I am not sure if you are using indicators in your system. For instance, if you are using 2 MAs and you optimise the parameters to 'suit' the market, then it will likely fail sooner or later. I would say most strategies have some degree of curve-fitting to fit the market's behaviour. Without it you will not gain statistical advantage. However, the curve-fitting should be made in such away that it exploits the overall market behaviour for a significant amount of period where it gives you a concrete statistical advantage. Curve-fitting on a limited amount of period will only reflect the market's behaviour in that particular period and will not likely work in other periods.
Thanks. Those are all valid points worth remembering. I did make a minor tweak that is helping me stay out of the chop zone for the night. Truth be told, my understanding of backtesting is rather limited. Sometimes, I wonder if it's worth spending so much time and energy for something that has no guarantee that it would work.
You're on the the right road. Truth be told, you're mastering a great skill. Anyone who's able to religiously stick to a system is light years ahead of most of the people out there. Just stick with it.
OUT 2008.75 -0.75 CUM P/L: 3.00 Wrapping up for the night. Although it ain't as good as my discretionary trading, the result wasn't too shabby.