- FanGraphs Baseball - http://www.fangraphs.com/blogs -

Is Bill James Right about Ground Ball Pitchers and Injuries?

When Bill James speaks, many in the baseball committee listen intently–as they should. James, while certainly not always correct in his theories (and, really, who is), can always be counted on to provide the larger community with excellent food for thought.

In this most recent case, James claimed that ground ball pitchers have essentially been overrated. Per Rob Neyer:

Any analyst can give you a long list of reasons why ground ball pitchers should be the best pitchers. The problem is, they’re not.

Make a list of the best pitchers in baseball. Make a list of the best pitchers in baseball, in any era, and what you will find is that 80% of them are not ground ball pitchers. They’re fly ball pitchers.

What I have never understood about ground ball pitchers, and do not understand now, is why they always get hurt. Show me an extreme ground ball pitcher, a guy with a terrific ground ball rate, and I’ll show you a guy who is going to be good for two years and then get hurt. I’m not saying this about Chien-Ming Wang and Brandon Webb; I was saying this before Chien-Ming Wang and Brandon Webb. They’re just the latest examples. Mark Fidrych. Randy Jones. Ross Grimsley. Mike Caldwell. Rick Langford. Lary Sorensen. Clyde Wright. Fritz Peterson. Dave Roberts. They’re great for two years, and then they blow up. Always.

Now, there is a lot that can be teased out here, but I want to focus on the last part of James’ claim–that ground ball pitchers are more injury prone. Are ground ball pitchers (specifically, extreme ground ball pitchers) more injury prone?

We can look at this a few ways. First, let’s take the top 25% of ground ball hurlers since 2002 and compare their rate of injury against the other 75%. The 75th percentile for GB% over this time period is 48% for all pitchers with a minimum of 40 IPs (with no discernible differences by starter/reliever, but I did restrict to pitchers in the same role the year prior). Thanks to Jeff Zimmerman, we have injury data going back to 2002. Looking at starting pitchers with ground ball rates over 48% in a given season from 2002-2012, we see that actually their odds of injury appear lower than those with less than a 48% ground ball rate (injury rates of 14% vs. 20%, respectively):

Type of Pitcher Odds of Arm Injury
>= 48% GB% .16
< 48% GB% .25

What about even more extreme ground ball pitchers? Say, those in the top 10%? Turns out the 90th percentile for GB% over this time period was 52%. Even when restricting the data based on this number we don’t see any strong evidence for ground ball pitchers being more injury prone. In fact, the results are almost identical (injury rates of 15% vs. 19%, respectively):

Type of Pitcher Odds of Arm Injury
>= 52% GB% .18
< 52% GB% .23

In the interests of being comprehensive, maybe the issue isn’t just on of throwing ground balls but really of pitch type. Sinker ball pitchers tend to induce more ground balls, so let’s see if anything changes when we replicate the analysis based on PITCHf/x data.

Here, I looked at all pitchers with the same role two years in a row with at least 40 IP. The first set of numbers includes all pitchers, regardless of role, using a sinker percentage of 44% as the cut-off for extreme sinker ball pitchers:

Sinker % Odds of Arm Injury
>= 44% 0.16
< 44% 0.18

Again, no confirmation of James’ hypothesis. What if we just restrict to starting pitchers?

Sinker % Odds of Arm Injury
>= 44% 0.12
< 44% 0.22

Still nothing. And, again, it appears the effect works in the opposite direction. Starters that throw a higher percentage of sinkers have less risk of arm injury than those that don’t.

UPDATE: Some have asked whether it isn’t about the frequency of injury, but the severity. Very fair question. I looked at the frequency of large numbers of days missed due to arm injuries and compared them to ground ball pitchers in the 75th percentile, 90th percentile, and those below the 75th percentile. Whether we look at the percent of arm injuries that resulted in over 100 days lost or over 150 days lost there really doesn’t seem to be a “penalty” for extreme ground ball pitchers:

>= 75th P-tile >= 90th P-tile < 75th P-tile
Arm Injuries: % >99 days missed 28% 29% 31%
Arm Injuries: % >150 days missed 17% 14% 17%

Furthermore, if we just compare the rate of extreme arm injuries between pitchers with ground ball rates above and below the 75th percentile we see that extreme ground ball pitchers only account for 14% of such injuries, whereas those below the 75th percentile account for 86%.

Bottom line, this initial look does not suggest that extreme ground ball pitchers are more prone to catastrophic arm injuries compared to their counterparts.

Now, this wasn’t a comprehensive study–it was a quick look at the data on a Saturday afternoon. Others should certainly jump in and both replicate this quick look and delve into other aspects of James’ hypothesis. But, for now, the initial evidence doesn’t appear to support the injury risk claim.