Which WFM filtering is best?
2 weeks ago #274075
1. About WFM filtering: which is better, recommed WF conditions or WF stability of avg. profit per day>50% (from SQX video course)? Or any other more suitable filter conditions?
2, I think WF score is not very good to judge the robustness of the strategy, because the robustness of the original strategy is difficult to have a unified baseline (for different original strategies), and the score can only reflect how much better the optimized out of sample is than the original strategy ( Or worse), does not directly and accurately reflect the robustness of this strategy itself. I think WF stability seems more convincing to the robustness of the strategy. Don’t know if my feeling is right?
3. In addition, what is the calculation formula for WF stability of avg. profit per day? Not very sure. I don’t understand SQX’s tips: performance in run vs in optimization part
1) you can check for stability for different metrics. Mostly the net profit and avg trade are used and make sense
2) it is only one metric to measure robustness. It is important to do various other Rtests. Not use just a single one to evaluate robustness
3) there is an extensive descriptive article regarding WF metrics and evaluations – check https://strategyquant.com/doc/strategyquant/description-advanced-walk-forward-values-can-used-filters-databank/#wf-stability
Viewing 2 posts - 1 through 2 (of 2 total)