Thanks a lot. Finally an argument - not just a statement. I agree that we should find a big enough sample in one milion hands using this formula.
If this is what you believed - then you would analyse your sample hands to see how often you won against big stacks vs how often against small stacks. You could compare this to general data regarding how often your allin was vs a big stack/small stack. This is my point (which you don't seem to be able to comprehend for some reason) - whatever your hypothesis is (regarding how they rig their software) you can devise a test for it. Always. This is the brilliant thing about scientific approach - as opposed to the rigtard approach - of just talking s**t.
I am well aware of this. My point was that this wasnt done in the hand analysis we are talking about. So there still is a possibility of people being right about rigged games. ( By the way, I dont think its rigged ).
Totally incorrect- I repeat - you cannot have your cake AND eat it too. By this I mean - the poker site cannot have legitimate deals AND also rig the deals. There is no possible way to do this and hide it from ALL types of analyses.
ALL types of analysis..... But it is not possible to make a statement about apples when You analyse tomatoes ( thats what you learn when you attend the 2nd day of preliminay school too). The Analysis we talk about analyses winning percentages. Not the EV. By analyzing the EV You would spot this of course but if You are so aware of the scientific method You should know that You have to DO the analysis before talking about its results. But this hasnt been done, so that You might find that after analysing the same sample again You would see correct winning percentages and screwed EV, so that the same sample of hands that indicated a fair game before now would indicate a rigged game. You dont get around this problem by calling me stupid, You would have to analyze the EV.
And one last weakness of this analysis is that once the poker rooms are aware that You are using it they will use 2 RNGs. One for hands in which at least 2 players are allin, working legit. And one for the 93% of the other hands that could be screwed up totally without Your analysis ever spotting this. Analysing the complete set of hands would reveal this of course, but since you havent analysed these hands at all you can make no statement about them.
So there'S plenty of room left for the "rigged online poker conspiracy".
The point is , it really doesn't matter how they "achieve" it, I can promise you there would be a way to statistically determine whether it is being done or not. I challenge you to provide any example, that could be effectively hidden. You will not be able to do so. It is completely illogical (and mathematically impossible) to think that you could.
I'm not saying it cant be hidden completly. But it could very well be hidden from the kind of analysis that is referred to here as my AA vs. KK shows and as the 2 RNGs example shows. Both cases would not have been covered up in the analysis done, so it's true that there is a way of analysing the hands that would uncover it. But it's completly wrong that the analysis done proves these cases wrong.
I hope You get my points now:
1)Just analysing the Win% of all-in hands is far from a complete analysis. ( No EV check, all other hands can be rigged ) 2)Just because a complete analysis MUST reveal rigged games an incomplete one must not. 3)And therefor the analysis done is not good enough to decide wether online poker is rigged, though it "prooves" ( there always is a close to 0 chance that you analyze a set of hands in which variance strikes ) that allin hands are played fair. 4) Though I tried hard I was not able to find another evaluation than the Win% 5)I dont believe the games are rigged, but they still might be unless a complete evaluation shows us correct % and correct EV in all kind of hands.
------------
------------
OMG - I am a Genius ( if you are right )
The point is , it really doesn't matter how they "achieve" it, I can promise you there would be a way to statistically determine whether it is being done or not. I challenge you to provide any example, that could be effectively hidden. You will not be able to do so. It is completely illogical (and mathematically impossible) to think that you could.
Below is an analysis of 1 million hands from which the preflop allins were taken. First number is the number of hands. Second is the expected number of wins/losses, third number the actual outcome, then the differrence and at last the standard deviation. Ahead 36101 24002 24093 +91 88 Behind 33524 11726 11738.5 +12 85
As long as the results are within 2 standard deviations the results can be called "very close" statistically. In this case the analysis shows us a fair game. But still the poker room has a "margin" of 352/340 hands it could screw up without changing the reseults of our analysis.. You would still be in the 2 standard deviations You need to considder the results as accurate.
The deviation can range from -176 to 176 without implying anything else than a fair game. Remember that these are all in hands. Say You play NL100 then this could mean that the poker room has taken 352 times $100 from You to keep the depositing donkeys happy while still being able to "proove" the game was fair.
So cheating within the deviation can not be spotted. And this is also true for EV, so even further analysis wouldnt uncover this.
And as You cannot analyze a neverending amount of hands you will always have a standard deviation bigger than 0.
You can send my Math Nobel Prize Money ( as I have just done something mathematically impossible ) to any charity of your choice !
And here is one for the true believers of rigged poker: NEVER check the "Do Not store Hand Histories Locally" because then the room knows You cant spot them cheating !
------------ wherers my last update
------------ the margin is just half as big of course.... 176/170 hands, because 2 satndard deviations are from -88 to +88 and not -176 to 176. My fault, but still enough to steal YOu $17600 in NL100 and make a poker room pro lough all the way to his free WSOP seat.
------------ And on the other hand this means that the +91 deviation is out of the standard deviation. So by Your criteria I have just posted a 1 million hand sample that prooves the game is rigged, because the value should not be above 88. Considdering that we have analyzed 9 times the minimum sample this is statistically relevant.
The interesting thing is that opposed to what people are ranting about this indicates the games are set up in favor of the favorite hand as the deviation is above the limit for the hands in which you are ahead when going all in.
Joined: Apr '09
Location: Australia
Age: 40 (M)
Posts: 6483
Posted by kraemer:
As long as the results are within 2 standard deviations the results can be called "very close" statistically. In this case the analysis shows us a fair game. But still the poker room has a "margin" of 352/340 hands it could screw up without changing the reseults of our analysis.. You would still be in the 2 standard deviations You need to considder the results as accurate.
It is quite late here and I am fairly drunk by this time (yes, on a monday night..) So I will reply to this only.
The poker rooms can't use the "spare" standard deviations to manipulate the results. Why? Because any manipulation they caused would be added on top of the natural deviation. It would push it beyond the threshold allowed.
Anyway, I'll reply properly tomorrow. I respect that you have thought out your post properly anyway. So I'll make a better effort (sober) tomorrow to reply in kind.
------------ one more thing, fwiw - like you I am not saying it CANNOT be rigged. I am simply saying, I am yet to see anyone provide any evidence it is and personally I find the motive (of the poker sites) vs the risks (reputation, license, etc) to be lacking. As such, I find it odd that people make such severe claims without any empirical measurements to back it up.
Joined: Oct '10
Location: United Kingdom
Age: 48 (M)
Posts: 3
EZY, I BUILT UP A TIDY BANKROLL ON THERE , N DIDNT GO TO DEEP IN HIGHER STAKES BUT FOUND THAT FOR EACH ALL IN I CALLED I LOST, THAT WAS IN TOURNEYS , AND IN CASH GAMES, HAD TO WAIT FOR RIVER TO FALL B4 MADE ANY BET AND HAD TO FOLD TO EVERY BEY WICH WAS MORE THAN FRUSTRATING, SO NOW CLOSED MY ACCOUNT WITH THEM, NOT FOR LACK OF WINNING JUST COZ GOT TO MONOTINOUS N COULDNT PLAY EVEN THE BEST OF HANDS
Joined: May '09
Location: India
Age: 37 (M)
Posts: 4873
sorry jess, not leaving much for you to reply to. enjoy your hangover lol
Posted by kraemer: So there still is a possibility of people being right about rigged games. ( By the way, I dont think its rigged ).
Yeah just like There is possibilty that your computer gonna blow up in 3 secs There is possibility that your car/ to be car is a giant robot alien There is possibility that right below you there is a nuclear bomb. There is possibility that your neighbour is from mars
If you get my point, lets not talk about whats possible or whats not but how much is the possibility ?
Posted by kraemer: And one last weakness of this analysis is that once the poker rooms are aware that You are using it they will use 2 RNGs. One for hands in which at least 2 players are allin, working legit. And one for the 93% of the other hands that could be screwed up totally without Your analysis ever spotting this. Analysing the complete set of hands would reveal this of course, but since you havent analysed these hands at all you can make no statement about them.
Are you really trying to say poker sites are using two RNGs First rigged rng that starts the hands and as soon as two players go allin they shift to another rng which is random??? If yes then rofl at your logic And if you have pocket cards of both players of hands that didna reach showdown, you can analyse all of them too. Btw though im not that much into programming, maybe jess will make you realise how much more calculations site will have to make and regarding the server costs they might end up paying more than they gonna win from it. And if there are two rngs in a software licensing authorities will most probably notice it.
And lets suppose there are two rngs working in tandem, rng1 - rigged, rng2 - random As software cant determine which hand will go to showdown / not it has to start with rng1 and then shift to rng2 post flop. So either software has artificial intelligence ( chances of which are very very very remote considering the number of hands) to decide which rng to use or it has to shift rng when one player goes allin.
Going by that conclusion, for hands in which there is a showdown but none of the player go allin before river, it has to shift rng at river or analysis of these hands will distort figures of rng2. Ok so analysing hands which went to showdown but there was no allin before river, comparing winning percentages at turn and river will point out the riggedness and will not go unnoticed in analysis.
Regarding ev check - you can do a stack/pot basis analysis for it, will require more effort though.
Posted by kraemer: 2)Just because a complete analysis MUST reveal rigged games an incomplete one must not.
When the actual data is infinite in a system like this there are no "musts", no matter what you do you will never get 100 % results. And going by your logic - all medical tests are void and all medicines are useless. And so on
Posted by kraemer: 3)And therefor the analysis done is not good enough to decide wether online poker is rigged, though it "prooves" ( there always is a close to 0 chance that you analyze a set of hands in which variance strikes ) that allin hands are played fair.
Atleast someone has proof ( in statistics this is significant enough) that in some aspects rng is random but theres no proof whatsoever that it is rigged but only possibilities and theories.
Posted by kraemer: 4) Though I tried hard I was not able to find another evaluation than the Win%
Because most people don’t have time to prove something they already know for people who wouldn’t agree anyway and make up another rigged theory. Testing for allin % is a proof that rng is capable of generating hands random enough and adding any other code/ rng to make it rigged wont go unnoticed by certifying agencies and licensing authorities.
Posted by kraemer:
Below is an analysis of 1 million hands from which the preflop allins were taken. First number is the number of hands. Second is the expected number of wins/losses, third number the actual outcome, then the differrence and at last the standard deviation. Ahead 36101 24002 24093 +91 88 Behind 33524 11726 11738.5 +12 85
As long as the results are within 2 standard deviations the results can be called "very close" statistically. In this case the analysis shows us a fair game. But still the poker room has a "margin" of 352/340 hands it could screw up without changing the reseults of our analysis.. You would still be in the 2 standard deviations You need to considder the results as accurate.
The deviation can range from -176 to 176 without implying anything else than a fair game.
Ok you have a really poor understanding of statistical concepts and you don’t seem to make any sense. I don’t know the study design of this analysis but don’t get how you got that standard deviation value.
Standard deviation - it’s a measure of the dispersion of quantitative data. ( Quantitative variable - which have numerical values, like no of hands won) It gives indication of how much the individual values in a data differ from the mean ( I don’t think I have to define mean now lol) So if u take a data with n variables ( quantitative and many of which might be repeated too e.g sample of 6 variables - 2,4,4,4,5,2) and calculate the mean of that, lets suppose mean = M
Now you plot a graph with variable on x axis and their frequency of occurence in data on y axis From the data you can calculate standard deviation ( not going to write the formula cos I don’t know how to write the symbols here and anyway its irrelevant), lets suppose standard deviation is S
Significance of standard deviation Now in the graph, area under the curve corresponding to values from M-S to M+S will have 68.3% of variables area under the curve corresponding to values from M-2S to M+2S will have 95.4% of variables area under the curve corresponding to values from M-3S to M+3S will have 99.7% of variables
Standard deviation is always mentioned along with mean, without mean it doesn’t have any significance. So ur standard deviation of - 176 to + 176 has no inherent value.
in the hand analysis you posted, the variable is qualitative - i.e. whether the ahead hand won or not. And you don’t get standard deviation for single study of qualitative variables. But like I said there might be some specific analysis that guy must have done and without mentioning it your arguments don’t make any sense.
The only standard deviation which could be achieved in this study is if we conduct an infinte number of similar studies for 36101 hands and then no of hands won will be the variables. And if the system is random, we will get a bellshaped curve with mean = expected value by pot odds i.e. - 24002 Though it beats me how could he calculate standard deviation from that single study. But anyway let me analyse the given results properly for you.
So 95.4% of times the no of times hand ahead won will be between 24002-176 to 24002+176 i.e. 23826 to 24178 out of 36101 hands or 95.4% times difference in hands won per 36101 hands will be 0.5% compared to expected So 99.7% of times the no of times hand ahead won will be between 24002-264 to 24002+264 i.e. 23738 to 24266 out of 36101 hands or 99.7% times difference in hands won per 36101 hands will be 0.75% compared to expected
Regarding poker sites using that 2 sd (352 hands) to rig the site First like jess said its because of normal variance there is a difference and sites cant use it to their advantage as it will add more to the difference. have to agree that if they somehow rig the software for 80% hands to win by 79% it might not be evident statistically but it wont make that much of a difference too.
And if they could mess with it would be 176 hands not 352, 352 hands will make 4 standard deviations. Since we are not comparing two values on either end of standard variation but 1 value on end of 2 sd and mean.
And they gonna mess with 176 hands outa 36101 hands which is approx 0.5% hands Now don’t tell me you call that rigged , lmao
Posted by kraemer: Remember that these are all in hands. Say You play NL100 then this could mean that the poker room has taken 352 times $100 from You to keep the depositing donkeys happy while still being able to "proove" the game was fair. My fault, but still enough to steal YOu $17600 in NL100 and make a poker room pro lough all the way to his free WSOP seat.
Ok lets assume for some unknown weird reason poker site don’t like you and decide to rob you of your money but within limits ( acc to your theory within 2 SD) So u play 36101 hands allin against player B ( who is loved by site for some reason) and instead of 24002 hands u win only 23826 So they take 17600 from you to give to person B But wait a sec - what did poker sites win??? NOTHING, they got same rake they would have got if u had won. And more importantly just see how much u won during that process - 23826 x 197* = 4693722 $ So u had a profit of 4693722 - (36101 x 100) = 1083622$ So for profit of almost 1.08 million they rob you of 17.6k ( 1.6%) and you are complaining, sigh
* assuming there is 3$ rake cap
Posted by kraemer: And on the other hand this means that the +91 deviation is out of the standard deviation. So by Your criteria I have just posted a 1 million hand sample that prooves the game is rigged , because the value should not be above 88. Considdering that we have analyzed 9 times the minimum sample this is statistically relevant.
NO, you just proved that you have a very poor understanding of statistics but still love making wild assumptions.
Posted by kraemer: You can send my Math Nobel Prize Money ( as I have just done something mathematically impossible ) to any charity of your choice !
If you can prove any day pokersites are rigged, sue them . You will be up for multi million dollar reward
Posted by kraemer: OMG - I am a Genius
To be honest, you are just making a complete fool of yourself. ------------------------------------------------------------------------------------------------------------------- And from my side - this matter is closed since you don’t have any backings to your theories except your poor understanding of statistics/ faulty logic and lame assumptions. Come back when you have something substantial to backup any of your claims.
And they gonna mess with 176 hands outa 36101 hands which is approx 0.5% hands Now don’t tell me you call that rigged , lmao
I call it rigged if only one single hand in a billion is not random.
And if you think its OK to steal from people as long as they are in profit you have a strange understanding of property.
Just as strange as not getting the point here, like that i was not saying they use 2 rng but just that a win% analysis of allins wouldnt spot that.
First You say its mathematically impossible. Then You admit they could take up to 300 hands from You. But thats not cheating because it's only a small number of hands. WHich of your posts is the one I should take seriously?
I would love to hear your reply when your boss cuts your wage in half, telling you not to complain because he is still paying You...
"From the day You learn to Read - You are going to be absolutly free! "
Joined: May '09
Location: India
Age: 37 (M)
Posts: 4873
Posted by kraemer: And they gonna mess with 176 hands outa 36101 hands which is approx 0.5% hands Now don�t tell me you call that rigged , lmao
I call it rigged if only one single hand in a billion is not random.
And if you think its OK to steal from people as long as they are in profit you have a strange understanding of property.
Just as strange as not getting the point here, like that i was not saying they use 2 rng but just that a win% analysis of allins wouldnt spot that.
"From the day You learn to Read - You are going to be absolutly free! "
messing of hands and stealing from people - which is just based on assumptions and delusional theories i never said they cud take 300 hands from you, i said based on your theory even if they take 300 hands it is too small it can be ignored.
you are still missing the point - its so small it is statistically insignificant
but like discussed before claiming they cud rig 0.0001% of hands and as there is no way to prove it , people like you will never believe system is random no matter what is done.
and regarding the stealing you have no proof and it is just your hypothesis. anyway if i had to invest 2 million and get 1 million sure profit but with a remote possibility that they might take 1% off from my profit i'll do it anyday. and you have a strange understanding of business. (google risk vs profit in investment, might help you understand it)
regarding ur job question - if ur boss takes 1% off ur pay? do u quit ur job?
Joined: Feb '11
Location: Canada
Age: 39 (M)
Posts: 1660
Wow, what happened to this thread. It just exploded with all kinds of conversation. I'm pretty far out of the loop (and too lazy to read it all, sorry at least I'm honest).
Is this another rigged thread? Or are we discussing legitimate statistics here? Feel free to ignore my post seeing as I'm not really playing ball here.
Joined: Dec '09
Location: United Kingdom
Age: 42 (M)
Posts: 2087
yes it seems a full blown rigged thread which has dragged in supernoob this time lol sometimes its just hard not to reply lol
what i find funny and hard to believe about these rigged things is there can be over a 1/4 million people playing on stars at one time playing many tables at the same time alot of them, im no programmer far from it but what alot of rigged people say is that the sites have fav players that win the money and others its rigged to lose. That is got to be some crazy s**t programming it baffles the mind and im sure would even for a really good programmer but i maybe wrong lol.
