Schizo Greybox Trading

Can you be 100% free of curve-fitting? If so, please enlighten me.

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.
 
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.
 
Last edited:
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.
 
Back
Top