Batted balls, defense, and the Marlins pitching staff

Rally’s historical WAR database has an interesting method for determining pitcher WAR, involving subtracting prorated defensive runs based on the number of balls in play allowed by a given pitcher. As far as I know, Rally does not attempt to break up team defense into more granular responsibilities, even though it is clear that ground ball pitchers would be more affected by their infield defense than their outfield defense, for example.

On the heels of an excellent article by BtB colleague and friend of the Maniac Harry Pavlidis on the move up the minors of a ground ball pitcher, I figured I’d take a look at some numbers and see how our defense might be affecting the run allowance of the pitchers in front of them. Since my partner in crime John Herodl is working on a series of articles of future Marlins ground ball pitchers, it may be useful for us to know how our current defensive iteration is going to help or hurt them.

Methodology

Using the NL linear weights for batted balls provided by Pavlidis in the linked article, I took 2007-2009 data on batted balls provided by Baseball-Reference (via Retrosheet I believe) and stripped the home runs out of the fly balls. The value of the fly ball including the homer was 0.06 runs per BIP, but since the home run is not affected by defense, I had to remove the value of the home run. Leaving the HR linear weights at 1.40 (a close enough guess), I get a value for the fly ball without home runs as -0.108 runs, very similar to the average ground ball worth-0.08 runs.

To estimate the approximate defensive value added or taken away by the team’s defensive play, I added up projected defensive values for infielders and outfielders and divided by an estimate on the amount of plays each unit is responsible for. All grounders are the responsibility of the infield, while all fly balls not categorized as popups and around 90% of all line drives were attributed to outfielders. The line drive guess was just that, a guess. I’m not sure if that is correct in the least bit, but I’ll go with it.

First off, let’s start with the defensive contributions. Adding up the projections and fudging with the numbers a little bit, I got a defensive projection of around -24 runs per 1843 grounders for the infield, a rate of -0.013 runs per grounder. In other words, the infield defense is around 0.013 runs worse than average on a ground ball. The outfield, on the other hand, was estimated at around 10 runs per 1930 responsible plays, a rate of 0.0052 runs per play.

Here’s a table summarizing the various weights used:

Keep in mind that I switched the sign conventions on the defensive weights in order to keep them in scale with the other ones. Remember that since we are discussing run prevention, negative numbers are good, while positive numbers are bad.

A couple of pitchers for examples

I decided to look at a few starters to see how their batted ball profiles would be affected by the team’s defense. First, let’s start with Chris Volstad, a player with a history of ground balls and home runs. CHONE projects 165 innings, but to estimate balls in play, we need an estimate on batters faced. Volstad has faced around 4.3 batters per inning over his career, which would give us 709 batters faced in that time. At that rate and his projected strikeouts and walks, you are looking at 499 balls in play and 22 home runs. From 2007 to 2009, according to the data on fly balls vs. line drives, about 12% of home runs have occurred on line drives. Since I only want to look at the effect on fly balls, we’ll call that 20 homers on fly balls.

In total, those balls in play lead to around 0 runs above or below average allowed on just those balls. Of course, there are still a number of line drives and popups to consider, which would likely push that value further up. It can be pointed out, however, that if Volstad could improve his ground ball rate further, at the expense of his fly balls, the home run rate would drop and would be a bigger benefit than the avoidance of the team’s poor infield defense.

But that’s a pitcher with a HR/FB rate that was very high. What about one that was pretty low? Let’s take a look at Josh Johnson. For Johnson, CHONE has a projected 130 innings, but I’m going to bump it up to 180 for a conservative but more realistic guess. At 180 innings and an approximate 4.15 batters faced per inning, you get an estimate of 747 batters faced. In those same 180 innings, CHONE would have projected 162 strikeouts and 162 K’s, 55 BB’s, and 4 HBP, giving him 526 balls in play and 14 home runs, 13 from fly balls.

In Johnson’s case, the numbers were quite similar to Volstad’s, and the major difference was home runs allowed. This total was some 11 runs better than average. Let’s try a final example, involving Ricky Nolasco. The difference with Nolasco is that he is generally a fly ball pitcher, whereas the first two were worm burners. What’s the effect? I’ll bring CHONE’s projections up 180 innings again, giving Nolasco around 518 balls in play and 22 homers from fly balls.

Though JJ and Ricky offer similar number of balls in play, they have very different GB%/FB%. And the differences in home runs allowed once again dominate the run allowance of these BIP, such that the differences in defensive play are minute compared to the allowance of home runs. For our team, it appears that there is no reason to attempt to play away from the infield, because ultimately ground balls are just going to be much more valuable unless your defense is miles better or worse than the average. Even a difference of 20 runs around the average apparently is not enough to make a huge difference for individual pitchers. Team defense, however, is another story.