Statistical/Probability Analysis How-To

[mtardif]

Hello,

I was wondering if anyone was willing to share their statistical/probability analysis methodologies with us? What data do you collect, how do you analyze it, how do you go about filtering out the garbage trades, etc to finally arrive at a tradeable system.

And do you look for a natural bias of probability, or do you create one from the data? What I mean is, does the probability have to be greater than 50/50 "naturally", already existing in the "natural" price action, or do you create a greater than 50/50 probability by using indicators?

I'm having some difficulty producing a system that performs as well outside the test period as it does inside the test period. I figure it's my analysis methodology that is flawed. It's probably some minor detail that I screwing up or not paying attention to, but I guess I need someone else to tell me what that is.

Thanks.