Julien's Faster than Light Blog
Jump to top
Wednesday, January 19, 2005
from alec:
Just saw the "intuitive" explanation of walconpow. good. what about a
more technical explanation? why, exactly, is "con/4 + con*pow/2 a
good batting average predictor for slow guys"? and how do you come up
with your "hitters vs. pitchers" numbers? etc.
the technical explanation is proving to be more involved than i thought. basically, it's contained in the history of this blog, more towards the beginning when i was running regressions.
yeah. so that's where the batting average predictor came from. course, in the early days, i was using all kinds of wacky formulas (formulae?) for pow. so it may be hard to follow what's going on. but if you reed from the beginning, you'll see where the changes happen.
hitters v. pitchers. pick a percentage. wal, for example. let's say the hitter's wal is h, the pitcher's wal is p, and the league average wal is a. the formula is:
h v. p = (h*p/a)/(h*p/a + (1-h)*(1-p)/(1-a))
it's a mathematically determined formula, good for any percentage stat. bill james had some mathematician work it out for him. any statistician or probability theorist could tell it to you.
Comments:
Post a Comment
