clock menu more-arrow no yes

Filed under:

Pitchers' K/BB Ratios....and I'll get around to the Astros

New, 5 comments

I have always enjoyed reading Bill James' articles, partly because he is a good writer, but also because I like the innate curiosity exhibited in his studies.  Bill James had a recent article, "More Walks, More Strike Outs," in which he expanded on his view that pitchers' walks are more important than strike outs. This led me to think about the way we compare pitchers using K/BB ratios. Because James' article is available on a subscription web site (, I can't link to the article.  But I will describe some of his findings, and consider how they might affect our evaluation of Astros' pitchers.

Here are the high points:


  • James challenges the notion that a strike out is equal to a walk, which is the presumption behind statistics like subtracting pitcher walks from strike outs,  It also seems to me that this is implicit in using a K/BB ratio. 
  • In James' view, avoiding walks is more valuable than striking out batters.  His study finds that low strike out-low walk pitchers in general have lower ERAs than high strike out-high walk pitchers.  This relationship was more or less consistent across similar pitcher groups.
  • By using a K/BB ratio to compare pitchers, we assume that a pitcher can offset his poor control by striking out batters. This may be true to an extent, but James' study suggests that 1 strike out doesn't necessarily offset 1 walk.
  •  One explanation for the results may be what James terms the "fellow travelers" of walks.   Pitchers who give up more walks also throw more wild pitches, hit more batters, and commit more balks.  Similarly, James' "high strike out group" (compared to the "low walk group") had 28% more strike outs, which was accompanied by 59% more walks, 70% more wild pitches, 19% more balks, and 19% more hit batsmen.


Now that I've set up a backdrop, we can get around to some practical implications for the Astros.

Strike outs are flashy, but does it lead us to undervalue control?

In some game thread discussions, the reduction in Bud Norris' fastball velocity has come up.  Norris continues to have good velocity, even though it's a few ticks lower than previous seasons.  Yet one possible explanation is that Norris has attempted to improve his control by possibly backing off a bit on the velocity.   Norris has reduced his walk rate by 20% this season.  Norris' strike out rate hasn't decreased either.  Norris' K/BB ratio has increased from 2.05, in 2010, to 2.62 in 2011, largely due to the decrease in BB rate.  No doubt this goes a long way in explaining his ERA reduction of more than 1 run.

Perhaps another example of the walk vs strike out trade off is Wandy Rodriguez.   Wandy's strike out rate has declined 11% in 2011.  Yet an 18% decrease in his walk rate resulted in a higher K/BB ratio in 2011.  Despite the lower K rate, Wandy's ERA declined from 3.62 in 2010 to 3.25 in 2011.

For the Astros' 2011 pitching staff, BB rate has a higher correlation with ERA than K rate (.26 R squared vs. .19 R squared). Haven't had enough numbers yet?  In the table below, I calculate a strike out minus walk statistic for each Astros pitcher.  I then calculate a weighted strike out minus walk rate statistic, based on a 1.5 weighting for walks.  (This is similar to the relationship between walks and strike outs in the Fielding Independent Pitching formula.)  The difference between strike out rate and walk rate will decline because the walk rate has been given a higher value.  The percentage decrease column is supposed to tell you which Astros' pitchers' statistics are hurt the worst if avoiding walks is assumed to have greater value than strike outs.  I'm not sure if this table is useful, but since I spent time preparing it, here it is.  Lyles, Norris, and Wandy fare the best.  Happ and most of the bullpen---look even worse.


Weighted Percent
K/9 BB/9 K - BB K-BB Decrease
Happ 7.6 4.4 3.2 1.0 68%
A. Rodriguez 6.4 2.9 3.5 2.1 41%
Myers 6.2 2.6 3.7 2.4 35%
Lyles 7.0 2.9 4.2 2.7 34%
Norris 9.4 3.6 5.8 4.0 31%
W. Rodriguez 7.3 2.6 4.7 3.4 27%
Del Rosario 5.02 4.54 0.5 -1.8 473%
Escalona 7.63 4.7 2.9 0.6 80%
Abad 6.86 4.12 2.7 0.7 75%
Melancon 7.62 3.69 3.9 2.1 47%
Lopez 7.34 3.08 4.3 2.7 36%