I have started playing on a small US site and had some strange results. It seems like my good hands are not holding up as often as they are supposed to. I want to use my hand data to test the "randomness" of the cards.

Using only hands where I am heads up, all in pre-flop, my hypothesis is that my average equity over these hands (as computed by pokerstove) should roughly equal my win % in these hands as played. So that if I get it all in preflop with AA against KK 10000 times, I should win about 8000 of those hands.

I have two questions: First, is this a valid hypothesis, or is there some variable I am missing. Second, if valid, how many trials would it take to achieve statistical significance? If the numbers diverge, what amount of divergence and over how many hands until the results could no longer be explained by short term variance?

Thanks in advance for any thoughts.

Hello,

This is a great question to ask in the probability forum. There are some extremely shrewd math guys in there who could help you with this.

I have started playing on a small US site and had some strange results. It seems like my good hands are not holding up as often as they are supposed to. I want to use my hand data to test the "randomness" of the cards.

Using only hands where I am heads up, all in pre-flop, my hypothesis is that my average equity over these hands (as computed by pokerstove) should roughly equal my win % in these hands as played. So that if I get it all in preflop with AA against KK 10000 times, I should win about 8000 of those hands.

I have two questions: First, is this a valid hypothesis, or is there some variable I am missing. Second, if valid, how many trials would it take to achieve statistical significance? If the numbers diverge, what amount of divergence and over how many hands until the results could no longer be explained by short term variance?

Thanks in advance for any thoughts.

its P-p/(sqaureroot)p(p-1)/n

then we can take the z value we get and look up this z value for statistical significance.

Thanks ejames209

winni could you possibly define those variables for me? Thank you.

https://stattrek.com/online-calculat...omial.aspx#faq

on this page, you will find the basics you must understand. also it has a calculator, that you can plug your values in.

This has come up before, so I'll post my standard reply.

So I recommend:

1. Import your HU AIPF hands into a spreadsheet
2. Enter the equities from a poker equity website (if they are not already in your database)
3. Perform the above calculations for Expected Number of Wins and Standard Deviation of Wins
4. Tally your actual number of wins (consider a split-pot to to 0.5)
5. Calculate how many standard deviations your actual wins are different from the expected number of wins
6. Convert this to a "probability" using statistical tables/website (this step is not absolutely necessary)
7. Post the results in this thread.

Let us know if you have any questions.