In addition to setting a max return to manage outliers, how can I set a minimum return? I don't see a way to do so in the dataset settings, but would appreciate any help!
What's the use case for a min return?
Thanks Marco,
Two key uses: 1) remove outliers from shorts 2) improve accuracy on longs
For 2) I want to make the system pay more attention to the long than the short side. This is a quick and dirty way to do so. Some of my factors work great for shorting (but don't benefit the long side much). I think the AI models are picking up on this and reducing overall error by making the short side very good at the expense of the long side (see lift chart below). I'd prefer to encourage them to do the opposite. My impression is that most P123 users are focused on long-only strategies and will also want to "tilt" their models to focus more on the long side vs. the short side.
