https://drive.google.com/file/d/18LojDj6Wj8KJ7Y1tahb9eqYhvbxuv01-/view?usp=sharing
Attachment Screenshot_1.png added
Attachment Screenshot_1.png added
Attachment Screenshot_2.png added
Attachment WF Optimization - Strategy 331259.sqx added
Attachment Screenshot_1.png added
Attachment WF Matrix - Strategy 53260.sqx added
Status changed from New to In progress
1) different results every time you run the matrix - it depends kon your configuration and number of possible total combinations. If you have 10 million possible combinations of parameters you need to use at least 5-10.000 max optimizations so that genetic engine can search the available parameters space. Even then, only 10.000 of all 16 mil. possibilities are evaluated, so there is a chance that the results will be slightly different every time you run it.
The key is to ooptimize as little parameters as possible, not everything, to keep number of total possibilities low.
2) Walk-Forward results display is confusing, we remade it for the new version. Until now it mixed result of one WF run with result of full WF matrix.
I'm attaching how the WF Matrix result looks now.
comment #5 - there is a % of how many conditions must pass in order for one WF result to pass. So one of the conditions could still fail and your WF result can pass - it depends on setting. If you want to force all conditions to pass set robustness score to 100%.
comment #6 - difference in number - we'll look at it, probably only different rounding used in both places.
comment #7 - PASSED/FAILES result in the databank is set after the optimization depending on your setting in Ranking. You can modify the conditions in WF Result screen as well, and result will be recomputed, but it will not change Filter result in databank.