Hi everyone,
Which filters or methods do you use to test the robustness of your models before going live?
Any feedback appreciated!
Hi everyone,
Which filters or methods do you use to test the robustness of your models before going live?
Any feedback appreciated!
This line could be added to your universe to get a more realistic idea of how your model will perform before going live:
Min (LatestActualDays, DaysSince (Max (LatestFilingDate, LatestNewsDate))) > 3
The goal is to correct for some of the delays in updating of FactSet data for small-caps. > 3 can be changed of course, and I do not know the best number to use. It probably depends on your universe, your beliefs about FactSet data delays, and the specific features you use.
I made some suggestions here: Break Your Strategy: How to Stress Test Your Quantitative Models - Portfolio123 Blog