>>/49266/
I think you get a good database that would be decent and covering most of the r but it would need a lot of extra columns besides raw data from a PCS or whatever.
It would be bad at outliers like a Mollemer but I bet it would optimize the backpack problem (put most stuff with limited amount of space or, for instance, get most points with the limited budget you have) for you so it would put in there Powless or Matthews, probably Christen too, it would surely list them high in EV before applying the backpack problem solution.
It also wouldn't just completely forget I have Barre picked and a shitload of points left to get someone better who is not on a downtrend, trends it would be especially good at.