F?!
Jb
Big fan of Ralph's approach. But here's a question.
When you analyze some results, these are results generated with a given f. The f you were using in those trades.
The optimal f then, does this take into account what f was used to generate the results, by dividing by biggest loss?
I don't see how that can work. Surely you'd need to factor in what f you are currently using and calculate optimal f from that, not from just the trading results.
Even if it does take that into account say you use 1% a trade, the approach doesnt seem to take into account that you might have say 20 of these trades open simultaneously so dd could be 20x higher than calculated on a per trade basis. As far as I can see you need to use account dd not realized loss and number of simultaneously open trades seems very important. Even with one instrument one might have 5 or 6 trades as you enter into a position or try to move average position closer to current price.
Maybe I miss something or over complicate ?
Don't get me wrong I'm not knocking it. I think it is fantastic but seems it could either benefit with being more realistic or more clear.
Since I am not in US I have never attended the course so I can't make head or tail of the java or r files since I lack an example data set so i'm doing it the old way and doing one condition at a time and checking dd. Rather than finding optimal f, how would i find optimal leverage?
I'm using a nice system that has made about 5-10% a month consistently for 10'months now with real funds. It is in desperate need of optimization. Right now I'm using sub accounts and changing f in real trading to find optimum. It's slow but I can't get useful numbers from optimal f as it stands. Any pointers welcome.
Regards
Jake
Sent from my iPhone