What is better for strategy building: Longer IS period or More trades?
3 replies
murty
6 years ago #192049
Is it better to create strategies with:
- Longer In-Sample period but fewer number of trades, OR
- Shorter In-Sample period but with more number of trades?
Example:
Let P: IS Period in bars, N: number of trades in IS period
Strategy S1: P1 = 1000 daily bars, N1 = 300 trades, Net Profit = 5% per year
Strategy S2: P2 = 400 daily bars, N2 = 900 trades, Net Profit = 9% per year
For a profitable and prosperous future, would you rely more on Strategy S1 or S2 ? Of course, we cannot rely on the one that gives the highest In-Sample profit 🙂
tomas262
6 years ago #192083
Hello,
900 trades is pretty good statistical sample but still 400 daily bars represent about 1 and half years I consider this to be too little for having reasonable statistical significance. What happens to the strategy when tested on previous 400 bars? Does it stay profitable?
murty
6 years ago #192365
No, it does not 🙁
So, S1 or S2 ?
tomas262
6 years ago #192616
No, it does not
So, S1 or S2 ?
I would like to see it profitable at least after re-optimization. If both strategies pass the robustness testing I would keep them both. Still as I want only the best I would let them trade minimum lot-size for some time
Viewing 3 replies - 1 through 3 (of 3 total)