Pure Tin Foil Hat Stuff - U wuz warned

[FlFishOn]

What's that odd sound? I know, it's some one obviously uninformed and talking out their ass.

[FlFishOn]

"Which seems perfectly reasonable. "

Let's have none of the 'reasonable' sh+t here. Flames, ill informed rants OK, but forget reasonable. This is mostly a bunch of poorly schooled 20sumthns and I'm not looking to wise them up.

[kdog]

[ QUOTE ]
I haven't mentioned this in many months

[/ QUOTE ]

Yeah, and I must say, you hooked more than you usually do with the troll this time.

[ QUOTE ]
We get new folks here. A few actually have some objectivity

[/ QUOTE ]

Objectively, over the years here, all posted statistical analysis, as well as the more logical arguments, support the not rigged side. But you already know this. So, this time I'll issue the challange you never accept...Lay out your data so that we may prove or disprove your assertion. Otherwise STFU.

[Catt]

[ QUOTE ]
This comment more or less proves you unqualified to comment.

[/ QUOTE ]

I'm not the one who had to make a trip to the public library to undertake a "painful reacquiantance with statistical methodology" in order to draft an astonishingly ridiculous message to "a poker writer." I am not the one who starts the message with "I have gathered a lot of data" and then goes on to cite his 419 SNG hands. I'm not the one who stated essentially "My data show a winning % of X; there is a 95% certainty that this data does not show a winning % of X+1." I am not the one who determined the expected outcome by a PokerProbe simulation and "some random sampling" and then acknowledges that it needs to be more "rigorously explored." I am not the one who, despite acknowledging that "I don't get a lot of data points for rigorous statistical analysis" states "the data is clearly biased." And finally, I am not the one who inserts gratuitous and ridiculous statements like "While this may strike the layman as inconsequential, when subjected to statistical tests these are troubling numbers indeed." Were you smoking a pipe when you drafted that part? Did you put on a stiff collar and dip quill to ink?

You had to go to the library to "refresh" your understanding of confidence intervals, and you adopt the posture of a statistical maven hard at work on a Cray supercomputer crunching numbers that a layman would think inconsequential -- oh, to be one of the cognoscenti.

I'm glad you found a hobby that you enjoy. Keep crunching the data! But stick to the random card space at PP -- comparing the "work" you are doing to a political poll by Zogby or others indicates that you may need a few more trips to the public library.

[FlFishOn]

I'm gonna mark you down as 'no opinion', OK?

[Aytumious]

The main problem with these type of idiotic and slanderous posts is that integrity is such an important concept for successful online gambling and people who may otherwise decide to give it a shot may read this type of drivel and decide perhaps it is not such a good idea. Provide at least something marginally interesting in the way of proof -- and no, your preliminary data is not useful -- and then we can actually have some sort of intelligent debate.

[J_B]

[ QUOTE ]
Proposition: Party and skins deal nonrandom turn and or river cards some small % of the time to benefit underdogs thus sending chips from good (winning) players to fishies.

Rebuttal: There's no way a bunch of programmers could keep such information under wraps. The dirty secret would get out (there are others, most are spurious or logically flawed).


[/ QUOTE ]

Ok, If you are a good player then move to another site. If you are a fish, keep at party please. more fish for me either way!

[jedi]

[ QUOTE ]
[ QUOTE ]
And they have a margin of error of 2-4%, too. Your sample size is even smaller (in a domain where the standard deviation is higher), so you'd expect to have a larger margin of error. Suddenly, your numbers end up with in that margin. Oh noes!

Just because you drew a sweeping conclusion from 3 hands doesn't make it right, or even well informed (hell, even somewhat informed). In fact, people have been shown to find bias where there isn't one.

Show me 4,190 hands, or 41,900. Then maybe you have something.

[/ QUOTE ]

He isn't saying this is evidence of rigging though. He is just saying this is worth looking into with larger samples. Which seems perfectly reasonable.

[/ QUOTE ]

Well, then let's see the larger samples before he says anything. One side has already shown random data, (if not fair) with a fairly large sample size. This guy has shown us under 500 hands. Let's get some real data to debate. I'm all for it. I don't think it can be done, but if it can, let's see it.

[Orpheus]

If we look at only one TXHE hand and the board, there are (52*51 * 50*49*48*47^46) deals, or 52!/45!, divided by (2)*(5*4*3*2*1) order permutations (a hand is functionally identical, regardless if the order of the cards in it, so there are 240 unique deals that boil down to the same hand for a single player) That's approximately (8.0 × 10^67)/(1.2 × 10^56) divided by 240 = 667 billion/240 =2.778 billion distinct outcomes. That's about 25 times the number of US citizens over 18.

If we look at all 10 hands at a full table, it's 52!/27! (27 cards remaining undealt out of a deck of 52) unique deals divided by (2)*(2)*(2)*(2)*(2)*(2)*(2)*(2)*(2)*(2)*(5*4*3*2*1 ) order permutations. That's a stunning 8x10^39 unique deals divided by 122880 order permutations -- for atotal of 6.5x10^34. That's more than 1000x the total number of PROTONS in the entire adult US citizenry.

And you want to assess that mega-universe with 419 hands?

Answer this: how many coin-flips (the simplest possible test case) do you think you'd need to know if a coin was biased, just by looking at the number of heads? Unlike poker hands, there are only two possible outcomes, not some unspeakable number. How many flips would it take before you'd have a reasonable expection of 50/50 in your test

Do you think you could do it in 10 flips? Absolutely not. Suppose the coin is exactly fair. There are 1024 possible outcomes of 10 coin flips, and less than 1/4 of them have exactly 5 H and 5 tails, while 220 combinations have 6H and another 210 have 6T. Would 100 flips be enough? Well, of all the equally possible combinations of 100 flips only about 7.9% have exactly 50 H/50T. In fact, th larger your sample size, the LESS LIKELY it becomes that you'll get exactly 50/50 in the sample size (though it does grow more likely that you'll get CLOSE to 50/50 results)

My numbers may be a wee bit off, because the equations involve terms with values of over 10^100, which overload my trusty old scientific calculator, but they are close. You can look up "binomial distribution" and do the calculations yourself.

My point is this: even with something as simple as a coin flip, small deviations from the theoretical mean are EXPECTED. They are not suspicious, because, even in theory, YOU CANNOT EXPECT TO GET Exactly THE THEORETICAL MEAN. Increasing the sample size actually DECREASE your chances of getting that theoretically perfect value (which peaks at a sample size of 2, believe it or not!). Standard deviation isn't an "error", it's an essential and real property of any statistical population. To determine if a coin is biased, you must calculate the mean and SD of the ideal result, and then separately calculate the probability that your observed results came from a 50/50 coin as opposed to a 60/40 , 51/49 or 50.1/49.9 coin.

Poker hands and outcomes are far more complex (many more unique outcomes, and much higher variance) than a coin flip.

Your result does not even SUGGEST that there is rigging. It is completely consistent with a fair game at your sample size. I don't expect you to understand or accept this, but the number you got is, statistically "spot on" for a FAIR game at your tiny sample size. If you take similarly sized samples of hands from any site you trust, or from hands you deal yourself with physical cards, you will see LARGER disparities in MOST of your 419-hand samples. If you forced a statistician at gunpoint to say if this result supported a fair game or a rigged one, s/he'd have to say it was consistent with a FAIR game at this sample size. (But you'd need a gun to get them to express a professional opinion on such a tiny sample

Don't take my word for it. Deal yourself 500 hands. Record theem and calculate the results. You'll see. If you don't like the result you get, try it a few more times. You'll begin to see the light long before you get a single set that's as close to the theorectical expected value as you demand.

If that's too much work, try flipping a coin. A 100-flip run should take under 3 minutes, so you'll easily have time for several runs. Report back and let us know how often you came within 1% -or 2%- of your expected value for even this umtimately simplified example.

[strogrules]

My wife and I have often had discussions as to wether the cards on poker sites are rigged. I dont know if this has been touched on by other people, but why would poker sites who are raking in millions of dollars from rake going to risk their reputation by rigging hands. Its a lose lose situation for them.