Near the end of last year, I went back to class to learn how to program in VBA. Just yesterday, I finished a program to calculate a mechanical mathematical system similar to what you posted. I then used it to backtest SPY going back to 1993. First few runs kind of verified your statement that (for some unknown reasons), the best exit/stop is equal % and at a modest %. A tighter stop actually made things worse. Also, adding a holdover to the next day improved the outcome. The system generated only a modest CAGR profit over 25 years.In the last 21 years of trading mathematical mechanical systems, I have traded and forward (actual trade) tested every possible combination of profit goal versus stop amount, and each combination was traded and forward (actual trade) tested several years per combination.
The combination that works the very best for my annual profit is:
Modest Profit with Equal Value Stop amount.
This setup produces the highest W/L ratio of any Profit/Stop combo.
Overall, the profit/loss is very sensitive to whatever assumptions I put into generating IV, to initial option values, to assumptions made on profit/loss exit, time to expiration, to calendar period....
I suspect that either my program still has bugs or the system is too dependent on situations not within my control (I maybe missing some rules). Or perhaps I am fooled by randomness again?