From Russell Carleton at Baseball Prospectus on May 10, 2018:
Last week, I discussed how the stock story about a player having a good first month that contains the line “we’ve gotten to the point where Player Y has X plate appearances and so we can trust his stats” amounts to statistical malpractice. The numbers that mark those thresholds are meant to mark the point where the distribution of all players overall becomes “stable enough.” The problem is that all anyone wants to write about are the straggler data points on the edges of the graph. Those are the players having a really good (or really bad) first few weeks of the season. They are, by definition, the exceptions to the rule. They’re also the most interesting. No one wants to read about how Player Y is doing exactly what he did last year.
But that sets up a problem. The players being picked on the edge of the distribution are the ones who haven’t (yet?) come back to the pack. Obviously, something happened there. One possibility is that they got lucky. Another is that they found a new talent they never knew they had before. Maybe a little bit of both. If it is luck, luck is—by definition—random, so counting on it to stick around is dangerous. The player is likely to go back to what he’s historically been. If it’s skill, on the other hand, we assume skill is somewhat stable. In that case, we might expect the player’s newfound performance level will continue.
Read the full article here: https://www.baseballprospectus.com/news/article/39862/baseball-therapy-breakout-real/
Originally published: May 10, 2018. Last Updated: May 10, 2018.