Originally Posted by Spike View Post
Originally Posted by unowme View Post
Bayesian inference might allow APs to adjust their play based on observation of outcomes by uncovering a lack of true randomness in the game itself or by overcoming a lack of knowledge on the inner workings of the game.
.
I have always said the reason I can do this is because there's only 37 or 38 pockets on a wheel which isn't very many. If there were 60 or 90 I couldn't do this. I've gone to random.org and gotten random columns of numbers that are 60 or 70 or 80 numbers long and I can't do what I can do with the 36 numbers in roulette.
One method would be to make a histogram of the 38 pockets. If the frequency of occurrence of certain pockets are much higher than the rest and those frequently occurring pockets are adjacent to one another on the wheel, then you have identified a biased wheel and can bet that cluster of adjacent pockets. I have no idea if that is the method you use and am merely identifying one of the methods people use to make money on roulette. Clocking the decay rate of the ball due to friction is another, but most people need a computer for that.