Corbin Carroll Is Even Better Than Advertised

Not every out is created equal. Take this fly out from Corbin Carroll, for example:

A lot of things can happen when you make an out with the bases loaded. You could strike out, leaving every runner in place. You could hit into a double play, an inning-ending one in this case. You could ground out some other way, or hit an infield fly. But Carroll’s here was the most valuable imaginable; with one out, he advanced every single runner, including the runner who scored from third.

Mathematically speaking, you can think of it this way. The average out that took place with the bases loaded and one out lowered the team’s run expectancy by a massive 0.61 runs in 2024. That’s because tons of these outs were either strikeouts (bad, runner on third doesn’t score) or double plays (bad, inning ends). But Carroll’s fly out was far better than that. It actually increased the run expectancy by a hair; driving the lead runner home and moving the trail runners up a base is exquisitely valuable.

That’s not the only way this could have gone. Consider a similar situation, a groundout from Aaron Judge:

Like Carroll, Judge batted with a runner on third and fewer than two outs. In this situation, the average out is bad, lowering run expectancy by 0.514 runs. But Judge’s was obviously worse. It cost the Yankees all the expected runs they had left in the inning, naturally, which added up to just a bit more than 1.15.

One thing you could say is that both players made outs on this play, and that since players don’t control when and how their outs come, we should just give them all the same value. That’s a reasonable stance to take. FanGraphs’ marquee offensive stat, wRC+, treats all outs as identical. It’s not just us, though. Batting average, on-base percentage, slugging percentage, wOBA, you name it; “an out is an out is an out” is a core part of the way it works. Notwithstanding minor exceptions for “intentional outs” like sacrifices, pretty much everyone treats outs as interchangeable.

I wanted to dig a little deeper than that, though, so I started making some assumptions and doing some math. First, I looked at how costly an out was in every base/out state in 2024. Then, I broke outs into four categories: balls caught in the air, grounders that didn’t result in double plays, double play grounders, and strikeouts. For each base/out state, I calculated how much better or worse each out type was compared to the “average” out in that base/out state.

Let’s take another example, because I’m talking about a lot of numbers here and it’s easy to get lost. With the bases loaded and nobody out, the average out lowered run scoring expectancy by 0.565 runs in 2024. The average strikeout in that situation lowered run scoring expectancy by even more, 0.765 runs. This makes good sense to me; the “average” out scores a run fairly often, and a strikeout doesn’t. On the other hand, a fly out with the bases loaded lowers run expectation by only 0.494 runs. It’s slightly better than the average out – there are sacrifice flies to think of.

Continuing on, a non-double-play grounder is a great outcome, lowering run expectancy by a mere 0.164 runs. Since we know only one out was made, a groundball is amazing. Maybe the runner scored, maybe there was an error, maybe a couple runners advanced to boot. On the other hand, a double play is a disaster, lowering run expectancy by 1.091 runs. Sometimes that’s the runner at the plate plus the lead runner, sometimes it’s a classic 6-4-3 with a run scoring, but on average, hitting into a double play smarts.

You might wonder why I grouped all fly balls together but kept double plays separate from all other grounders. It’s a matter of accounting. Imagine hitting a medium-depth fly ball that the right fielder easily corrals. With a fast runner on third, that’s probably a sacrifice fly. With a slow runner, it’s probably just an out. But the batter doesn’t have control over that, and we already give baserunners credit, so giving the batter points only if a sacrifice fly is completed feels wrong. Instead, I merely assigned the average value across all fly balls, with baserunners getting their credit added or subtracted for their contributions separately.

Double plays aren’t like that. The batter exerts a ton of control over them. Take Carroll, for example. He’s no stranger to grounders, but he’s also hilariously fast. Like, turn-out-the-lights-and-be-in-bed-before-it-gets-dark fast. Carroll hit into just three double plays last year. Manny Machado, famously not fast, hit into 25. You also can’t hit into a double play if you don’t hit the ball on the ground, so fly ball artists do better than grounder-heavy types. The point is, hitting into a double play really is something you can put on the batter, so I do. Everything else, I merely point to the average outcome.

If you’re either very bored or reasonably good at computer programming, you can go through every single out that major leaguers made in 2024 and note how much better or worse they were than the average out in a given base/out situation. That’s more or less what I was doing above, but to be slightly more specific: A fly out with the bases loaded and one out is 0.224 runs better in expectation than an average out, so I credited Carroll with those extra runs. A double play grounder with one out and runners on first and third is 0.708 runs worse than an average out, so I docked Judge that amount. I did this for every single play in the 2024 season and then just hit sum.

My selection of Carroll to lead this column off wasn’t just random. His outs were a lot better than the average out made in the situations that he batted, to the tune of a whopping 8.5 runs in aggregate. Only one other player, Jackson Merrill, even topped six runs. This is a zero-sum game: For every out with better-than-average results, there’s an offset somewhere. It’s also a low-volatility game; 80% of players fell between -2 runs and 2 runs across the entire year.

Carroll’s game is made to break the mold of “average out,” though. That blazing speed means he almost never hits into double plays, which is a huge part of his score. He rarely strikes out – strikeouts are better than double plays but worse than everything else. He hits a decent number of grounders, too; they just never become double plays. That means he’s frequently advancing baserunners but almost never creating extra outs.

Merrill, the other standout of the productive out, succeeds for similar reasons. He rarely strikes out. He’s fast. He hit into only two double plays all year, albeit with fewer groundballs thanks to his batted ball mix. But it’s a similar formula: Put the ball in play but avoid the worst type of ball in play.

Those are reasonably large adjustments. Eight runs is the better part of a win; if you were to credit Carroll for his good outs, his value last season would’ve been closer to 5 WAR than the 4 WAR we marked him down for. I think there’s a good argument that he shouldn’t get all the credit: He doesn’t control who’s on base and how many outs there are when he bats, for example, and he had a ton of runners on third base when he batted with less than two outs — 36 to be exact, 20 more than Steven Kwan, to name a hitter with a similar batted ball distribution. But I feel comfortable saying that Carroll’s batted ball distribution and speed out of the box meant that his outs were less damaging than average, and in a way that none of our current hitting stats tabulate.

The other side of the coin? It’s Judge, of course. Judge cost the Yankees a whopping 8.8 runs relative to average with his deleterious outs in 2024, one of only three hitters who lost six or more runs in this accounting (Machado and strikeout king Tyler O’Neill were the others). Honestly, it wasn’t the strikeouts so much as the fact that Judge hit into a whopping 22 double plays.

Some of that was out of his control, because he led the majors in double play opportunities. Batting after Juan Soto will do that. But in 200 opportunities, he hit into 22 double plays. José Ramírez was second in the majors in double play opportunities – and he hit into only nine out of his 155 chances. Bryan Reynolds hit into only seven in 148 bites at the apple. Heck, Soto himself had 138 chances to hit into a double play and ended up with only 10 of them. Judge’s performance in double play situations was bad, and it happened frequently.

As in Carroll’s case, I’m not sure how much of the blame from those -8.8 runs truly belongs to Judge. A lot of it is situational, and those situations tend not to repeat from one year to the next. It’s also not a big deal in the grand scheme of things; if you insisted on counting every drop, Judge would have produced about 10.3 WAR last year. (This would have made for a much more competitive MVP race between him and Bobby Witt Jr., but even with this adjustment, nobody else would’ve been at their level.)

Still, these run adjustments are real. Different ways of making outs have always been treated mostly the same (batting average never gave you an extra out for a double play), but they’re slightly different in practice. You can’t consider Carroll’s offensive prowess without noting his ability to advance his team’s agenda, at least a little bit, even when he fails. Likewise, Judge’s season inarguably involved some painful failures that weren’t completely accounted for in his batting line.

Once again: Most of these come out in the wash. It’s likely that Luis Arraez’s -1.4 runs, Shohei Ohtani’s +2.6 runs, and Soto’s 2.4 runs are all just noise that will cancel out in the long run. But I like calculating things, and I particularly like hunting for hidden value. Avoiding double plays while still advancing runners is unquestionably valuable – and now, that value is at least a little bit more out in the open.