There are books out there documenting 24 hours, 48 hours, and 72 hours of decisions at the craps table and roulette wheel. Do you believe a set of rules on those results would be valid?
In any case, the more rules the system has, the more likely it is to be data mined. Because each rule could have been created to account for a particular situation or situations found in the back-tested data. If you break the data up into 21 groups of 400 trades, how consistent is the performance between those groups? That might also give you some idea of the level of data mining. For example, if 2 or 3 groups represent the majority of the profit, the system probably isn't a very good one. On 4/17/08, si00si00 <[EMAIL PROTECTED]> wrote: > > My question is, does anyone know if the data-mining bias can be > considered irrelvant when the sample size is so large? (in this case, > the sample size is roughly 8400 trades). Put another way, with so many > observations, how many different rules would have to be back tested in > order for data-mining bias to creep in?
