Attempting to Predict Fernando Tatis Jr.’s Outfield Defense

Fernando Tatis Jr. is now an outfielder for the San Diego Padres. Despite his recent stint on the injured list — his third of the season, and his second related to his shoulder — he might end up collecting the NL MVP provided he can stay productive and healthy, all while moving away from the only position he’s played in professional baseball. In his return to action on Sunday, he raised his wRC+ to 172 and got his first playing time in right field. There wasn’t much to be gleaned from the four balls hit his way, however, leaving us to wonder how well he’ll handle the position going forward. Mike Petriello at MLB.com has covered the unprecedented nature of this move, but I want to take a look at what we can actually expect from Tatis defensively. I’m not the first to consider the question. Last week, Michael Ajeto highlighted Tatis’ defensive ability in an article for Baseball Prospectus and delved into the analytical precedent for shortstops who have recently made the conversion to the outfield. Today, I’ll look at what we might predict about Tatis’ outfield defense given some of the other data we have about his speed.

But let’s set the stage. The biggest flaw in Tatis’ game since his debut has been his defense. After a 2019 characterized by 18 errors and poor defensive metrics, he seemed right the ship last season. But his defense has since regressed, and while his -4.3 Def runs at shortstop this year hasn’t made him unplayable at the position, it’s certainly not ideal. Still, this move wasn’t necessitated by his play at short, but rather by Tatis’ recurrent shoulder problems and the Padres’ stated desire to shield him from further injury, though as Ajeto noted in his piece, whether that will actually work is an open question.

That’s how we ended up with Tatis in the outfield. Before getting too far ahead of ourselves, however, it’s important to know how his numbers at short are derived. His penchant for throwing errors is well documented and they do serve to drop his DRS, UZR, and Outs Above Average (OAA) considerably. We can peek at the component parts of UZR to separate the errors from Tatis’ range, but we aren’t even close to a three-year sample and thinking about UZR sample size is a Pandora’s Box that should be kept closed for today, anyhow. For our purposes, we’re not going to pay much attention to throwing errors or even Tatis’ range on grounders as a shortstop. The baseline we care about is how well he can play right field and quick twitch and speed are the parts of the shortstop skillset that seem mostly likely to translate to the outfield grass.

Judging Tatis’ reads in the outfield will require far more batted balls being hit his way but we can still start estimating based on some plays he’s already made at short. Since his debut, Tatis has posted a -9 OAA, which puts him in the bottom 10 among qualified shortstops. But on plays ranging back, he’s posted 2 OAA, which places him in the top five. Small samples to be sure, but we can see that there’s some ability to track fly balls going back. Here’s what Tatis’ range has looked like on plays with an expected success rate below 50% and the ball at least 45 feet away, per Baseball Savant:

Those are all the end points on plays where he’s tracked down a ball 45 or more feet away. For some of these, it’s apparent that he wasn’t starting in the conventional shortstop position. But that wall of points jutting into the outfield is promising, showing he can get back on the ball comfortably.

Play-by-play defensive measures are rather scant, and measures from Statcast like first step, “Burst,” and route running efficiency are only available for outfielders. That’s why I wanted to look at something we have for all players: measures of speed.

The relationship between sprint speed and OAA is somewhat established for outfielders. The correlation between the two is r = 0.43, which isn’t strong but is certainly substantial. Translating speed to the outfield can be tricky for players who we’re used to seeing on the infield dirt, however. I fed sprint speed and the 90 feet splits provided on Baseball Savant into a Random Forest model to predict OAA. A bit overkill, perhaps, but the results were in line with the sprint speed correlation. I got an r-squared of 0.17, meaning those variables could only explain 17% of the variance in Outs Above Average for outfielders. I was hopeful that by including the 90 foot splits, it might highlight those who have above average or better early acceleration, and that that, in addition to sprint speed, would help paper over some level of the first step or “Burst” metrics that would be useful in projecting a player with no outfield data. But there’s more to it that simply isn’t publicly available. We’ll have to take another route. What we can try to answer is how his speed sets his defensive ceiling.

Sometimes for outfielders, being good (or at least average) means converting all of the easy and routine plays; basically, it pays to never miss high probability opportunities. Borrowing again from the Statcast pages, we can look to see how often outfielders made outs on 5-Star opportunities (0-25% catch probability) and 4-Star opportunities (25-50%), and note its relationship with sprint speed:

Tatis’ sprint speed (29.2 ft/s) is marked by the dotted line. There’s certainly some relationship here but it’s similar to the correlation with Outs Above Average, r = 0.45. Once again, we see that there is a skill gap when it comes to playing the outfield, one that a player like Ender Inciarte (26.8 ft/s, the slow outlier) has shown. Whether Tatis is good will depend on his instincts in the outfield and how well he can read jumps, but his speed will likely make that a smaller skill gap to close. If we go back to the model and try to predict the 4- and 5-Star conversion rate using sprint speed and the 90 foot split times, we see a better r-squared of 0.24. That’s still not ideal when you’re trying to make accurate predictions and small sample caveats apply (the sample only involves a few hundred players), but what was interesting to me was which variables the model placed more weight on when making predictions:

Focusing only on the chart on the left, the model saw the first five feet split to be about as important in making predictions as sprint speed. That first five feet split is measured as the amount of time it takes the batter to get five feet out of the batter’s box, whereas sprint speed is measured as the speed a player ran during their fastest one second interval. There are reasons why you could imagine the five feet split being both useful and problematic. I included batter handedness to try and control for the fact that lefties would post better times. You can also see how players with big swings would post slower times. There’s a rabbit hole of qualms and caveats we could go down, but the model is almost seeing the five feet split (in conjunction with the other splits) as a proxy for acceleration. Taking everything into account (and with the caveat that the model obviously has its flaws), Tatis is projected to convert on 29% of his 4- and 5-Star plays given his sprint speed and split times. I plugged him in as a right fielder out of respect to Trent Grisham and in deference to how the Padres used him on Sunday, but the model doesn’t care that much about the outfield position. The speed plays, and far more often than not, that speed is what is needed to make low probability catches.

There isn’t enough public defensive data to accurately project Tatis’s defensive production in the outfield. But what we can say is that he is a superstar talent with all the elite speed and acceleration that great outfielders tend to have and that is generally necessary to making low probability catches. It sets a great ceiling for his defensive potential and leaves us largely to wonder how well he can handle the routine and the skills inherent to playing the outfield.