I am curious. On your http://TradingSolutions.com site, you display models and their performance.

Question 1) is this realtime performance or backtesting?

Question 2) Since you are showing examples (a handful of stocks) and not the results of everything, what is the average case performance of your stuff. For example, did you run the test on thousands of stocks? Or just the ones listed?

Question 3) For these stocks, how many different models did you run on each stock - in other words, is this the best performing chart from having run dozens of different models on the same stock?

Question 4) Are you showing results that are looking back with after retraining ("gee, this is what this stock would have done in the past, using the new model that we retrained to get rid of some losses"??)