I don't think research like this should be dismissed, but this guy is no Nate Silver. The author missed a few key statistical concepts, like overlapping contexts (multicollinearity) and not enough contexts (omitted variable bias). Example: you can have a scenario where the passer could be in multiple contexts identified by the author. Being "pressured" and "out of pocket" are not necessarily separate events, but the author is attributing the ball placement to singular events. Poorly designed study imo.