Bellinger, Devers, and MLB’s Most Improved Position Players
FanGraphs contains multitudes. Multiple flavors of Wins Above Replacement — one in which the pitching component is driven by FIP, the other by actual runs allowed (RA9-WAR) — for one thing. Multiple projection systems (Steamer and ZiPS) and ERA estimators (FIP, xFIP, SIERA). Multiple measures for defense, pitch selection, and plate discipline, borne of different data feeds. Multiple ways of measuring playoff odds and projected won-loss records. Multiple depth charts, now that we’ve brought Roster Resource on board. There’s a lot of cool stuff… if you know where to look.
One of the cool but relatively new and lesser-known features is our Season Stat Grid, introduced just over a year ago, and in the planning stages for longer (I know that I’m one of the people who lobbied for the tool). The grid allows the user to view 11 year worth of data in a single category, and to track and rank year-to-year totals and changes based on thresholds of plate appearances and innings. It’s hours of fun, and occasionally fuel for an article. So after highlighting the exceptional, breakout season of Rafael Devers and noting that — at the time it was written, at least — he had the majors’ largest year-over-year improvements in batting average, on-base percentage, and WAR, while ranking second in his gains in wRC+ and fourth in slugging percentage, I figured the topic was worth a league-wide look.
Towards that end, I chose 10 statistical categories where we might look for significant changes, namely the aforementioned five plus walk and strikeout rates, out-of-zone swing rate, fielding (UZR plus positional adjustment, if any) and, for a nod towards win expectancy, Clutch. To qualify, players had to reach 400 plate appearances last year and 300 thus far this year. I then took the top 30 players whose changes went in the right direction (higher in all cases except for strikeout and chase rates), awarding 30 points for first place, 29 for second, and so on. When two or more players were tied — even if it was just a virtual tie, where we can’t see what’s to the right of the displayed decimals — I split the points evenly among the tied; for example, James McCann and Yoán Moncada, who have increased their batting average by 66 points apiece, occupy the third and fourth spots and thus each get 27.5 points. I doubled the impact of WAR and wRC+ (60 points maximum), even though components of those are included elsewhere within the survey, on the belief that those two stats drive the lion’s share of our understanding as to who has improved.
I’ll make no claim that this is science, and one can certainly quibble with my choices of stats and methodology until the cows round third base and head home. For one thing, this method steamrolls any sense of scale; Tim Anderson’s 93-point improvement in batting average is worth just 2.5 points more than the gains of McCann and Moncada, with Devers’ 86-point improvement worth just one point more than the pair. Nonetheless, here’s the top 20 that my methodology yielded, with the points they received in each category:
| Rk | Player | Pts | AVG | OBP | SLG | Fld | wRC+ | BB% | K% | O-Sw% | Clutch | WAR |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | Cody Bellinger | 256.5 | 23 | 27 | 30 | 17 | 54 | 19.5 | 29 | 0 | 0 | 57 |
| 2 | Rafael Devers | 245.5 | 29 | 30 | 28 | 9.5 | 59 | 0 | 30 | 0 | 0 | 60 |
| 3 | Carlos Santana | 209.5 | 26 | 25.5 | 22 | 0 | 47 | 0 | 0 | 14 | 23 | 52 |
| 4 | Scott Kingery | 198.5 | 21 | 25.5 | 24 | 0 | 52 | 8 | 0 | 25 | 0 | 43 |
| 5 | James McCann | 195.5 | 27.5 | 28 | 27 | 0 | 59 | 0 | 0 | 0 | 0 | 54 |
| 6 | Yoán Moncada | 191.0 | 27.5 | 16.5 | 20.5 | 16 | 43 | 0 | 24.5 | 0 | 0 | 43 |
| 7 | Ketel Marte | 179.5 | 25 | 22 | 23 | 9.5 | 43 | 0 | 0 | 0 | 0 | 57 |
| 8 | DJ LeMahieu | 167.0 | 24 | 24 | 13 | 0 | 56 | 0 | 0 | 0 | 0 | 50 |
| 9 | Tim Anderson | 157.5 | 30 | 29 | 10 | 0 | 50 | 0 | 18.5 | 0 | 20 | 0 |
| 10 | Trey Mancini | 148.5 | 9.5 | 19 | 15 | 0 | 38 | 10 | 9 | 0 | 0 | 48 |
| 11 | Marcus Semien | 138.0 | 0 | 13.5 | 11 | 0 | 34 | 19.5 | 17 | 14 | 5 | 24 |
| 12 | Carlos Correa | 126.0 | 19.5 | 10.5 | 25 | 0 | 47 | 0 | 0 | 0 | 0 | 24 |
| 13 | George Springer | 124.5 | 9.5 | 8 | 19 | 8 | 36 | 0 | 0 | 10 | 0 | 34 |
| 14 | Evan Longoria | 123.0 | 0 | 13.5 | 0 | 22 | 6 | 27.5 | 0 | 0 | 24 | 30 |
| 15 | Nelson Cruz | 116.5 | 22 | 20.5 | 26 | 0 | 40 | 0 | 0 | 2 | 0 | 6 |
| 16 | Mark Canha | 108.5 | 0 | 23 | 0 | 0 | 29 | 30 | 0 | 26.5 | 0 | 0 |
| 17 | Kole Calhoun | 96.5 | 0 | 13.5 | 12 | 0 | 32 | 0 | 0 | 0 | 0 | 39 |
| 18 | Nick Ahmed | 86.5 | 17.5 | 16.5 | 0 | 0 | 2 | 4.5 | 0 | 28 | 0 | 18 |
| 19 | Josh Bell | 82.0 | 0 | 0 | 29 | 0 | 23 | 0 | 0 | 0 | 0 | 30 |
| 20 | Miguel Rojas | 79.5 | 17.5 | 20.5 | 0 | 0 | 0 | 16 | 0 | 19.5 | 0 | 6 |
That feels like a pretty good list, with some obvious breakouts and rebounds that have drawn varying degrees of attention. Category-wise, only a small handful of the players who improved the most in some areas — clutch, fielding, and strikeout rate, in particular — have made improvements on a similar scale in other areas, but those who scored highest via this junk-drawer methodology were the ones who tapped into those areas.
Let’s go through the top 10, lightning-round style. All stats are through Wednesday.
1. Cody Bellinger, Dodgers
After a flawed but nonetheless quite solid sophomore season (120 wRC+, 3.6 WAR), the 24-year-old Bellinger has been unreal this year after restoring his mechanics over the winter. He’s posted significant improvement in all three slash stats, cut his strikeout rate from 23.9% to 16.2% while improving his walk rate from 10.9% to 14.1%, and given the Dodgers elite defense in right field (9.0 UZR, 19 DRS) instead of playing regularly (and well) at first base. He’s one of only two players in this exercise who scored in eight out of 10 categories (the 11th-ranked Semien was the other). He entered Thursday leading the NL in homers (42) and WAR (6.9), with slight edges over reigning MVP Christian Yelich. He may well win some hardware of his own thanks to his across-the-board improvements.
2. Rafael Devers, Red Sox
As I noted on Tuesday, Devers turned in something of a dud last year (90 wRC+, 1.0 WAR), his age-21 season, but has significantly improved his two-strike approach while cutting his strikeout rate from 24.0% to 16.7%. He’s now among the top three in improving each slash stat and wRC+ (145), and his jump to 5.4 WAR is the widest in baseball. If it weren’t for some guy named Mike Trout, he’d be a true MVP candidate.
3. Carlos Santana, Indians
Santana’s one-year foray to Philly wasn’t as bad as his .229 batting average would lead you to believe (109 wRC+, 1.9 WAR), but he’s enjoyed a strong bounce back since returning to Cleveland. Thanks to a more disciplined approach, the 33-year-old first baseman’s .413 on-base percentage is second in the league, and his .552 slugging percentage is sixth. He’s been on fire lately, with six homers in a 10-game span, and leads the resurgent Indians both in homers (30) and WAR (4.4).
4. Scott Kingery, Phillies
A well-regarded prospect, Kingery flopped as a rookie (62 wRC+, -0.1 WAR), that while generally playing shortstop, a position where he had almost no previous experience as a professional. With César Hernández still a fixture at second base, the 25-year-old Kingery has bounced around the diamond again, spending most of his time in center field — where Odúbel Herrera‘s domestic violence suspension and Andrew McCutchen’s injury have thinned the ranks — and third base, where Maikel Franco has again disappointed. He’s cooled off considerably since the start of the season, but his 108 wRC+ and 2.0 WAR have put his career back on track.
5. James McCann, White Sox
McCann spent the past four years as the light-hitting regular catcher for the Tigers, meandering around replacement level, but after being nontendered following a dreadful season (58 wRC+, -0.5 WAR), the 29-year-old backstop has played his way into his first All-Star appearance. His boost from last year has been driven by top-five improvements in his slash stats, as he’s gone from hitting .220/.267/.314 to .286/.339/.470, accompanying a 98-point jump in BABIP (from .282 to .380) with a doubling of his home run rate (from once every 57 PA to once every 27 PA).
6. Yoán Moncada, White Sox
The player who topped our 2017 prospect list at age 22 didn’t embarrass himself in his first two seasons in the majors, but he’s taken some major steps forward this year. A more aggressive approach — his O-Swing% has jumped from 23.3 % to 33.2%, and his overall swing rate from 41.1% to 48.1% — has nonetheless resulted in a significant drop in strikeout rate (33.4% to 27.6%) accompanied by across-the-board improvements in slash stats; his .301/.358/.535 line, with a 134 wRC+ and 4.0 WAR, heralds his arrival as a star.
7. Ketel Marte, Diamondbacks
At bat, on the bases, and in the field, Marte has improved every major area of his game. While taking up center field (where he’s been outstanding to the tune of 7.6 UZR), he’s gone from a modest 104 wRC+ (.260/.332/.437) to a sizzling 141 (.319/.381/.574), while nearly doubling his home run output (from 14 to 27). His 5.7 WAR is topped only by Yelich, Bellinger, and Trout. Welcome to flavor country.
8. DJ LeMahieu, Yankees
In a season where every intended Yankees regular except Gleyber Torres has hit the injured list, the versatile LeMahieu has emerged as the team’s unsung hero. While splitting his time between second base (56 starts), third base (34 starts), and first base (21 starts), he’s shown that his unimpressive Coors Field stats and splits weren’t his true measure. With substantial improvements in all three slash stats (from .276/.321/.428 to .334/.381/.534) despite coming down from altitude, he’s gained 53 points of wRC+ (from 86 to 139), and with his .334 average, he’s neck-and-neck with Michael Brantley for the AL lead, giving him a shot at becoming the first player to win a bating title in both the AL and NL. What’s more, he’s hit .410/.450/.600 in 121 PA with runners in scoring position.
9. Tim Anderson, White Sox
Despite missing a month due to a right ankle sprain, and walking nearly half as often (2.6%) as last year’s meager 5.0% rate, the 26-year-old Anderson has significantly cut his strikeout rate and enjoyed a 102-point spike in BABIP (from .289 to .391), raising his batting average an major league-high 92 points (from .240 to .332) and carrying the rest of his slash stats with it. He’s already set a career high with 2.6 WAR; on a prorated basis, he’s been about twice as valuable as last season (2.0 WAR in 153 games).
10. Trey Mancini, Orioles
Though he hit 24 homers in each of the past two seasons, Mancini had enough holes in his game that he slipped below replacement level last year (-0.2 WAR, with a 93 wRC+). Driven by a nine-point drop in groundball rate (from 54.6% to 45.8%) and much-improved strikeout and walk rates, the 27-year-old Mancini has become one of the few bearable Orioles to watch. His 125 wRC+ leads the team, while his 2.1 WAR is second.
There’s more to be said about all of these players and the ones below them in the rankings, but I’ll cap it here. One surprising facet of these rankings is that it yields a complete lineup, with McCann catching, Santana at first base, LeMahieu (or Marte) at second, Anderson at shortstop, Devers at third, Mancini in left field, Kingery (or Marte) in center, and Bellinger in right, with Moncada and Marte (or his aforementioned alternatives) left over. In all, I think it’s a defensible, well-rounded list of players who have improved, and if I’m not tarred and feathered for this admittedly junk-drawer methodology, will return at some point with a similar one for starting pitchers.
Brooklyn-based Jay Jaffe is a senior writer for FanGraphs, the author of The Cooperstown Casebook (Thomas Dunne Books, 2017) and the creator of the JAWS (Jaffe WAR Score) metric for Hall of Fame analysis. He founded the Futility Infielder website (2001), was a columnist for Baseball Prospectus (2005-2012) and a contributing writer for Sports Illustrated (2012-2018). He has been a recurring guest on MLB Network and a member of the BBWAA since 2011, and a Hall of Fame voter since 2021. Follow him on BlueSky @jayjaffe.bsky.social.
This is one of the few lists that Trout can’t make, the price of consistency.
Jay, if you’re reading this, are you aware that your JAWS values correlate very strongly with career WAA? I compared the top 50 in career WAA with their JAWS values, with an r of .976. I expected a good correlation, because both are correlated with WAR, and both give more credit to obtaining the same WAR in fewer years. But I was surprised at just how close it was.
SHHH!!!! Trout will read this and put up a 15 WAR season next year.
Yes, and there was a point long ago where I gave thought to a WAA-based Hall metric. One reason for preferring the replacement level-based measure is that it does a better job of rewarding volume (career length), which is still the primary driver of who gets into the Hall and therefore worth accounting for in a metric that tries to define standards.
If you have two players with vaguely similar WAA, the one with the higher WAR is the more likely to be in, because his career was longer and his counting stats probably conform better to voter expectations. Not a perfect example but take Scott Rolen (44.1 WAA via B-Ref) and Brooks Robinson (39.7 WAA), both of whom hit for power and won several Gold Gloves while playing for many postseason teams including at least one champ. Robinson (78.4 WAR) was a first-ballot guy because he fleshed out his career totals while Rolen (70.2), who was done at 37, hasn’t cleared 18% in two tries.
Thanks. On that note, interesting that Grich, Whittaker and Utley (who probably won’t make it) all have higher WAAs than Sandberg, Alomar, and Biggio.
Fun stuff, thanks.
Are you doing pitchers too? Or are you just going to write a post about Lucas Giolito?
He mentions at the end that he is at least contemplating a for-pitchers version.
Fairly certain that the question was rhetorical, and rather than seeking an answer, MikeS was just pimping Lucas Giolito, who by all means, should be thoroughly pimped.
I understand breaking it out by different statistical categories instead of just ranking by WAR. It’s interesting to see where people have improved. I don’t understand summing the points, since WAR does that already but is better at it.
The cow is round third base, Dino Ebel is ringing the bell sending ’em home!
Considering that “Clutch” isn’t likely a repeatable skill and is calculated with a weird denominator that renders it uninterpretable, I’m not sure what we gain by looking to see who has “improved” in it.
Agreed. I think this is a fun exercise, but I would have left out Clutch.
I figured Kepler would be there.
He was 22nd, just missed the cutoff for the table. Points in AVG, SLG, wRC+ and WAR.
Hopefully I’m wrong but Kingery’s sell-by dates seems to have passed.
Unsurprisingly most of these guys improved from at or below league average to being pretty good.
Bellinger and Springer particularly stand out for being in this group despite having been well above average the previous season. Even Correa had a down year back to about a 100ish wRC+ in 2018 (following a 150 in 2017), ditto Santana.
A related idea, loosely based on the above, heavily cribbed from someone else’s comment in a different article or chat (who nailed it on Urshela) and entirely reliant yet another’s persons work (Zips; the combined Depth Charts leave too many guys out.)
Here are the 20 guys from whom the ROS wOBA, per ZIPS, is most improved from the pre-season wOBA (in percentage terms). Apologies for the formatting:
Name OldwOBA ROSwOBA % change
Giovanny Urshela 0.275 0.320 16.4%
Bryan Reynolds 0.280 0.320 14.3%
Eric Sogard 0.272 0.306 12.5%
Tommy La Stella 0.287 0.320 11.5%
Will Smith 0.278 0.306 10.1%
Roberto Perez 0.269 0.296 10.0%
Yordan Alvarez 0.322 0.354 9.9%
Fernando Tatis Jr. 0.302 0.332 9.9%
Hunter Dozier 0.290 0.318 9.7%
Austin Nola 0.262 0.287 9.5%
Hunter Pence 0.294 0.322 9.5%
Mitch Garver 0.296 0.322 8.8%
Danny Santana 0.298 0.324 8.7%
James McCann 0.277 0.301 8.7%
Christian Vazquez 0.286 0.308 7.7%
Hanser Alberto 0.281 0.302 7.5%
Cameron Maybin 0.307 0.329 7.2%
Ketel Marte 0.329 0.352 7.0%
Howie Kendrick 0.315 0.337 7.0%
DJ LeMahieu 0.316 0.338 7.0%
Interesting to me that the sample size is still too small to move the needle on someone like Aquino (who ZIPS still projects as a terrible hitter – .294 from .282 – despite destroying the ball across 400 professional PAs this year).
The list is not enormously different if you look at it on a total points basis instead of %:
Name OldwOBA ROSwOBA pt change
Giovanny Urshela 0.275 0.320 0.045
Bryan Reynolds 0.280 0.320 0.040
Eric Sogard 0.272 0.306 0.034
Tommy La Stella 0.287 0.320 0.033
Yordan Alvarez 0.322 0.354 0.032
Fernando Tatis Jr. 0.302 0.332 0.030
Hunter Dozier 0.290 0.318 0.028
Hunter Pence 0.294 0.322 0.028
Will Smith 0.278 0.306 0.028
Roberto Perez 0.269 0.296 0.027
Mitch Garver 0.296 0.322 0.026
Danny Santana 0.298 0.324 0.026
Austin Nola 0.262 0.287 0.025
James McCann 0.277 0.301 0.024
Ketel Marte 0.329 0.352 0.023
Christian Vazquez 0.286 0.308 0.022
Cameron Maybin 0.307 0.329 0.022
Howie Kendrick 0.315 0.337 0.022
DJ LeMahieu 0.316 0.338 0.022
Peter Alonso 0.333 0.355 0.022
At the risk (certainty?) of a cluttered data dump, here’s the other end of the spectrum. (Sidenote: wow, Jeff Mathis). I could probably triage this down to just players that matter – ie, the Peter Bourjos collapse is a lot less interesting than the Khris Davis and Matt Carpenter ones – but I’ve posted (almost) enough spam already:
Name OldwOBA ROSwOBA % change
Max Stassi 0.295 0.258 -12.5%
Aaron Altherr 0.310 0.273 -11.9%
Carlos Gonzalez 0.326 0.290 -11.0%
Erik Kratz 0.294 0.264 -10.2%
Zack Cozart 0.314 0.282 -10.2%
Scooter Gennett 0.334 0.304 -9.0%
Jeff Mathis 0.240 0.219 -8.7%
Bobby Wilson 0.234 0.214 -8.5%
Steve Pearce 0.336 0.308 -8.3%
Khris Davis 0.355 0.326 -8.2%
Travis Shaw 0.346 0.318 -8.1%
Tyler Austin 0.323 0.297 -8.0%
Jesus Aguilar 0.350 0.322 -8.0%
Peter Bourjos 0.264 0.243 -8.0%
Lucas Duda 0.322 0.297 -7.8%
Jung Ho Kang 0.331 0.306 -7.6%
Ruben Tejada 0.266 0.246 -7.5%
Ryan O’Hearn 0.311 0.288 -7.4%
Luke Maile 0.256 0.238 -7.0%
Matt Carpenter 0.363 0.338 -6.9%
If you sort this one by points, you get guys like Ramirez and Votto popping up:
Name OldwOBA ROSwOBA pt change
Max Stassi 0.295 0.258 -0.037
Aaron Altherr 0.310 0.273 -0.037
Carlos Gonzalez 0.326 0.290 -0.036
Zack Cozart 0.314 0.282 -0.032
Scooter Gennett 0.334 0.304 -0.030
Erik Kratz 0.294 0.264 -0.030
Khris Davis 0.355 0.326 -0.029
Steve Pearce 0.336 0.308 -0.028
Travis Shaw 0.346 0.318 -0.028
Jesus Aguilar 0.350 0.322 -0.028
Tyler Austin 0.323 0.297 -0.026
Jose Ramirez 0.386 0.360 -0.026
Lucas Duda 0.322 0.297 -0.025
Jung Ho Kang 0.331 0.306 -0.025
Matt Carpenter 0.363 0.338 -0.025
Joey Votto 0.383 0.359 -0.024
Ryan O’Hearn 0.311 0.288 -0.023
Nick Williams 0.324 0.302 -0.022
Tyler White 0.336 0.314 -0.022
Last one: top 40 pre-season woba, sorted by % change to ROS woba.
Name OldwOBA ROSwOBA pt change % change
Nelson Cruz 0.360 0.378 0.018 5.0%
Anthony Rendon 0.359 0.376 0.017 4.7%
Cody Bellinger 0.372 0.389 0.017 4.6%
Christian Yelich 0.384 0.399 0.015 3.9%
Alex Bregman 0.369 0.376 0.007 1.9%
Max Muncy 0.357 0.363 0.006 1.7%
Anthony Rizzo 0.371 0.375 0.004 1.1%
Charlie Blackmon 0.376 0.380 0.004 1.1%
Trevor Story 0.358 0.361 0.003 0.8%
Kris Bryant 0.366 0.369 0.003 0.8%
Ronald Acuna Jr. 0.364 0.366 0.002 0.5%
Freddie Freeman 0.383 0.385 0.002 0.5%
Mike Trout 0.424 0.426 0.002 0.5%
Justin Turner 0.365 0.365 0.000 0.0%
Joey Gallo 0.376 0.375 -0.001 -0.3%
Jose Altuve 0.361 0.360 -0.001 -0.3%
Michael Conforto 0.354 0.352 -0.002 -0.6%
Josh Donaldson 0.376 0.373 -0.003 -0.8%
Shohei Ohtani 0.364 0.361 -0.003 -0.8%
David Dahl 0.354 0.351 -0.003 -0.8%
Vladimir Guerrero Jr. 0.354 0.351 -0.003 -0.8%
J.D. Martinez 0.385 0.381 -0.004 -1.0%
Rhys Hoskins 0.363 0.358 -0.005 -1.4%
Nolan Arenado 0.378 0.372 -0.006 -1.6%
Eugenio Suarez 0.357 0.351 -0.006 -1.7%
Francisco Lindor 0.366 0.358 -0.008 -2.2%
Giancarlo Stanton 0.376 0.367 -0.009 -2.4%
Juan Soto 0.404 0.394 -0.010 -2.5%
Mookie Betts 0.384 0.374 -0.010 -2.6%
Manny Machado 0.357 0.345 -0.012 -3.4%
Aaron Judge 0.384 0.371 -0.013 -3.4%
J.T. Realmuto 0.356 0.342 -0.014 -3.9%
Paul Goldschmidt 0.364 0.349 -0.015 -4.1%
Bryce Harper 0.395 0.378 -0.017 -4.3%
Daniel Murphy 0.371 0.354 -0.017 -4.6%
Eloy Jimenez 0.364 0.343 -0.021 -5.8%
Joey Votto 0.383 0.359 -0.024 -6.3%
Jose Ramirez 0.386 0.360 -0.026 -6.7%
Matt Carpenter 0.363 0.338 -0.025 -6.9%
Khris Davis 0.355 0.326 -0.029 -8.2%
What about Danny Santana , Tommy LaStella, Gio Urshela, and Michael Brantley? All seem to having seasons well beyond anyone’s expectations.
I’m not sure if Santana had enough PA last season?
Mark Canha, Matt Chapman and Marcus Semien have all improved as well, perhaps Chapman is the least surprising on raw talent but the three have carried the team in the absence of and due to the mysterious (or blatently obvious injury) to Khris Davis.
There is no doubt there is something beyond wrong physically with Davis but Mark Canha has been full on beast mode this season.