Pitching Repertoire: Visualizing Success

The Boston Red Sox enter 2015 with two starting pitchers who have never thrown an inning for Boston, two who have logged fewer than 200 innings for them, and Clay Buchholz, who is the poster child for inconsistency. This series will also look at Rick Porcello, Justin Masterson, Wade Miley, and Joe Kelly. To get a little better idea of the pitching repertoire we might expect from these players, Ian York has studied the data and created visualizations of their 2014 pitches, based on PITCHf/x.

Pitches cover a lot of territory and these images require a little context to be understood. Using Buchholz’ four-seam fastballs (abbreviated “FF” in the images) as an example, we can see the distribution of all these pitches, both to left- and right-handed batters, superimposed on the official strike zone:

We are looking at the plots from the umpire’s viewpoint, so the batters are standing “in between” the two plots. Like most pitchers, Buchholz targeted the outer part of the strike zone with his fastball.

Of course, the rulebook strike zone isn’t the one that umpires actually call. We can overlay this zone with the ones that umpires called in 2014:

Note that Buchholz’s fastball to left-handed batters was not on the outside of the de facto strike zone, as it seems to be when using the rulebook strike zone. Because the strike zone as called for LHB is shifted to the left, most of his fastballs actually ended up not far from the center of the de facto zone.

Now, let’s cut the rulebook zone up into smaller boxes of a more useful size. We’ll look at pitches that fall into 25 specific subregions, in and just outside the rulebook strike zone:

The main question we have is whether, and how hard, particular pitches were hit. That means we will only examine pitches that were strikes, hits, and outs (excluding balls, pitchouts, intentional balls, and hit batsmen). The distribution of these selected pitches could be quite different from that of all the pitches thrown. We’ll represent this distribution as circles, with their area proportional to the number of pitches in each of the boxes:

Inside the strike zone, this does not show much more than the distribution plot. However, we can see that Buchholz induced a moderate number of swings on pitches that were not inside the strike zone, especially outside and above the zone to RHB, and inside to LHB.

Finally, the circles are color-coded to indicate batter success. This is based on total bases per pitch, again in each sub-zone. Grey represents the baseball-wide average total bases per pitch for pitches that were not balls (in 2014, about 0.321 as calculated from PITCH/fx data); blue is less than that, red is higher – that is, bad for the pitcher:

From a pitching viewpoint, we would like to see all blue ‒ and the deeper the blue, the better. If some circles are red, we would like them to be small circles – that is, if the pitcher has a weak spot, but rarely throws the ball there, potential harm is minimized.

Buchholz’s fastball to right-handed batters, for example, is at least moderately blue in the region where most of his pitches end up ‒ outside/middle ‒ while the reddest of the bubbles is the small one low and inside. But to left-handed batters his most common location ‒ again, outside/middle ‒ gets hit substantially harder, resulting in more total bases per pitch than average and ending up in the red side of the spectrum.

Above, we examined just his fastball, but this is reflected in his overall splits: In 2014 RHB hit .259/.307/.389/.696 against him, while LHB hit .284/.356/.437/.793 – almost 100 points of OPS higher.

By dissecting the data on Buchholz’s fastball we can see – and confirm – what the statistics tell us is happening. By studying the locations and outcomes, there is a framework to understand what a pitcher needs to do to be successful.  We will next look at in succession each of the Red Sox starting pitchers for 2015: Buchholz, Rick Porcello, Wade Miley, Joe Kelly, and Justin Masterson.

Follow Ian on Twitter at @iayork