from Amazon
Tuesday, January 05, 2010
If you have UZR and TZ, should you EVER use FRAA?
Steven in the BPro comments wrote:
Two things at work in the player comments as far as the fielding metrics: Christina and I encourage the writers to debate with the stats (all the stats) where appropriate. Second, I’m not really satisfied with any one fielding metric and consult a range of them, and I know the others do the same. Until a grand unified theory of fielding stats comes along, that is going to be an area where what the writers see and what the stats say aren’t going to match up every time.
There is nothing, nothing at all, that FRAA adds to the discussion, if you have UZR and TZ. FRAA is a subset of these two. It would be like using ZR and UZR. ZR is a subset of UZR. Once you have UZR, you no longer need ZR. A similar analogy would be to still use OBP or SLG, if you have wOBA or EqA. wOBA and EqA already combine OBP and SLG. Therefore, you don’t need those two, unless you specifically need them for their individuality. FRAA has no such individuality.
So, guys, please, stop using FRAA. It adds nothing to the discussion.
The question is TZ v UZR. While accepting that UZR is better overall (because it uses more granular information), if there is something that TZ does differently from UZR, then that’s your case for using TZ. That’s my case for WOWY: I actually look at the identity of the batters and pitchers, and I rely on the fact that over a few years, the stringer bias will overcome some of the signal.