Proceeds from the ads below will be donated to the
Bob Wuesthoff scholarship fund.
Sunday, February 08, 2004 |
Rob Neyer, Witch Doctor and Sycophant
Jon over at Dodger Thoughts points out a Rob Neyer story discussing the A's poor choice of platoonmates for Scott Hatteberg. Thanks to groundbreaking 1988 work by Bill James to which Neyer's column refers, we know ability (for righty Karros) is more accurately predicted by righty-righty matchups than vs the much smaller sample size lefty-righty matchups. I won't bore you with the details; read Jon's or Rob's column if you want them.
Here's the wow part from Neyer:
MGL gets into things like "correlation coefficient" and "y-t-y correlation," neither of which I would explain in this space even if I perfectly understood them ...Neyer makes his coin on a reputation as a stathead, a guy who believes numbers tell the difference between wine and hogwash. But here he admits he is as a first grader trying to integrate a fifth-order polynomial. Does anyone believe such stuff? Imagine a lawyer claiming ignorance of the murder statutes, a chef not knowing how to dice tomatoes, or a politician needing instruction on how take a bribe. The calculation and meaning of Pearson's r -- correlation coefficient -- is first-year statistics. With this admission, Neyer joins the ranks of witch doctors, a sycophant and publisher of press releases for any propellerhead with a sufficient audience.
Newer› ‹Older
 |
Post a Comment
Note: Only a member of this blog may post a comment.