Here are some more charts from Wednesday's session. This time from the first few hours of trading. When the Primary Momentum (based on the tradable) moves up or down and the Secondary Momentum (the model) does not confirm the move, price will adjust to liquidate the difference. As I mentioned before, this is not a complete trading system. The two oscillators show that inefficiencies in the market exist and that price moves in the direction of the model to close the difference.
I agree, backtesting is just another way of curve fitting - selecting trading rules that perform well on a particular data segment. Seven years? LOL. The model ( with the current frequency (that can be adjusted)) would generate some 100k signals or more. That would be very statistically significant. Actually, I decided to post the charts after monitoring the model in real time for long time.