1. Do you have any guidelines of re optimizing your auto trading model ?
2. Re-optimizing implicitly means that you are assuming that the last data gives the best change of predicting future movements. Then a very important question is what is “the best” set of data ?
I never seem to settle on a regular best way of doing anything! 🙂 Every week nearly I am trying a new angle!
Your questions prompted me to check recent performance (past 4 weeks) of the version I posted here
https://www.prorealcode.com/topic/di-tema-trendfollowing-strategy-on-dax-5-min/#post-106645
I noticed from the equity curve that most of the gain was during that long downtrend (up to Feb 19) so it is understandable why the strategy may not perform well in the last 4 weeks choppy sessions?
Anyway another new angle today … I optimised over 50k bars and then backtested over 100k bars and attached is the result.
So in answer to your Q.2 above … I don’t re-optimise over a recent period only, but I do want strategy settings to suit recent market movements so when performance drops then I re-optimise and put emphasis on recent periods … using Walk Forward as required.
Mods … should not this Topic be in the Pro-Order Support Forum?
