September 8, 2013: Canton, OH – You know that statistic that shows up in the “pitching” section of box scores listing the number of groundouts and fly outs?  Ever wonder if these numbers have any impact on how the game turned out?  Well, we know that pitchers do have some influence in what direction the ball will go when contact is made by the hitter in regards to the batted ball turning into a ground ball or a fly ball (for the sake of simplicity, we are lumping pop-ups and line drives into this category).  We know this because pitchers who tend to get a ton of ground balls as measured by their GO/FO% also tend to maintain similar numbers through their career.  Beyond the assumed increase in double plays that result from the simple fact that the ball is hit on the ground more often, offering the defense more chances to turn double plays, what else is impacted by this “pitcher skill”?

From talking to a few scouts around the league, I found that the GO/FO% statistic is measured and used in evaluating the prospects of these pitchers.  A few scouts said that they consider the statistic as helpful in measuring the pitchers susceptibility to giving up home runs, while other scouts claim that it is helpful for managers to know if a pitcher is a particularly extreme ground ball or fly ball guy in setting the defense behind that pitcher.  What I am wondering is if ground ball percentage relates to success for pitchers within the PEBA and, if so, how strong is this correlation?

I set out to answer this question by analyzing this year’s pitching performances up through yesterday.  The first thing we need to do is set the parameters on who is included and who is excluded from the data set.  I set the qualifier mark at 140 innings pitched.  This resulted in 99 pitchers qualifying across the league for the study (N = 99).  Next, I wanted to solve the issue of how to define “success”.  So I decided to include correlations for GO/FO with ERA, FIP (fielding independent pitching) and pitcher VORP (value over replacement player).  I figure we can see if any correlation exists between the data, how strong that correlation is, and, accepting “correlation does not equal causation”, perhaps we can at least see if there is a relevance between GO/FO and success.  In addition, I wanted to look at the correlation between GO/FO and home runs allowed per nine innings pitched (HR/9) for pitchers since it was mentioned by more than one scout that they believed a correlation existed.

First, a word on correlation coefficient (represented by the symbol “r”).  A quick and dirty explanation of what we are looking at when we see the numbers below.  The coefficient ranges from “1” to “-1” but, really, we are looking for a “r” near either number if there is any correlation and a “r” near “0”, meaning there is no or a weak correlation.

The correlation coefficient (r) is .09, which can be interpreted as saying there is basically no correlation between ERA and GO/FO, or that success with regards to ERA does not consistently occur with a high (or low) GB%.

I was somewhat surprised by this, but it is plain to see that the data doesn’t support any relevant connection between keeping the ball on the ground and preventing earned runs.  It is interesting to see a string of high GO/FO pitchers (>.55%) running across the top of the graph, demonstrating an almost equal number of pitchers with ERAs below 4 as there are pitchers above.  Kind of hammers the point home.

The r = .04, which once again indicates no significant relationship between GB% and FIP.  There is something of interest in relation to this graph, but let us finish off the other graphs first.

Here, r = -0.09.  So the trend now is that amongst ERA, FIP and VORP, we are not seeing any significant correlation between a pitcher’s ability to keep the ball on the ground and his success as measured by any of these three measures.  What’s even more disturbing in the VORP graph is that the dots in the upper-left both sit in the high GB% rate and the negative VORP section, indicating that this specific group is good at keeping the ball on the ground and terrible at providing value to their team.

Here, r = -0.26.  We have a weak, but at least small, correlation between keeping the ball on the ground and avoiding home runs.  When I mentioned that the FIP graph was interesting above, it was because home runs allowed is a component of FIP, yet the skill at inducing ground balls does not significantly impact the statistic despite have a small impact on HR/9.  This further illustrates that though there is a correlation, it is small and should be considered so when looking at GO/FO when evaluating pitchers.

In summation, it looks as although GO/FO may have some value in evaluating a pitcher’s ability to avoid home runs, but its value is limited by being only a small correlation.  There isn’t any correlation between keeping the ball on the ground and success as measured by ERA, FIP or VORP for pitchers.  It might be worth looking at the data over a few seasons to improve confidence in these findings, but with this season’s data as the data set, the value of GO/FO is pretty limited as an assessment tool.

– Steve Youngblood regularly writes for the Fargo Gazette