from Amazon
Thursday, January 24, 2008
K/BB ratio or K/BB differential?
This is a cut/paste from a Baseball-Fever post I made, which I don’t think I posted here. In response to someone who said “Although this would make a 300 K/150 BB guy better than a 195 K/50 BB guy, which I’m not so sure I agree with”. Here’s what I did:
Since 1993, of all pitchers with at least 12% of their TBF as walks, the top 10 guys in K/TBF gives you this:
13.5% BB/TBF
26.5% K/TBF
3.72 ERA
Of all pitchers with at most 5% of the TBF as walks, the top 10 in K/TBF gives you this:
4.4% BB/TBF
16.7% K/TBF
3.98 ERA
So, the first group has a 2/1 K to walk ratio, while the second group has 3.8 to 1 K/BB ratio. The first group has triple the walk rate as the second group.
(Exactly matches your parameters.)
The differential is slightly higher for the first group (0.13 per TBF compare to .123 per TBF).
The actual ERA is a testament that the differential works just fine.
***
The first group: .33 extra base hits per hit.
BABIP: .281
BA/OBP/SLG: .219, .330, .352
The second group: .33 extra base hits per hit
BABIP: .294
BA/OBP/SLG: .269, .304, .429
***
By the way, I used a minimum 1000 TBF since 1993. That’s 728 pitchers in all.
***
The K, walk differential works just fine. It’s why FIP works so well, which is basically the K walk differential, with the HR thrown in. With the HR removed, K minus BB is the best-fit. There is a dependence of HR to K And BB, and if you take out the HR, then you need to rebalance K and BB. And as it turns out, you get the best fit giving K and BB exactly a coefficient of 1, as opposed to FIP where you give 50% more weight to the walk.