ok college/uni guys, failed traders, poor people in general here is a simple way to make a little money at your local roulette table and its an edge which i have no problems in disclosing because itll never ever dissapear. The system involves playing roulette. ah yes so you say that the game is perfect? the casino always wins? and that u cant gain an edge? i am going to reveal an edge to u that i figured out last week. simple statistics really, any comments or system criticisms are appreciated. I am assuming that there is always dealer bias on a roulette table. An assumption which is not unrealistic. The question is how can we detect dealer bias? the answer is simple. 1st year statistics is all you need. ok here it is: THE SYSTEM Go to your local casino, wait for change of dealers at the roulette table. Then take a sample of his/her first 15 numbers. Construct a test statistic. Only bet on the sections 1-18 or 19-36 which pay 2-1. U have just under 48% chance of winning thanks to the 0. Null hypothesis: u = 0.48 u>0.48 construct a test statistic (X-Xbar)/standdard error. Find the critical value under a t-distribution with n-1 degrees of freedom. The alpha level u choose will be up to your own personal risk preferences. if there is a bias then bet on that bias.