We May Somehow Still Be Underhyping Shohei Ohtani
If looking back at the phenom’s one-WAR week made your eyes pop, try looking ahead: One projection system thinks the Angels’ two-way star could be a hybrid of Kris Bryant and Noah Syndergaard … or maybe even of, unfathomably, Giancarlo Stanton and Clayton Kershaw
Two weeks into the season is far too early to think about All-Star selections, but indulge me for a moment. Here’s one potential set of All-Star starting lineups this July:
Catcher: Buster Posey in the NL, Gary Sánchez in the AL
First base: Cody Bellinger, Edwin Encarnación
Second base: DJ LeMahieu, José Altuve
Shortstop: Corey Seager, Francisco Lindor
Third base: Nolan Arenado, José Ramírez
Outfield: Rhys Hoskins, Yoenis Céspedes, and Andrew McCutchen; Justin Upton, Andrew Benintendi, and Adam Jones
The lineups are a little light on star power and missing Mike Trout and Bryce Harper, but they’re not too outlandish. A few extra injuries and untimely slumps for the league’s best players could result in All-Star ballots that look just like that one. Oh, and there’s another thing that all of those players have in common: One prominent projection system says Shohei Ohtani is already a better hitter than any of them.
Move to the mound, and the forecast from that system, Steamer, is even more bullish on the young Angel’s prospects: It predicts just five starting pitchers in all of baseball to record a better ERA over the rest of the season—Clayton Kershaw, Chris Sale, Noah Syndergaard, Corey Kluber, and Max Scherzer. And it already thinks Ohtani will produce the best strikeout rate of any starter over the remainder of 2018, bumping Scherzer down to second place and Sale to third in that metric.
Those assessments might sound outrageous so early in the season, before Ohtani has even reached 25 career plate appearances or pitched against a team other than the Athletics, but it’s impossible to imagine anything Ohtani can’t accomplish at this point. He’s hit three home runs and slashed .368/.400/.842 in five games at the plate, and he’s posted a 2.08 ERA and struck out 18 batters in 13 innings on the mound. He was worth a full win above replacement in just one week, and he’s been the best player on a team that stars Trout.
But rather than temper the excitement, as they are often wont to do, the projections confirm the ebullience surrounding Ohtani’s debut. As purely mathematical, rational algorithms, they can’t be swept up in the hype or the spectacle of the first real two-way player in a century or all the “... since Babe Ruth” modifiers that designation brings, yet Steamer still believes that Ohtani is a top-six pitcher and top-25 hitter in the sport. Already. As a 23-year-old just two weeks into his MLB career.
To be fair to Ohtani from an expectations perspective—and to, say, Altuve, who won last year’s MVP award and now has a worse batting forecast than that of a rookie with 20 career plate appearances—Steamer is the most bullish on the pitcher-slash-DH among the major public projection systems. Ohtani’s outlook at PECOTA, which Nate Silver invented and is hosted at Baseball Prospectus, has barely budged since the preseason, and ZiPS, which Dan Szymborski operates, thinks the world of him as a pitcher but is the least encouraged by his bat.
The three systems are in agreement, at least, that Ohtani is a top pitcher. By a small margin, Steamer is the most enthused (particularly concerning his best-in-baseball strikeout projection) because it considers fastball velocity as part of its calculations. That’s why Steamer was optimistic about his K potential before he threw a single MLB inning, when Ohtani already ranked fourth among starters in Steamer’s strikeout rate projection.
“Coming into the year, we were expecting him to have a 97.5 mile-an-hour fastball, so our expectation around that would be that he’s going to be pretty good,” says Jared Cross, who runs the Steamer engine. “And then we’re adjusting that based on his strikeout rates in Japan.” Per the Japanese sabermetric company DeltaGraphs, Ohtani led the NPB in strikeout percentage in 2014, 2015, and 2016 (minimum 140 innings) before missing much of the 2017 season due to injury, so that adjustment only confirmed what his elite velocity already suggested about his talent level. Once he came to the majors and both started striking out opposing batters at a prolific rate and threw even harder than expected—his 98 mile-per-hour average fastball velocity ranks third in the majors thus far, behind only Luis Severino and Syndergaard—that projection received a slight boost, propelling Ohtani to the no. 1 spot.
But again, the other systems are in general agreement about his acehood. They differ on his hitting projection, though, and in particular how to adjust in response to his early home run exploits. Once again, Steamer is the most enthusiastic. It is generally the twitchiest of the prominent projection systems, with the most sensitive responses to in-season events—but even by those standards, Ohtani stands out for the Steamer strides he’s made. Before the season, the system forecast Ohtani to finish with a 116 wRC+, meaning he would be 16 percent better than a league-average hitter after adjusting for park and league context. After just a few games at the plate, though, his rest-of-season projection now calls for a 128 wRC+, giving him the most drastic change of any player who’s batted this year.
Much of that massive jump comes not from his impressive box-score results but from the underlying batted-ball data that supports them. “If it weren’t for Statcast data, I don’t think we could have a jump that large for any player,” Cross says. Steamer began incorporating Statcast data “in a simple way” last year, so the system doesn’t yet use launch angle or sprint speed but does include exit velocity as a factor in its projections.
“If a player is hitting the ball hard, we did some tests and said, ‘Hey, his projection should come up … even beyond what the results of those hard hits are,’” Cross says. Of particular relevance to Ohtani is that Steamer gives extra weight to the most extreme exit velocities, and Ohtani has already hit two batted balls in the air at 112-plus miles per hour. Only eight other players have managed that feat this season.
Entering play Tuesday, Ohtani had also posted the highest average exit velocity of any player with at least 10 batted balls this season, and his early, powerful performance “tells us something right off the bat even in a small number of plate appearances,” Cross says. That’s especially important given the limited batting data available on the two-way player before this season: Many of his plate appearances in Japan came years ago, when he was a teenager, and he reached 250 in a season only once. Even the best prospects who advance through the minor leagues stateside will have recorded 500-plus plate appearances in the year preceding their promotion.
Entering the season, then, Ohtani’s hitting projection came with considerable variation between the best- and worst-case scenarios, and that level of uncertainty means “just the small number of exit velocities he has in the majors this season is given considerable weight,” Cross says. It’s hard to fake hitting the ball as hard as Ohtani has, so he essentially made the best possible first impression in his first few trips to the MLB plate.
ZiPS, however, doesn’t use Statcast data to inform its in-season projection adjustments, and PECOTA doesn’t use Statcast at all, so Ohtani’s early exit velocities haven’t registered at all with those systems. The difference in approach stems in part from sheer computational concerns—Szymborski says via email that it’s easier to crunch all the data in-season with a simpler model—and in part because the Statcast information is new enough that it’s less clear how a factor like exit velocity affects a player’s future outlook than, for instance, strikeout and walk rate.
That contrast explains why PECOTA would have exhibited practically no reaction to Ohtani’s early performance, no matter how impressive. As Baseball Prospectus’s Rob McQuown says via email, “PECOTA’s historically been the most conservative of any projection system. And the weights used on preseason projections are weighted far more than any in-season stats.” In other words, McQuown says, PECOTA is somewhat undervaluing Ohtani’s in-season stats when constructing his rest-of-season projection, but that inertia has nothing to do with Ohtani specifically. “If any other batter who was projected to go .269/.335/.451 [PECOTA’s preseason slash line for Ohtani] started off like Ohtani has, his projection wouldn’t change much, either,” he adds.
Plus, PECOTA was already bullish on Ohtani before the season even began based on his statistics in Japan. Because Ohtani had experienced such statistical success against relatively tough competition there, his baseline rookie-year projections were rather enthusiastic compared with the usual output for a top prospect (Matt Wieters infamously excepted). PECOTA pegged Ohtani as a top-20 starting pitcher and top-60 hitter before the season, so even though he hasn’t improved on those valuations yet, it’s not as if this system doubts his abilities.
But let’s consider the possibility that his first two weeks are reason enough to tilt his projection upward, and that his loud velocity—both pitching and hitting—averages are suggestive of his potential, and that his rest-of-season numbers from Steamer aren’t too outlandish. What would that look like for his full-season numbers?
Fold in Ohtani’s existing at-bats, which entering play Tuesday had produced a 279 wRC+, to his rest-of-season batting projection, and his total wRC+ for the season settles at 137. For comparison, Kris Bryant’s rookie wRC+ was 136. Fold in Ohtani’s existing innings, meanwhile, to his rest-of-season pitching projection, and his total ERA for the season settles at 3.23 (with a 3.26 FIP to match). For comparison, Syndergaard’s rookie ERA was 3.24. Steamer thinks Ohtani boasts the combination of rookie Bryant’s bat and rookie Syndergaard’s arm. No amount of hyperbole is excessive in expressing such a phenomenal concept.
And yes, those numbers are likely too eager given that Steamer carries the most enthusiastic projections for Ohtani with the most drastic adjustments after his first two weeks. But a projection system like Steamer functions by generating a wide range of possible outcomes and touting the average scenario as its single-data-point projection that appears on public websites and powers playoff-odds calculations. If Steamer’s outlook is accurate, that phenomenal concept is Ohtani’s average projection—meaning he’s just as likely to do even better than that than he is to fall short. “Even better” than half-rookie-Bryant, half-rookie-Syndergaard is unfathomable. What does such a thing even look like?
The best way to answer that question is to look at Ohtani’s 90th-percentile projections, which represent the nearest approximation of his ceiling. Cross shared those newly updated numbers Tuesday. These are the most optimistic outcomes from the optimistic projection system, so Ohtani is unlikely to approach them. But still, just swirl them around and see how delicious they taste: His 90th-percentile rate hitting stats are about what Giancarlo Stanton did last year, and his 90th-percentile rate pitching stats are about what Clayton Kershaw did last year. That’s absurd. That’s unimaginable. That’s two MVP players in one. But it’s not impossible. For Ohtani thus far, nothing has been.
Projections accurate through Tuesday morning.
