I hope there's a way to estimate from empirical data how much of the time he's actually right, using a boolean algebra model or something.
Because I'm starting to see a phenomenon, the oppisite action of the decision you make is usually the correct one.
Shouldn't profits against trading him be good enough indicator on that? But yeah i agree, empirical data would be nice. In fact some sort of model that's tracking all the influencers and their trackrecods out there would be very cool.