Hello everyone.
I’ve been testing an algo with walk forward and I’m happy with the results. The variables will need to be refreshed every so often but I’m struggling to find the best timeframe to rerun the optimisation.
What is the best method for finding how often I need to optimise please?
I’m testing 5m timeframe back 1 year, should I change the iterations or how far back the test goes. Each test can take 10 mins so trial and error is taking ages.
Thanks.
What is the best method for finding how often I need to optimise please?
Depends of the results you have in your OOS periods.
Example, if you are optimizing one year with 4 IS+OOS periods through, each IS+OOS period would represent 3 months.
If you are happy with the result, then:
-> if your IS is 70%, then you should re-optimize each 2.1 months to trade the strategy 0.9 months
If not, try to change the iteration quantity and so on…
Thanks. I had it set to 5 iterations and each OOS represented about 6 weeks. So I should check it every 6 weeks then? This is what I thought but good to have it confirmed.