Why does Pujols regress to the mean?

Like I promised in the thread about relief pitcher WAR, an article about regression to the mean that addresses the problem faced with regressing players who are only at this point potential major leaguers.

over 2 years ago cwyers 5 comments 0 recs |

Story-email Email Printer Print

Comments

Love it Colin

F*** Billy Beane... actually, I kinda like Holliday

by vivaelpujols on Aug 7, 2009 1:47 AM EDT reply actions

I was reminded of your regressed bell curve as I was reading this article by Dave Allen:
http://www.fangraphs.com/blogs/index.php/making-it-work-with-lots-of-whiffs/

Is the aesthetic similarity a coincidence, or does this sort of curve persist across many different stats?

by Tommy Bennett on Aug 7, 2009 5:38 PM EDT reply actions

Yeah, it's pretty common.

The central limit theorem explains why it crops up so often. (Which is convenient for us, because it lets us get a lot of use out of the normal distribution.)

I has a blog. | Ask a sabermagician!

by cwyers on Aug 7, 2009 5:59 PM EDT up reply actions

Either I wasn’t clear above or I don’t understand the central limit theorem (about equally likely).

Does it explain the asymmetry of Allen’s curve and your shifted Pujols OBP curve?

by Tommy Bennett on Aug 7, 2009 6:52 PM EDT up reply actions

Ah.

Dave’s curve is asymmetrical for a much simpler reason – because one cannot have a contact rate higher than 1. (You can certainly have a contact rate below .7, of course, but you aren’t very likely to stick around the majors with one.)

I has a blog. | Ask a sabermagician!

by cwyers on Aug 7, 2009 10:14 PM EDT up reply actions