Numbers don't lie: Umpires are certainly biased ... but racist?

One of the smartest writers out there is David Leonhardt, whose new Times-related microsite is opening for business very soon. I’ve been in touch with Leonhardt a few times over the years, so I know he’s a baseball fan.

When I saw this tweet, though, I didn’t figure it was about baseball:

It is. That link goes to an essay by Brayden King and Jerry Kim, in which they use PITCHf/x data to analyze umpires’ strike-zone calls from 2008 and ‘09, which included more than 700,000 pitches. The over-arching conclusion? Roughly 14 percent of non-swinging (that is, taken) pitches were called incorrectly.

Before we go any farther, I will suggest that figure is probably inflated. Between calibration issues and the inability of any system to perfectly judge a hitter’s natural strike zone, I think it’s appropriate to cut the umpires a bit of slack. Or a lot of slack. So let’s say it’s only 10 percent they’re missing.

Yes, that still seems like a lot. But consider that major-league pitchers are really, really good, and they’re trained to aim for the edges of the strike zone; anything else is playing with gasoline-drenched matches. Also consider that major-league hitters are, I’m willing to bet, taking more close pitches than ever. All of which means lots of marginal pitches, traveling at high speeds or with hellacious movement. Even the best umpires on the planet -- that is, major-league umpires -- are going to miss some of the close ones.

What’s interesting is that the incorrect calls aren’t evenly distributed. Among the findings:

• Umpires favor the home team (but just ever so slightly);

• Umpires don’t like calling a strike on an 0-and-2 count; but ...

• ... they really don’t like calling a ball on a 3-and-0 count; and ...

• ... oddly, they seem to miss a lot more pitches in the ninth inning of a close game than in the first inning. That is, they choke.

Here’s the really tough one, though, and the reason for Leonhardt’s tweet:

The race of the pitcher, we found, also mattered, but not as much as other factors. Umpires were 10 percent less likely to expand the strike zone for African-American pitchers than for Caucasian pitchers, but race did not seem to influence whether an umpire called a pitch a ball when it was actually a strike.

In a way, it is surprising to find such nonrandom errors, because MLB umpires are keenly aware that their ball-strike calls are being scrutinized and that they will be evaluated accordingly. They have an incentive to be as accurate as possible. And yet they still make routine errors behind the plate. We think that the sorts of errors we observed are not deliberate and may reflect an unconscious and biased decision-making process.

Before we go any farther, I do have to mention that the results might be outdated. Why doesn’t this study include, or consist solely of, data from more recent seasons? While umpires are aware they’re scrutinized, they’ve been significantly more scrutinized in the most recent seasons, so it’s possible they’ve changed their behavior, whether unconsciously or otherwise.

That said, I do think it’s worth making a distinction between racism and bias ... especially when the bias might well be unconscious. It’s one thing to accuse umpires of unconscious bias, because nearly all of us are guilty of unconscious biases.

If you don’t believe me, just follow Slate on Twitter; every few days, they present the results of yet another study showing that blacks in this country are treated worse than whites, and quite often by people who have no idea they’re doing it and would probably be mortified to be presented with the evidence.

If you accuse the umpires of racism, though, is it even a small leap to accuse them of actually being racists? I don’t think it is, and I’m not willing to do that. I think most umpires are generally fair-minded, at least when they’re working. Just as I believe that I’m fair-minded, and you’re fair-minded, most of the time. Consciously, anyway.

But yes, it’s perfectly fair to accuse the umpires of (mostly) unconscious bias, against young players, against visiting teams, against ending plate appearances ... and, most unfortunately, against pitchers with dark skin. While most of these biases will even out over the course of a season, being black doesn’t even out.

Historically, do we need to re-evaluate the worthiness of black pitchers? Was Bob Gibson even better than we think? Did bias cost Luis Tiant enough wins to materially impact his Hall of Fame résumé? Has CC Sabathia already done enough?

I don’t have any idea. But these do seem like fair questions.

Oddly, while the supposition that these are unconscious biases makes the umpires look better ... it’s also an argument for taking these decisions out of their hands.

If they could somehow be trained to call the pitchers more fairly, that probably would have happened already. All of these unconscious biases seem the result of human nature, which is strong enough to resist even the most concerted training and education.

So what’s to be done? If I were a black pitcher -- or even a white pitcher, I’d like to think -- I might well talk to my union about exploring the automation of the strike zone. Of course, the problem is that all the automation in the world won’t come soon enough to help LaTroy Hawkins, and probably not Sabathia, either.

Oh, except automation might not help Sabathia at all. Because as a veteran pitcher with a tremendous career behind him, he also benefits from unconscious biases.

This stuff’s tricky and unattractive, and we should resist the urge to demonize people for being people. But that doesn’t mean we should pretend it’s not there.

You can follow Rob Neyer on Twitter, where most of his offensive tweets are unconscious.