Yes. More filters usually means more backfitting, which can create great results from past results, but not so good results on future races. A recent quote of Bhagwan's is spot on - "trying to find patterns is like analysing shifting sands".

Filters need to be created out of a logical process first, and then applied to the data, rather than randomly trying things and keeping them if they happen to increase the profit over a set of past results.

Shaun: "Guys the Holy Grail is out there and being used by a select few that don't post on here" ... any more details on that one??