Not yet, but I’m working towards that (in a way). But i’d be interested in hearing others’ thoughts on this.
Here’s some of mine, in no particular order.
It’s summer, I’m very old school, so I really tend towards “Sell in May and go away, come back in on St. Leger’s Day”. While it’s not exactly the best strategy, it does make for a pleasanter life
Right now I have about 14 different different, mostly technical analysis, signals, and I’m putting them through Numerai individually to see the effect of their feature removal, etc.
My signals are actually calculated in terms of relative returns, and also contain a number of tickers that aren’t in the live_universe. The submissions for Numerai are taken from those, cleaned up, and normalized to (0,1), but it would be the relative returns I would trade off of, not the submitted ones.
I had about 5100+ tickers on my last submissions, and I’m not rich enough to trade on all of them Or even very many of them. But most anyway are clustered around an expected return of 0, and/or have a high degree of uncertainty (i.e. conflicting results from different indicators). So those I would ignore.
The ones I am most interested in trading are those that lie roughly in the top or bottom few percent consistently across indicators. I check these out by hand for now, and (unsurprisingly) they are better than I am on my own. That is, of course, a very low bar, but it is promising.