So you think entirely endogenous statistical analyses are “actual proof “?You can make up a billion non-fact based reasons for player performance changes. Doesn't mean any of them are right. Or should be talked about like they matter in the absence of actual proof.
Probably 50% of reliever performance is small sample size noise. As evidenced by Walden approximating a viable major league pitcher for 2 months before his true lack of talent level took over. But that didn’t stop all the analytic folks here from signing the praises of the mythical Brian Bannister magical analytics bus and projecting those 2 lucky months Walden had onto some analytical breakthrough that had somehow eluded 10 to 15 other sets of managers and pitching coaches Walden had