like you need to be able to observe all players' hole cards even if they fold at any time during a hand.

And, conversely, you'll not be able to accurately test your hypothesis (likely getting very skewed results) if you only observe a selected set of hands where the selection depends upon the quality of one or more players' hands.

Several people in the past have been tripped up by this "selection bias" and I don't want you to fall into that same trap.

You really don't have any evidence, just some vague feeling.

What do you think this accomplishes?

whosnext: I do realize that there is a risk for selection bias here, but maybe we are not quite in agreement as to how that selection bias would come about. I'm not following the second paragraph of your last post when you talk about hand quality. I think you may have meant hole card quality? When I say "hand" I mean the entire dataset on any given board, from dealing the hole cards to the final result on that board for every player involved.

What I described is the only way I can think of to mine the correct data to test the hypothesis.. unless there is another way to do a related hypothesis test. If I can't see the entire board (or enough of the board to tell who wins for sure) then the data point is speculative at best. I would not be treating any hand as being higher or lower quality. It is either useful or not. A hand where no cards make it into the board is useless, as is one that just goes to the flop (unless the result is a royal flush on the flop, for example).

For alternative hypothesis tests, I could for example collect a number of hands where aces are dealt to a player. I could count how many times the aces win and how many times they lose. Next I could look at how many times the aces win when played by a recent chip loser and how many times they win when played by a recent chip winner. Likewise, the same would be applied to the losing cracked aces. If there is a significant difference in cracking or holding between the two populations of players, we would have a confirmed hypothesis.

Didace: try providing something constructive for this discussion. If all of this is real, how would you go about proving it?

In terms of value to players, I can suggest that there is already value right now even without the study being started. Any player can simply look at their hand histories and see if their bad beats happened right after a significant transfer of chips took place on the previous two boards, especially if your villain was involved. It's just a suggestion on my part for now. Take it or leave it.

You surely realize that the quality of hands (however you want to define that) influences whether a deal goes to the river and which players go to showdown.

Of course, the converse must hold as well. Deals that go to the river with multiple players going to showdown are not a random selection of all deals.

And, I am sure you realize, players muck losing hands at showdown which can create its own selection bias issues.

That's why most studies looking into whether online deals are truly random have focused upon the distribution of hole cards dealt (in non-selective deals) and board cards dealt after all-ins.

P.S. I just saw your most recent post. I fundamentally disagree that spouting a view that the Pokerstars RNG is biased is a good thing in the absence of any real evidence or analysis. My personal view is that such unsubstantiated views can do real harm and can be a disservice to the online poker community.

"Bad beat," while it seems descriptive is not that specific a term. Is losing as a 51/49 favorite a bad beat? Technically I guess, but is it really? If someone bets 95% of their stack pre-flop as a 4-1 underdog and then in the flop gets it in after flopping a set and is a 8-1 favorite - is it a bad beat if they then get 2 outed on the river etc. etc.

Your study (whatever it is) is not going anywhere other than to make you feel better, and it seems that believing people are having collaborative chats with you is part of that mental soothing, when really they are trying to point out some pretty glaring flaws in your general approach.

Again, in the end it is your time to consume, so use as much as you want, but you will at least make whatever it is a bit more robust if you do the following:

- State very clearly specifically what every vague term means. Bad beats, good hands etc. in the context of your study.

- Explain why a site would do this (ie: how does it increase rake and the costs and risks needed to avoid being detected and exposed)

- Show the methodology of all of your work.


Realistically, if you ever do any significant research on this (unlikely) the odds are it will not be done properly as you seek to find the answer you are looking for, and that is said only because that is the path taken pretty much by others like you before you (granted nearly all of them never went past the chatting phase that you are in now, so I do not expect much beyond this).

Not trying to be mean, but a lot of very experienced people have done extensive analyses of massive databases and they have not seen the things you believe to have seen in a pretty tiny sample. As a general rule I tend to believe those guys in cases like this, but best of luck with your adventures if you carry on with them past the chatting phase here.

You may also find this guy's work interesting as well. He was one of the few to go past the chatting phase, but he never found his elusive rig in the end.


https://forumserver.twoplustwo.com/2...hould-1299786/
https://forumserver.twoplustwo.com/2...isive-1293249/
https://forumserver.twoplustwo.com/2...untry-1123425/
https://forumserver.twoplustwo.com/2...uency-1116343/
https://forumserver.twoplustwo.com/2...omaly-1007942/
https://forumserver.twoplustwo.com/2...l-play-985031/

I would prefer to, hopefully, help you to think critically about what you are attempting.

I don't want to go on a tangent of explaining why any site would do this. As I mentioned previously, it could be intentional or an unintentional consequence of programming. If this is all real, it is almost impossible to prove, because of the nature of the game - data is frequently incomplete and anyone can argue it is variance or a small sample size causing the observations.

Based on the last few posts that I have read, it seems that the idea of putting together a home game and playing out every single hand would be the most effective way to avoid selection bias described by whosnext. However, it is incredibly inefficient and would produce thousands of hands at most unless multiple groups collect this data.

Riddle me this: if I am selecting only the hands where the board is complete, why is this selection bias significant and why would it negatively affect my study if I am considering every single hole card combination at the table? Or is the selection bias beneficial because it will show me exactly which players would pull off unlikely wins on the board as it runs out to the river? Is it selection bias or is it prudent selection of data points? High action hands where there is a significant amount of chips being exchanged?

For a moment, let's say I do obtain a large number of these types of hands that can be analyzed and studied for corellation and significance.

Can you suggest why, in this subset of high action hands, the (biggest) losers from the previous hands would be more likely to win and the winners from the previous hands be more likely to lose? Keep in mind, by "win" I mean absolutely win the board with the hole card combination that beats that of every other player at the table.

Monteroy, thanks for the links. I looked at those threads but what I am studying is just not anything comparable to the type of data being studied in those threads. Again, I understand historical attempts to find "rigs". I want to find a solution to this particular hypothesis in this thread.

Assume none of us are actually playing a random game when we play online poker. Assume the result of each board is fixed before any of us make a bet, and that losing players are favoured to win next over winning players.

If the above is true, how would I prove it or how would I disprove it if it is in fact randomness?

Two more final table broadcast transcripts. If I get the same results as with the first one, that would be three times as much data as before. It would give me more reason to study this further.

There is only one way to prove that the pokerstars software is rigged. That is to find their source code and go through every line, and find the place where it is rigged. What you are hoping to do is find supporting evidence of your theory.

But here is the thing - if you read about rigging theories you will find people who are certain that 1) games are rigged to make winners win more; 2) games are rigged to make losers win more; 3) games are rigged to cause action; 4) games are rigged to punish people who make a cash out; 5) games are rigged to reward people who have just recently signed up; 6) games are rigged to reward recreational players; 7) games are rigged to make draws come in more often; 8) games are rigged to punish professional players; 9) games are rigged so that someone who lost the last hand will win the next one; 10) games are rigged so someone who won the last hand will win the next; etc. etc.

I have been reading about all of these rigs for over a decade, and I can assure you that there are many people for each one claiming with certainty that it is true (except for yours, I think you're the first person I've seen with this one). I have played on many sites, and on every single one the people yell and scream about how whatever happened is just further evidence of a rig.

And yet, in all these years, with millions upon millions of hands being analyzed, not a single person that I know of has ever come up with any evidence at all to support ANY site having the deal of their cards being rigged. (there have been other issues, like superusers, but that is a different issue).

Here is a possible explanation for what you think you have seen. In the late stages of a tournament, anyone who loses a significant portion of their stack is likely to end up the short stack. In this situation, they rarely wait for the blinds to come around, and so they end up shoving with a wide range of hands. It isn't unlikely that they might do this within the next few hands, as soon as they get a face card, pair, suited cards, connected cards. They often get called by large stacks, because these players know the SS is shoving wide, and with the blinds and antes in the pot, calling doesn't hurt them much, and they have a chance to take out an opponent.

Does the short stack tend to win these hands more often than one would expect? It depends on what you would expect? You seem to think that in a 9 handed game they should win about 11% of the time. But in fact the probability is a lot higher than that, because they will never get called by all 8 opponents. If they get in against one player, a random hand still has a 38% against a top 25% hand. A king with a random card has a 44% chance against a top 25% hand; even any random queen is a favorite against a random hand. Since the medium stacks often fold, and the big stack is often the one who recently won a bunch of chips (your other rig), the SS might get it in against that recent winner a lot more of the time and win a lot more of the time than you would have expected.

Secondly, when the short stack shoves, gets called and loses, you don't notice it, because that is what is supposed to happen.

By the way, I have put your data into a statistical software package and played around with it. I looked at whether or not people are more likely to win within the next 3 hands following a big loss and I saw nothing to support your theory. I only looked at winning and losing, because I don't think it is relevant what would happen if all hands and all cards went to the river - because there are just far too few hands where we see the river, and also because holdem isn't played in such a way to make that relevant.

I do appreciate this response. Could you please re-run the data but including all hole hands. Wins by folding do not count. The issue is precisely that the winning hole hands (those that would win if they stayed through to the river) get folded by players after losing their chips in the previous hand. This is because the winning hole hands often look like trash hands.

If it is rigged, why would they broadcast the rig in the biggest tournament they have with hole cards face up though?

Assuming it really is rigged, two possibilities:

1. They do not know it is rigged. It was programmed that way by accident or it is somehow an unintended by-product of their RNG's generation of "randomness".
2. They know it is rigged and that the rig is so very subtle that they do not expect anybody to ever find it. And, if I am to give in to what most people are saying in this thread, there is no way to prove it.

Practically speaking, if it is there, this property of the RNG does not allow a player to necessarily win tournaments (unless somehow it can still be predicted in heads up play). It does, however, potentially help players to cash on tournaments when they may have otherwise finished outside of the cash-in cutoff. This isn't a matter of knowing that you have lost chips and now your next hand is going to be the stone cold nuts (at least I don't think it is). It is more of a matter of knowing when the hole cards that you are dealt are likely to be no good against an opponent and, on the opposite end of it, when your hole cards will likely perform well on the board that is to come.

At least that is what I am expecting at this point.. if it is there at all.