QuoteDislikedthis is an old technique (like 30 years old!) and its prone to failure. this is addressed extensively in the market wizard books as well as fred gehms book. if you search for thousands of patterns, obviously a few good ones will pop up. but what have you done other than predict the past? if you read my article i wrote for S&C, youll see that my opinion is that you are going about this backward.
This is precisely the same argument used in Perry Kaufman's book "Smarter Trading." He goes into length about how patterns that were historically successful might only be successful because of the current trend (hence only successful because of it), and how to go about identifying whether that was the case or not.
His approach is similar to yours it seems, he starts w/ a pattern and then seeks to prove or disprove it w/ data, rather than starting w/ data and searching for patterns.