Metrics for picking good heroes are poor #21
Description
My example team was: Anti mage, invoker, Axe, Wisp, and Bounty.
Every hero that was recommended has a better than average win rate vs these heros, but a resulting win rate of less than 45%
The "bad" picks all had a disadvantage vs those heroes, but all had > 55% win rate anyway
Realistically the picker should probably highlight the heroes with the best win rate vs those 5 - not the best change in win rate
