I have noticed I new feature in the backtesting with variables… Walk Forward Testing. My statistic/numbers mind is very happy with it but I am not sure of its real application. How do you interpret/use it?
Also I am confused with some results. I make the WFT with two variables and the results are worse than the system with fixed parameters I had establised before. Isn´t it looking for the best combination of variables in each period (5 in this case)? That should be better than fixed numbers for these variables the whole time, isn´t it?
I just found this post in the blog about this feature:
https://www.prorealcode.com/blog/learning/strategy-optimisation-walk-analysis/
Thanks Nicolas. Anyway it focus more in the instructions to use it or how it works and I am asking more about the interpretation
One question… the statistics in the Overview (Gain, Max drawdon…) of the WFT how is calculated? Wich set of variables or period use?
Sorry I’m late. I made videos in French already, they are in the blog and on the YouTube channel, but I didn’t have enough time to make them in English before the international release of this new feature.
I plan to make them in the next few weeks.
How to use it efficiently needs long and detailed explanations..
Walk Forward looks at best variables for each iteration of the optimization period to get the most profit for the whole period. The out of sample periods give information about the fact that your optimized variables are robust or not for each in sample period considered (WFE must be >50).
there is a page in pdf pro order about it ….
FYI, Walk Forward will be available for live IG accounts by the end of this week.