Ground Balls and Line Drives

This is a response to the article “Batted Balls: A Quick Study” by Nathaniel Stoltz at Call to the Pen, which argued that a big reason ground ball pitchers are desirable is that they allow fewer line drives and line drives are most likely to fall for hits. If true this would also suggest that ground ball + line drive specialists such as Austin Jackson might be a mythical breed.

It is true that there is a simple (negative) correlation between GB/FB rate and line drive percentage – just as there is a negative correlation between GB% and LD%. This is bound to be the case because line drives are, by definition, balls hit in the air. There is also a negative correlation between pop up percentage and line drive percentage, and between outfield fly percentage and line drive percentage. The correlations are mechanical, they don’t mean much.

If you look at how fly balls are distributed between outfield flies, pop ups and line drives the story looks very different. If we regress GB% (which is simply a transformation of GB/FB rate) we find that groundball pitchers – though they allow fewer balls to be hit in the air by definition – have a much larger percentage of line drives among fly balls. The coefficient is approximately 0.5, and meets statistical tests for significance with flying colors. The average percentage of fly balls classified as line drives is approximately 35% and the average percentage of all balls in play that stay on the ground is something like 45%. The numbers suggest that a hypothetical pitcher who allowed zero ground balls would only have a 13% line drive rate. An ‘average’ pitcher would have about 45% ground balls, leaving 55% in the air – 35% of which would be line drives. A pitcher who got 55% ground balls would have 45% flies, but 40% of those would be line drives.

For an average pitcher, then, a 10 percentage point rise in GB% would cause their line drive rate to go down (minimally) from 19.25% to 18%. But an extreme fly ball pitcher will have a lower line drive rate than a guy who allows slightly more fly balls than average.

One curious thing is that it seems to work exactly the same way for hitters as it does for pitchers: a coefficient of 0.49. So… the hitters with the highest line drive rates, all else equal, ought to be guys who hit just a few more flies than average. I’m not going to try to say that this issue has been resolved once and for all – but I would say that there is ample reason to rethink the doctrine of encouraging fly ball pitchers to keep the ball down. Porcello has a GB rate of 50% – for him, more ground balls means fewer line drives (and probably fewer hits). That won’t work for Phil Coke, with with a GB rate of 35% or Joel Zumaya with a GB rate of 37%. For those guys incrementally improving their GB/FB will mean more line drives and more hits. Not good. We want them to elevate it. Guys like Justin Verlander are at exactly the wrong place on the scale – a GB rate of 40%. Verlander needs to pick a side.