WAR’s Poor Utility for Utility Players

I realize that we’ve collectively agreed that Wins Above Replacement is the only baseball statistic that matters, and thus it’s conventional wisdom that Pete Crow-Armstrong has by now all but locked up this year’s National League MVP. Now, I think WAR is a useful stat, and I’ve frequently referred to it throughout my own writings on baseball, but I have a hot take: you can’t possibly summarize an entire baseball season with a single number.

Even WAR isn’t actually a single number: are we talking Baseball Reference WAR (rWAR), FanGraphs WAR (fWAR), or Baseball Prospectus WAR (WARP)? (I’m not even sure those are the only three formulations for modern players⁽¹⁾; they’re just the only three I ever see cited.) And even if we settle on one of those specific sites’ formulations (or agree to take the arithmetic or geometric means of two or three of them), they each have separate pitching and offensive/defensive components, and we’ve arbitrarily decided to simply add those together to evaluate the performances of players who do both. But should we? (Spoiler: No.)

Wins Above Replacement is generally good at comparing two players who play the same position. It is not as reliable at comparing multiple players who play different positions. The positional adjustment was introduced as a means to combat this, but it’s based on defensive metrics that surely even its staunchest defenders must admit is based on some rather sketchy data, even in the Statcast era. And such metrics are only reliable for players that consistently play the same position: adding players’ WAR at each position does an inadequate job of summarizing utility players’, well, utility.

Take Ben Zobrist. Over the course of his career, he played 6925 innings at second base, 3215⅓ in right field, 1779 at shortstop, 1074⅔ in left field, 203 in center field, 117 at first base, 44⅓ at third base, and 1 at pitcher. While his best position by far was second base, he provided above-average defense in the outfield and generally “good enough” defense in the other positions.

More to the point, Zobrist’s ability to switch positions mid-game allowed Rays manager Joe Maddon and Cubs manager Joe Maddon (sorry, couldn’t resist) to play other players who otherwise wouldn’t have seen playing time. As Foolish Baseball points out in his video on Zobrist, those players got WAR for their appearances, which is an inadequate way of measuring Zobrist’s value to his teams, since those appearances wouldn’t have happened had their teams’ roster not boasted a player as versatile as Zobrist.

Zobrist fell off the ballot in his first year of Hall of Fame eligibility without accruing a single vote, probably because he only had 44.7 rWAR, 42.7 fWAR, and 37.6 WARP in his career. But none of those numbers reflects his actual value to his teams. Baseball Prospectus itself called him “one of the most underrated performers of his generation,” and his Cooperstown ballot results proved them correct.

And what is a two-way player, if not the ultimate utility player? I have two numbers for you: five and twenty-two. Between March 26, 2026, and July 3, 2026 (Ohtani’s final pitching start so far this year), the Dodgers’ bullpen ranked fifth place among MLB teams in Expected Fielding Independent Pitching (xFIP), with 3.85. From July 4 to August 31 (when I wrote the bulk of this piece), it has ranked twenty-second in xFIP, with 4.41. Prefer FIP? They were fifth (3.62) when he was pitching and have been twenty-first (4.31) since. Their xERA has plummeted from eighth (3.72) to eighteenth (3.86), and even their ERA has plummeted from tenth (3.74) to twentieth (4.18). They’ve gone from sixth in home runs per nine innings (0.90) to seventeenth (1.14); they’ve gone from sixteenth in bases on balls per nine (3.74) to dead last (4.89). Only two stats have improved, and not meaningfully so: they’ve gone from third place in strikeouts per nine (9.56) to second (10.07), and they’ve gone from tenth place on batting average on balls in play (.277) to seventh (.267), both of which could easily be statistical noise given how little movement that represents.

A more digestible summary of the above information, with additional columns measuring how many places they dropped on the overall rankings, the subtractive difference in the value of each stat, and how many percentage points worse each stat got without Ohtani in the starting rotation. As we can see, Expected Earned Run Average is effectively a rounding error, but their other stats got substantially worse, especially their walks and home runs (thus explaining why their FIP and xFIP got so much worse despite their increase in strikeouts).

(I’ll avoid making the obligatory “nice” joke in response to the FIP shift, which isn’t nice at all.)

Ohtani hasn’t made a single appearance in relief this season (largely because of the so-called “Ohtani rule” allowing a two-way player who starts the game on the mound to remain in the game as a designated hitter after the bullpen takes over), so his presence or absence has had no direct effect on any of the bullpen’s stats.

And yet it’s difficult to overstate the impact Ohtani’s absence has had on their efficacy. These are easily measurable stats. Games are won and lost on the margins, and just in the last week, the Dodgers have lost four one-run games (including an entire series with the Braves, which, lest anyone say playoffs aren’t won or lost in August, will force the Dodgers into playing a three-game wild-card series should they wind up tied with the Braves for the NL’s second-place record). They’ve lost eleven one-run games since Ohtani stopped pitching. That’s roughly 23% of the games they’ve played in that span!

Now, I’m not saying they’d have won all eleven of those games if Ohtani had made every sixth start this whole time. But they’d surely have won at least a couple more. Had they won just four of them, they’d currently have MLB’s best record and would be able to breathe a bit more easily.

Even if Ohtani doesn’t return to the mound at all this season, his half-season of pitching looks similar to Éric Gagné’s 2002-2004 averages, which includes 2003, when he became the final reliever to win a Cy Young Award to date. Gagné’s 2003 numbers were exceptional (1.20 ERA, 0.86 FIP, 0.692 WHIP, 15.0 K/9, 337 ERA+), and Ohtani didn’t quite manage those (although for his first ten starts, his 0.787 WHIP was comparable to Gagné’s, his 0.74 ERA was even better, and his 546 ERA+ was overwhelmingly better). Across 2002-2004, Gagné posted a 1.79 ERA, 223 ERA+, 1.57 FIP, and 0.822 WHIP, averaging 13.3 K/9 and somehow pitching 82⅓ in each of those three years. To date, Ohtani has pitched 85⅔ innings with a 1.79 ERA, 227 ERA+, 2.59 FIP, 0.946 WHIP, and 10.0 K/9. So, basically, he’s been prime Éric Gagné on the mound in every measure except the strikeouts.

For those not inclined to compare starting pitching to relief pitching, then consider Paul Skenes’ numbers last year, when he won the NL Cy Young Award unanimously: 1.97 ERA, 226 ERA+, 2.36 FIP, 0.948 WHIP, 10.4 K/9, 187⅔ IP. Every single one of those numbers is within 0.20 of what Ohtani was doing before his injury sidelined him; several are mere rounding errors.

In what’s supposedly a down year for him, Ohtani’s offensive performance has been roughly on par with Juan Soto’s offense in 2025. Soto batted .263/.396/.525 last year, for a .921 OPS or 160 OPS+. Ohtani is hitting .280/.379/.529, for a .909 OPS or 150 OPS+. So Ohtani actually has slugged slightly better (although he probably won’t wind up with as many home runs as Soto) and has hit for a noticeably higher average, while Soto got on base more.

So, one player is effectively both prime Éric Gagné (or prime Paul Skenes) and present-day Juan Soto. How do we evaluate that? Wins Above Replacement tells us to simply add those numbers together. But that’s clearly wrong-headed: as I’ve shown, Ohtani’s absence from the mound has trickled down to the entire bullpen. There’s a simple reason for this: under MLB’s rules, a two-way player effectively gives his team an extra bullpen slot.

Now, Ohtani may be the only active two-way player in MLB, but for starters, three active MLB players won the John Olerud Two-Way Player of the Year Award in college: Skenes (Pittsburgh Pirates), Spencer Schwellenbach (Atlanta Braves), and Jac Caglianone (Kansas City Royals). Of course, it’s foolish to think those three would be as strong in both roles as Ohtani is, but there isn’t a team in MLB that wouldn’t benefit from getting, say, Max Muncy-level offense and Martín Pérez-level pitching from the same player.

So what should we do instead? I’m not a statistician, so I’m probably not the best person to answer this question. Multiplying those numbers together would be even more volatile than adding them, and I’m not sure it’s a valid approach either. (The multiplicative approach gives Ohtani 10.15 rWAR, 11.31 fWAR, and 11.04 WARP, for a geometric mean of 10.822; the additive approach gives him 6.5 rWAR, 6.8 fWAR, and 7.0 WARP, for an arithmetic mean of 6.767.) Maybe we just take the means of the additive and multiplicative approaches and call it a day (8.325 rWAR, 9.055 fWAR, 9.020 WARP), but that feels unsatisfyingly arbitrary. There must be a better way of quantifying utility or two-way players’ value to their teams’ roster flexibility than simply adding up their WAR at each position.

None of this actually says whether Ohtani has been more valuable than Crow-Armstrong this year. I lean towards thinking that a full season of Juan Soto-caliber offense and half a season of Cy Young-caliber starting pitching grades out above a full season of Aaron Judge-caliber offense and Willie Mays-caliber defense in center field, but I have no definitive means to prove that. One could argue that the Cubs wouldn’t be a playoff team without Crow-Armstrong, but anyone who makes that argument needs to prove that the Dodgers would be a playoff team this year without Ohtani, and given how much his absence from the mound has affected their bullpen, I’m not sure that case is as clear-cut as people think it is.

In any case, I can say that if a person thinks Wins Above Replacement is the only number they need to cite to make their case, that’s a clear sign they know just enough ball to think they know way more ball than they do. Don’t be that person.

—Aaron Freed
—Tallahassee, FL
—August 31, 2026 (revised September 2, 2026)

Back to top · My portfolio · Contact me · Website index

Endnotes

Back to top · My portfolio · Contact me · Website index