Hi, just started staking my first model, fingers crossed!
I was wondering why staking is applied so much later than predictions, which are already made on Monday. Doesn’t this put the metamodel at somewhat of a disadvantage as it is acting on data delayed ~3 days?
I know the time horizon is long, ~4 weeks, but still, large movements can happen in 3 days. And there are probably other hedge funds using machine learning (perhaps on similar underlying features that were used to generate the Numerai features) that apply their predictions immediately.