|Evaluating Pitchers: Groundballs and Homeruns|
Written by Brian James Oak (Contact & Archive) on August 14, 2007
If you buy into the DIPS theory that I went over in my last article and believe that a pitcher can’t control whether a ball in play becomes a hit, it would seem intuitive to believe that he can’t control whether a ball in play becomes a homerun either. It would seem that the power of the batter controls the destiny of the ball. While the batter definitely has some control over this—it is what separates Ryan Howard from Ryan Freel after all—the pitcher has an equal ability to prevent or allow homeruns. Comparing James Shields and Tim Hudson’s homerun rate might not be as sexy as the two Ryan’s but they too have a say in whether a pitch becomes a homerun and it plays an important part in the overall ability of those pitchers.
Before we get into how much a homerun affects a pitcher, let’s discuss why a pitcher is able to control them. For one thing the strikeout rate of a pitcher has an impact on the homerun rate. Why? Because every time a batter strikes out he is not putting the ball in play and those batters had no chance of hitting a homerun. (note: If you are reading closely, you may worry that I double counted the homeruns when I was figuring out the effects of strikeouts on a pitcher’s line in my last article, but this is not so because I normalized the homerun rates of the pitchers in the study before completing it.) This is one reason why great pitchers like Jake Peavy often have a low home run rate. Speaking of Peavy, though it is out of a pitcher’s control, a big ballpark like Petco can also help to bring down a homerun rate.
The other thing a pitcher can control is groundballs. Pitchers who keep the ball low in the strikezone and/or throw a sinker get far more groundballs than others. While groundballs themselves don’t necessitate a better ERA than flyballs (though the quality of infield vs. outfield defense on a team could make it so), they do make a pitcher far less likely to allow homeruns, which can heavily influence ERA. In fact, Nate Silver found that groundballs are actually a better predictor for a pitcher’s homerun rate than the pitcher’s own previous homerun rate.
Last time, I found that each K/9 affects a pitcher’s ERA by 0.14 and each BB/9 by 0.323. So how much does each HR/9 affect a pitcher? Well, first we need to find out how many runs each homerun hit in the majors is worth. In the early 1960’s, George Lindsey found a homerun to be worth about 1.42 runs, while Palmer and Thorn later confirmed that number, at least in the era from 1941 to 1977. While we don’t have a more recent number, we can assume that it is higher today due to the increased emphasis on on-base percentage. That being the case, we’ll round it up and assume that a homerun is worth 1.5 runs. While a typical homerun actually plates more runners than that, we have to account for the fact that some of those runners would have scored had the homerun never been hit.
An average pitcher allows 1.12 homeruns per nine innings. So a pitcher who allowed 2.12 homeruns per nine innings would typically have an ERA that was 1.5 runs higher (an average pitcher’s ERA would go from 4.50 to 6.00). That sounds like a lot, but most pitchers this side of Eric Milton tend to stay between 0.5 and 1.5.
To continue to look at Johan Santana who should have had a 3.64 ERA by just applying his 2006 walk and strikeout rate to an average pitcher, applying his homerun rate of 0.92 would give him the more appropriate ERA of 3.34.
While 3.34 is still higher than his actual mark of 2.77, we are not yet accounting for the quality of defense behind him, which contrary to popular belief does affect earned run average. Next time I will explore exactly how this works and why some of the more popular pitchers in baseball appear much better or worse than they really are.
You can contact the author at