Understanding Walk-Forward Matrix parameters
I have few questions about Walk-Forward Optimisation parameters:
In Sample days: say – 400
Out of Sample days: say – 200
Does it make the whole testing time frame equal to 600 days?
Does SQ count days by calendar (weekends included) or strictly by the datafile?
How starting date for the next shift is calculated?
Is it end of my 400 days date or 600?
… and I do not quite understand Walk-Forward Matrix Optimisation parameters:
What means values of “In Sample days”: Start 20 and Stop 20?
I understand what increment value is: Adding 20 days to some date but without understanding of start and stop values I do not know how to apply the increment.
The same goes for out of sample.
I did read examples on the web site, but they somewhat outdated and do not cover my questions in details I need.
Calling for the masters, because some help in interpretation is needed. 😉
hi, have you checked this article? https://strategyquant.com/doc/article/walk-forward-optimization.html
Yes, I did. That article does not cover settings when days are used.
But I already figured it out. So, it is clear to me that the total length is a sum of those two values.
Do not remember about weekends. But it is not that critical. It is most likely based on what the data file has.
The story with WFM is also become more or less clear to me. It is still a puzzle to me what start, stop and increment means but lowering values of “Walk Forward Runs” variables I increased number of days in one single shift. That is what I wanted, because strategies based on D1 charts did not have enough data with default settings and required longer periods.
Viewing 3 posts - 1 through 3 (of 3 total)