Size as a Statistical Field: Part II
by David Luciani
published January 23, 2005

Last time out in this space, I talked a bit about how we should use the size of a player as an actual statistical field even though many in the so-called scouting community might expect us not to do so.  I gave some results of an analysis that showed what hitters of different heights and weights did, on average, with their offensive performances.

This time, we continue the topic by looking at the pitchers.  Again, we're taking all pitcher performances since the 1969 season and splitting out the results based on the heights and weights of each player.  We've scaled the results to "per 1000 batters faced" so we can compare different heights and weights easily.  We're going to use the same weight and height categories we used for the hitters with a couple of exceptions.  The first change we'll make to the categories is that there are very few pitchers in the weight group <=149 and so we've set the cutoff for the lowest category in weight to be <=159 instead.

Here then is a summary of how each weight group performed per 1000 batters faced.  The ERA and WHIP columns are more precise than the rounded numbers you see displayed here and they use the actual non-rounded totals.  Unlike the batters, many of the categories have no meaning when you look at them in an aggregate (such as saves, wins or shutouts per batter faced, for example) so we are including here only the key categories that are components of the overall performance:

Unlike the hitter charts, the patterns here are not so obvious.  In fact, there are only a few obvious trends and even they are difficult to explain.  The ERA curve is relatively flat, dipping a bit in the 180-220 range or so but even at that, not so clear a pattern as any of the categories we saw on the weight curve for hitters.  We do see the quick decline in performance for a pitcher once he hits about 240 lbs or more and it seems that this is because his control declines along with a corresponding jump in strikeouts and home runs allowed.  One theory here, and this is more of a random thought than something we have data to prove, is that perhaps such a heavy pitcher throws harder.  That would cause a ball to be more unpredictable (thus more walks), leading to more strikeouts and more home runs (i.e. because the elasticity of the fastball).  I don't pretend to have the answers but it's one possibility here.  What was interesting is how consistent the hits allowed numbers are regardless of weight and the strikeout and walk totals really don't deviate much with the exception of the extreme weights.  It's not that the results come as much of a surprise but after seeing the hitting weight chart, we could be understandably disappointed if we had been expecting such similar and obvious patterns to emerge on the pitcher chart.  One result we can take away is that we could be justifiably cautious of a pitcher who comes into spring training one year with a weight increase that takes him above 240 or 250 lbs when he never was that heavy before.  Of course, we must also consider the possibility that pitchers in this weight class, during the period considered, simply weren't as good as the lighter pitchers and that weight had nothing to do with it.

We would expect the height chart to alleviate that disappointment.  After all, taller pitchers have more leverage and we would therefore expect more strikeouts. Again, I had to raise the bottom end of height here for consideration as there have been so few pitchers shorter than 5'6" or 5'7" and so the bottom level has been raised to 5'8" or lower.  Here's the summary of each group per 1000 batters faced:

Like the weight chart, the height chart for pitchers also disappoints us in our efforts to get an obvious pattern and in fact, those short pitchers did well.  Where hitters' weights and heights create obvious performance results, the pitching charts yield only a few clues.  Here we see only a few patterns such as control apparently improving a bit once a pitcher is 5'10" or 5'11" but then beyond that, the results are relatively stable.  If anything, hits allowed is surprisingly higher for the taller pitchers and the strikeout totals, while increasing some in the tallest pitchers, isn't as steady a line up as we might have expected.  The ERA and WHIP totals offer no exciting trends and if anything, it would seem both columns favor the shorter pitchers.

While I'm not going to draw conclusions on such an oversimplified study, I will summarize a few possibilities here.  First, on the hitting side, I strongly believe that height and weight are major factors in performance and the charts prove this better than anything.  On the pitching side, nothing replaces the pitcher's actual skill with his pitches and more importantly, his ability to deceive.  Robert Adair in his enjoyable but justifiably math-heavy book The Physics of Baseball gave us a possible explanation of why taller pitchers don't get an automatic increase in their velocity when he said that "tall pitchers have a longer pitching trajectory than shorter (short-arm) pitchers and then apply a little less force on the ball - albeit over a longer distance - to gain the same muzzle velocity."

When one considers the outstanding performances of the past thirty years or so, it's interesting that so many so-called "small" pitchers (by baseball standards) have been successful such as Greg Maddux, Pedro Martinez, Fernando Valenzuela, Ron Guidry, Mike Flanagan, Catfish Hunter and so on.

I plan to continue researching this topic in much greater detail but my preliminary hypothesis here is that on the pitching side, it matters more what pitches a pitcher throws and how he uses those pitches for deception than what his physical size is.  I don't think this is a new idea but the data seems to back it up and I am going to explore the actual correlations further and in greater detail.  In particular, a proper study here will need to find a way to isolate out relief and starting performances.  A reliever, on average, will have better numbers because he doesn't have to stretch himself out over a game and thus can exert tremendous energy over an inning or two without concern over how he'll feel a half hour later.

One related item does come to mind, though, and that is that some organizations in baseball shy away from small, amateur pitchers who get good results.  Understandably they may not "look like" potential big league players but I suspect that there are pitchers out there who have the stuff and skill to use it who could be getting overlooked by drafting teams.

As I say, these aren't conclusions as much as theories and I share with you the results for you to make your own decisions about the importance of height and weight.  There is no escaping that particularly for hitters, it's a key statistical field that we must consider when making a forecast.