Fernando Tatis Jr., Defensive Conundrum

At times, Fernando Tatis Jr. is an absolutely incredible fielder. Try to imagine someone other than Tatis, Javier Báez, or Andrelton Simmons making this play:

The speed, presence of mind, and arm strength needed to turn that ball into an out are simply breathtaking.

If Tatis can do that, then why does every fielding metric dislike his defense? DRS rated him as two runs below average over 731.1 innings in the field last year, and it was the highest on him. UZR saw him 5.8 runs below average, and Statcast’s new infield OAA pegged him as 13 outs below average. What gives?

If you’re a suspicious type, your mind might immediately go to the fallibility of defensive metrics. After all, they’re far less precise than offensive statistics. They don’t always agree with each other, for one thing, and they take forever to stabilize. Whoever invented the saying “lies, damn lies, and statistics” clearly wasn’t up on modern baseball fielding, or they would have fit UZR and DRS in there somewhere.

And sure, to some extent that’s true. UZR and DRS disagree by four runs on Tatis, nearly half a win of value. OAA is in outs, and that’s even more confusing. An out is less than a run in some cases (the difference between a leadoff single and a leadoff groundout comes out to around two thirds of a run) and more than a run in others (with runners on first and second and one out, it’s more like 1.2 runs). Leaving aside whether outs are the right unit to use when looking at defense (for what it’s worth, I think they are), you can see why it’s all confusing.

But here’s the thing: Fernando Tatis Jr. was a bad fielder last year. Yeah, I know, I just said all those things about defensive metrics. And I just praised his arm! And his range! And his presence of mind! Aren’t those all the things that go into defense?

They are. They absolutely are. But there’s something else that goes into defense, something we often forget about in the rush to come up with new and smart ways to classify it. Range is great, and a powerful arm is great, and also you have to field all the balls that get hit to you, not just a few.

Like, don’t do this:

Or this:

And heck, don’t do this either:

Some of it might be related to trouble coming in. Some of it is every shortstop’s kryptonite, roaming into the hole and backhanding balls. But look, let’s say it: Tatis had issues throwing the ball, and that’s a massive part of his defensive problems.

Errors are rightly regarded with disdain these days. They’re a limited statistic — aside from depending on the scorer’s judgment, they don’t take range into account, which is a huge liability. But that doesn’t mean that they’re pointless, because Tatis’ issue on defense is errors.

A very simplistic way to look at it is that every error is an out below not making an error. If the error wasn’t made, the runner would be out, and instead he’s safe. Of course, a run below completing the play is different than a run below average, because the average result of a play isn’t ever 100% an out.

Aside from the range and scoring issues, that’s a central problem of using errors as a metric. You need to know the conversion chance of the play the error is replacing to get a good understanding of how costly it is. Savant sees this play, for example, as only resulting in an out 61% of the time:

That’s not a hugely costly error. It’s almost a coin flip whether a shortstop makes that play, so plenty of them are bound to come up, well, short. But that was Tatis’ only error on a play that was less than 70% likely to turn into an out. A quick note: there were likely other difficult plays on which Tatis committed errors, but OAA is limited to tracked batted balls, and so only recorded 11 of his 18 errors.

Most of Tatis’ errors were throwing errors — 14 of 18, to be exact. And most of the throwing errors came on layups, gimmes. Take a look at this cookie:

Or, heck, this one:

Or this one:

Take into account the fact that these errors came on plays that weren’t particularly difficult (the average conversion rate on balls Tatis committed errors on was 84%, not too far away from the league average conversion rate of 87% on all balls hit to shortstop), and the picture looks pretty grim. He’s not turning remote chances into errors, in other words; he’s taking essentially league average plays and chucking them into the stands.

Among all non-Tatis major league shortstops this year, there were 10,123 balls in zone (per our handy defensive metrics), on which they committed 260 throwing errors. That’s one throwing error per 39 balls hit to them. Tatis had 207 balls hit to his zone, and made 14 throwing errors. That’s one per 15 chances — yikes!

Tatis’ chances weren’t noticeably harder overall than the major league average. In fact, Statcast sees his expected conversion rate as 88% on all batted balls. Of note, balls in zone and Statcast aren’t using the same definitions, and so there’s some disagreement, which is always going to be a limitation of using multiple defensive metrics. But it’s safe to say that Tatis is throwing away too many outs.

If Tatis committed throwing errors at a league average rate, he would have made between five and six last year. That’s eight or nine extra outs, which would make him an unremarkable shortstop instead of a disastrous one. That seems silly, like a childish answer to a difficult question: why does someone who looks so good rate so poorly? But there’s no way around it: Tatis is bad at fielding in a way that errors explain perfectly.

Think of it this way: on difficult plays (60% or less conversion rate), Tatis was an out above average in only 13 plays. He converted 4% more plays than the average infielder would have in those chances. The other shortstops at the bottom of the OAA leaderboard mostly weren’t making the hard plays.

Jorge Polanco, for instance, ranked worst among all shortstops. He was one out below average on difficult plays, converting only 12% against an expected 15%. Didi Gregorius, who was as poor as Tatis in the field, was no better: 14% conversion rate against a 19% expected conversion rate. Kevin Newman, who wasn’t quite as poor overall as Tatis but was the fourth-worst regular shortstop by OAA, converted a dreadful 5% against 19% expected.

If you needed proof that Tatis is a sparkling defender, there it is. On eight balls that got converted into outs 15% or less of the time last year, Tatis converted two. It’s a vanishingly small sample, but it’s also telling; Tatis has the spectacular in him. Báez, a man literally nicknamed the magician, had 23 such plays and recorded no outs. Simmons attempted nine and converted only one.

Fernando Tatis Jr. is one of the most exciting players in baseball. Some of that is because of his offense, of course; he puts the ball in play a ton, hits it hard, and is a dervish on the basepaths. Despite that, however, plenty of the excitement comes from his fielding. Some of the highlight plays Tatis makes are scarcely believable, the kind you have to see twice to believe. Just because he’s exciting doesn’t mean he’s consistent, however, and that’s been his problem so far.

So when you look at Tatis’ numbers and see a bad fielder, it’s not fake. He really is one of the worst defensive shortstops in the league at the moment. But at the same time, he’s capable of greatness. His highs are extremely high. His lows are extremely low. But if he cleans up the errors, whatever part of him it is that leads him to airmail throws under little or no time pressure, he could turn his defense around overnight. Keep an eye on this story next year, because Fernando Tatis Jr. getting better than his 3.6 WAR, injury-shortened 2019 would be an exciting development indeed.