What I don't love about analyzing relievers with those metrics, just like any other, is that the small sample size makes them noisy. If there's a week in which a reliever happens to face 10 struggling hitters, and 8 of them are susceptible to the 4-seamers up in the zone, it can appear to create patterns in the pitcher's season that aren't really there.