Quote:
Originally Posted by statmanhal
This almost seems like a trick question if I interpreted it correctly. If you are knowledgeable enough to calculate the standard deviation, I would think you are familiar with normal distribution probability theory, for 1000 samples allows for such theory to apply. But, I agree with the previous poster that the question is unclear.
Anyway, observing a sample mean 2.5 deviations above the population mean has probability of 0.6%. So, if the sample data were to be used to determine if the table is biased, you would reject the hypothesis that it was not biased.
This is a cross post from "Other Gambling Games" where he mentioned Bayes'. Since casinos have an interest in providing and unbiased wheel, a 2.5 standard deviation result is actually rather meaningless. It's far, far more likely that the result was due to luck than the player actually possessing an edge. I would argue it depends on the nature of the observed flaw, though. If it is easy to spot then it's likely nothing. If it is extremely difficult to spot then it is more likely to be meaningful.
