THE BOOK--Playing The Percentages In Baseball

Repeating for outfielders:

CF: 14-0, 16 middle = 0.733
RF: 3-4, 23 middle = 0.483
LF: 3-16, 11 middle = 0.283

RF v LF head-to-head: 21 better fielder in RF, 5 better in LF, and 4 close.

I think the only thing keeping the LF afloat are the backups in LF.  (We talked about this in the past.) LF is the new DH.

So, in terms of adjustments, and centering at 0, Fans are saying:
+7: CF
-1: RF
-6: LF

Something like that.  (Just the regular players.)

Combining the two groups by adding 4 runs to all infielders and removing 4 from all outfielders, we get:
+8: SS

+3: CF
+3: 2B
+1: 3B

-5: RF
-10: LF

Basically, they are all within the 2.5 runs margin of error from my generic spectrum numbers.

Would you feel at all comfortable converting the fan scouting report ratings to a UZR/150 type number?

Let’s say you think the best CF is +25 runs and the best CF rating is 4.6, would you feel comfortable having a mapping function that went from :

f(rating, #samples) = UZR/150 number.

doable?

vr, Xei

I’ve done it in the past (see link), and I will *definitely* do it this year:

http://www.tangotiger.net/scouting/scout_Results.html

Do you have splits for batted balls to the outfield? Like, 40%/30%/30% for balls batted to LF/CF/RF? Whatever happened to putting the worst outfielder (albiet w a strong arm) in RF?

#4.
Great, thanks!  Is it possible to make it downloadable in a csv or some other similar type of format?

Tango/#4

I think Rally (in the past, on his old blog) did something similar with the Fans’ Scouting Report in comparison to DER.

http://lanaheimangelfan.blogspot.com/2008/10/evaluation-of-defensive-projections.html