Cool stuff. One big caveat I have about applying this to Estrada is that even though he does indeed have a good changeup he, unlike the players examined in the article, doesn't actually have a track record of major ERA/FIP/xFIP discrepancies. His career ERA/FIP/xFIP rates are 4.00/4.13/4.01. There's barely any variance at all. I just seems more likely that all were seeing right now is a disproportionate amount of outs on flyballs over a small sample giving the illusion of a FIP beater.