Absolute Soulreading/Rigged thread #3

[lucky_mf]

[ QUOTE ]
Call me paranoid, but I do believe that all the posts here in defense of Absolute are coming from the same person. Certain accounts suggested I should use a standard deviation of 270 instead of 65 that I got from a winning player on pokerstars. They also suggested that I should use a t-distribution. Citation: "The t-distribution copes with uncertainty resulting from estimating the standard deviation from a sample, whereas if the population standard deviation were known, a normal distribution would be used." No one really seemed to know about statistics in this forum except drag who made some correct estimates and some other accounts that all were implying all the time by responding to each other that I should use a standard deviation of 270 instead of 65. Of course if you use a standard deviation that high the results get a lot more probable.

[/ QUOTE ]

You should not be using a SD of 65 - A SD of 65 may be appropriate for a player with somewhat typical stats, but a guy playing 90/80 or whatever is going to have a much larger SD. In absence of more information you should be using the sample standard deviation. As you are using a sample standard deviation, and not a population standard deviation, using the t-distribution is appropriate. In this case the distinction between using the t-dist and the normal distribution may not be very important as the t-distribution converges to the normal distribution pretty quickly (you don't need a huge sample before the differences become small).

Lucky