The Real Winners of Spring Training

Spring Training stats don’t matter. Right? Right. Well, mostly. It’s true — you shouldn’t be paying attention to things like batting average, ERA, or even home-run figures in the spring. There’s just too little time, too much volatility in the statistics, and too much uncertainty surrounding the quality of opposition for those numbers to have, well, any meaning. But last year, Dan Rosenheck’s excellent work in The Economist, later nominated for a SABR Analytics Conference Research Award, revealed that certain peripheral Spring Training statistics actually can have some predictive value for the regular season.

It needs to be stressed: even then, the effects are small. Nothing that happens in Spring Training should drastically alter your perception of a player. And for most guys, nothing should change. But, for the few players at the very end of each spectrum in these particular statistics, it’s okay to move your expectations up or down a tick or two.

You should read Rosenheck’s article and also view the slides he used to present his research at the MIT Sloan Sports Analytics Conference, but I’ll briefly summarize his findings here anyway:

• For batters, the three most predictive statistics that stabilize most quickly are: strikeout rate, walk rate, and isolated power on contact.
• For pitchers, the three most predictive statistics that stabilize most quickly are: strikeout rate, walk rate, and ground-ball rate.
• Each player’s Spring Training figures should be compared against his own projections to find the largest outliers.
• We can learn the most in the spring about younger players, who have less major-league playing time, therefore less significant data to fuel the projections, therefore more uncertainty within those projections

Using those four basic principles from Rosenheck’s work, we can fairly easily use the Spring Training leaderboards from MLB.com and our depth chart projections here on the site (a mix of ZiPS and Steamer projections with author-updated playing-time estimates) to find the players who changed their outlook the most this spring (though still not that much!).

The Hitting Winners

Jake Lamb

• 2016 projections: 25.6 K%, 8.4 BB%, .200 isolated power on contact
• 2016 spring stats: 24.2 K%, 19.7 BB%, .405 isolated power on contact

By this measure, Arizona’s third baseman Jake Lamb has had the single most encouraging spring of any batter in baseball. He’s still striking out more than the average batter, but he also has the highest walk rate of any qualified batter in the spring — more than double his projected rate — and he’s doubled his power output, perhaps thanks to a change in his swing path, inspired by his teammate, A.J. Pollock.

Phil Gosselin

• 2016 projections: 17.2 K%, 4.5 BB%, .126 isolated power on contact
• 2016 spring stats: 11.5 K%, 9.8 BB%, .375 isolated power on contact

Well, I guess Diamondbacks hitters have had a fruitful spring. Phil Gosselin seems likely to open the year as a reserve in Arizona, but that could change if he keeps this up. And while there’s no fancy swing change with Gosselin (that I know of), it’s really just an extension of what he did last year in Arizona, after being acquired from Atlanta for Touki Toussaint. Gosselin exceeded expectations at the plate with surprising power (granted, in 24 games), and he’s showed that same pop again in the spring.

Ketel Marte

• 2016 projections: 15.0 K%, 5.9 BB%, .134 isolated power on contact
• 2016 spring stats: 7.1 K%, 12.5 BB%, .244 isolated power on contact

Ketel Marte burst on the scene as a 21-year-old rookie in Seattle last year and has now positioned himself as the team’s shortstop of the future. He posted a high on-base percentage in 2015 thanks to a plus walk rate, though the projections didn’t see that as sustainable, seeing as he walked more than he ever had in the minors. Well, Marte did everything in his power to change that perception this spring, walking nearly twice as often as he struck out while hitting for some added pop, as he makes his case for Seattle’s leadoff man this year.

The hitting losers: Joc Pederson, Jon Singleton, Oswaldo Arcia.

The Pitching Winners

Aaron Sanchez

• 2016 projections: 16.9 K%, 11.0 BB%, 61.8 GB%
• 2016 spring stats: 25.3 K%, 4.0 BB%, 63.2 GB%

No young pitcher had a more encouraging spring than Aaron Sanchez. Sanchez put on 25 pounds in the offseason in an effort to smooth out his mechanics that led to such a high walk rate as a starter last year, and then had, by far, the largest positive deviation from his projected walk rate of any pitcher in the spring. He also had the third-largest uptick in strikeout rate. And he did all that without giving away any of his already-elite ground-ball rate, which was the highest of any spring pitcher. Doesn’t get much better than that. Now that’s how you earn back a rotation spot.

Shane Greene

• 2016 projections: 15.0 K%, 7.2 BB%, 46.6 GB%
• 2016 spring stats: 30.3 K%, 5.3 BB%, 40.0 GB%

Speaking of earning back a rotation spot, here’s Shane Greene, who was aided in doing so by an injury to Daniel Norris, but may give the Tigers a difficult decision upon Norris’ return if his spring is any indication. Greene had the second-highest strikeout rate of any qualified pitcher in the spring and is reportedly touching 97 with his fastball, a year after dealing with diminished velocity, perhaps due to early-season elbow problems and an artery issue in his throwing hand that caused numbness off-and-on throughout the season. Greene looks to again be healthy, and is pitching a lot more like the exciting Yankees rookie in 2014 that the Tigers thought they were acquiring when they traded Robbie Ray for him.

Anthony DeSclafani

• 2016 projections: 19.2 K%, 6.8 BB%, 43.6 GB%
• 2016 spring stats: 24.1 K%, 3.6 BB%, 48.4 GB%

Here’s a perfect example of why Spring Training ERA is meaningless: Anthony DeSclafani’s was 8.38. Don’t let that fool you; DeSclafani had one of the most impressive springs of any young pitcher. Unfortunately, a strained oblique has sidelined his progress, and he’ll open the season on the 15-day disabled list. Hopefully he’s able to regain his spring form upon his recovery.

The pitching losers: Jose Fernandez, Matt Shoemaker, Danny Salazar.

A couple caveats and clarifications:

1. Only the qualified batters and pitchers from MLB.com’s leaderboards were considered for this post. That gives us a player pool of 136 batters and 56 pitchers.
2. Because ground-ball rate isn’t published on MLB.com, we used Rosenheck’s calculation of 0.8*(GO/(GO+AO)) for a rough ground-ball rate estimate.
3. Because ground-ball rate also isn’t a projected statistic, I instead used a pitcher’s ground-ball rate over the last three years (if available) in place of a projection
4. Keeping consistent with Rosenheck’s finding that spring stats tell us less about older players with more certain 2016 projections, we only considered batters with fewer than 1,000 career plate appearances and pitchers with fewer than 350 career innings pitched (each about two season’s worth of playing time).