If I select:
Optimization period : 20
Test Period: 10
The the optimizer will use the first 20 days to optimize then test on the next 10.
It will then move forward 30 days(Optimize +Test) to the next period and again use the first 20 days in this period to optimize.
My Strategy uses 40 bar requirement to start. This uses a significant amount of the new test space on each walk forward.
If what I understand is correct it would be helpful if the walk forward only advanced for the length of the test period (10 days in my example) rather than 30 days.
You would then be optimizing on an overlap of the old test period and the 10 days of out of sample. I think this would give you a much better understanding of how the indicator will work on the underlying data.
This would be more inline with what most people do in real time.
Comment