This is a post that necessarily begins with an embarrassing story. Friday afternoon, I was doing some research into strike zones called against individual hitters, and full teams. I was comparing actual strikes against expected strikes, just as I’ve done with pitchers in the past. I didn’t know if I would find anything, but that was the whole purpose — it was statistical exploration. After some time, I gathered all my results in a spreadsheet, and then I remembered I already did this in October. Like the exact same thing, for this very website.
Memory jogged, I read the old post, and I read the old comments underneath the old post. Everything looked fine; I didn’t know what I was checking. All the results agreed, with the only difference being that this time I used a denominator of 1,000 called pitches instead of 1,000 general pitches. Hardly worth an updated post. But part of one comment did stand out to me, both because it was interesting, and because it meant I might not have to throw away all my data:
Dave S says:
October 19, 2012 at 10:39 pm
Do catchers get more calls than anyone?
It’s not immediately apparent why catchers might get a slightly different strike zone when they’re hitting than everyone else, but I figured there was enough there to pursue it, because it wouldn’t take a lot more work. If there was nothing, welp, that was only a few extra minutes. If there was something, hey look, there’s something! When catchers are in the field, there’s a catcher-umpire interaction, and there’s catcher familiarity with the strike zone. It makes some sense that there might be some sort of effect, then, that shows up when catchers are at the plate.
A quick refresher: it isn’t too challenging to gather pitch data, and strike data. I grabbed 2012 data from Baseball-Reference, since FanGraphs only has it for pitchers. I already had data for all players who batted at least 250 times. FanGraphs also has plate-discipline data, and from this data we can figure out how many pitches were thrown to each hitter in the strike zone, and how many swings each hitter took at pitches outside of the strike zone. From here, we can calculate expected strikes, and from there we can calculate the difference between expected strikes and actual strikes. Then we put it over a denominator of 1,000 called pitches, for purposes of standardization.
I separated catchers from non-catchers in the spreadsheet. For 2012, I had 33 catchers, totaling nearly 54,000 pitches, and 269 non-catchers, totaling more than 500,000 pitches. For all of the non-catchers, the average was a difference of -9 per 1,000 called pitches. That is, for non-catchers, per 1,000 called pitches, there were nine fewer actual strikes than expected strikes. The average isn’t zero because umpires are imperfect and the PITCHf/x strike zone is imperfect.
For all of the catchers, the average was a difference of -19 per 1,000 called pitches. That is, for catchers, per 1,000 called pitches, there were 19 fewer actual strikes than expected strikes. This separates them from the non-catchers by ten strikes per 1,000 called pitches. Looked at in another way, of the 33 catchers in the sample, 26 came out with a difference below -9. So, by this measure, in 2012, 26 were given more favorable strike zones than the average non-catcher.
When I looked at the initial list, I laughed when I saw Miguel Olivo getting one of the most favorable strike zones. I figured, he’s often behind in the count, when the zone shrinks, and also, if you’re an umpire, if Miguel Olivo doesn’t swing at a pitch, on some level you have to assume it was a ball. But this isn’t unique to Olivo. He’s just one guy in a much bigger sample.
The difference isn’t enormous — we’re talking one strike per 100 called pitches. It might not even be statistically significant. But there is at least some evidence now to suggest that catchers, overall, get more favorable strike zones when they’re hitting than non-catchers, overall. The reasons for this could be simple or complicated, and I won’t get into what they could be in here.
That’s because a whole lot more work could stand to be done on this before we say anything conclusive. We could, for one thing, get rid of the 250-PA minimum, since I just used that for convenience. We could look at more seasons than just 2012. We could perform a more direct and thorough PITCHf/x analysis, and we could try to control for count and batter size. What we don’t know, right now, is that umpires are kinder to catchers than non-catchers.
But they might be, as a general rule. Which would be interesting. There’s more yet to be done, but I’m satisfied with this starting point.