fitness compute with OOS
2 replies
Charles Umani
3 years ago #269109
hi, i noticed my fitness is compute on the global calmar ratio and not on my IS calmar ratio. So during generation out of sample seems absolutely useless ?!!
it’s definitively not a real out of sample.
PS : for this test I’ve use only calmar ratio to compute my fitness
is this normal for you ?
tomas262
3 years ago #269289
Hello,
thanks, we will be testing this as I see the results are not positively correlated now. Note: only the IS period data is related to the fitness value.
tomas262
3 years ago #269305
The fitness function is not only affected by the Calmar Ratio value but also by the number of trades. If the number of trades is low (around 100) the fitness value is affected accordingly to reflect this. If the number of trades is significant enough then the fitness function is strictly determined by the Calmar Ratio value. Check the screenshot
Viewing 2 replies - 1 through 2 (of 2 total)