Hi Myst,
I hope you will be feeling better soon.
I have shared two charts of Tezoz (XTZ) starting 14 6 20.
In the first, I optimised 15 9 20, using 3 months of data and then walked forward month by month without re-optimising the settings to see how well XDEV would continue to do in a scenario in which no further amendments were made to the settings.
It is reassuring to see a continued out-performance versus buy & hold using the initial XDEV settings, from using 3 months of data. XDEV performed well even when the settings were not changed.
In the 2nd sheet, I re-optimised the settings on 31 12 21 utilising the full strip of data back to 14 6 20.
Although the results in the first un-optimised play are impressive, the re-optimised results are more incredible still.
The question is whether re-optimising every month (or making small incremental changes weekly) brings one to this end result slowly over time?
I will, perhaps, try to do the same “walk forward” experiment again, this time re-calibrating the settings each month using the data from the months before.
Do you re-optimised over an (expanding) full length of data when you trade?
Do you re-start a new chart using 3 months (or 2 x buy or sell signals) at some point?
Have you observed whether it is better to use longer strips of data?
Kind regards,
Alex
AI
