clock menu more-arrow no yes

Filed under:

Clayton Kershaw, pulled early

Some say Kershaw is left in too long during his postseason starts. But not in Game 1.

2020 World Series Game 1: Los Angeles Dodgers v. Tampa Bay Rays Photo by Cooper Neill/MLB Photos via Getty Images

In Game 1 of the World Series, after a rough first inning that saw two of the first three Rays hitters reach on a walk and a single, Clayton Kershaw settled in. A strikeout and a groundout capped off Kerhsaw’s first inning, which saw him throw 20 pitches, all while looking relatively solvable. Kershaw only generated one swing-and-miss in the entire inning.

From the second on, however, Kershaw looked much more dominant. The slider command returned, and outside of one hanging breaking ball to Kevin Kiermaier, Kershaw completely quieted the Rays. Over his final five innings, four of them were one-two-three; in fact, Kiermaier represented the only baserunner Kershaw allowed following his rocky start.

Kershaw finished the sixth inning having thrown just 78 pitches, and his line was sterling: just two hits and one run allowed, with eight strikeouts and one walk. As the Dodgers added a bunch of runs in the middle innings — two in the fourth, four in the fifth, and two more in the sixth — the age-old question arose again: Keep a seemingly-cruising Kershaw in the game to start the seventh, or call it a day for the 32-year-old lefty?

Unlike in past postseason starts, when it felt like Kershaw had always been left in just slightly too long, Dave Roberts took no chances on Tuesday. Though he had thrown only 78 pitches and was likely good for one or two more innings, Kershaw’s night was done, and the dreaded “postseason narrative” was nowhere to be found.

Notably, Kershaw is having what is shaping up to be his most successful postseason. He has now made four starts, one in each of the Dodgers’ four series thus far, and the numbers look extraordinarily close to regular season Kershaw, if not better: a 2.88 ERA, 33.3 percent strikeout rate and a 3.2 percent walk rate all suggest dominance. The 3.15 FIP can again be attributed to some poor homer luck, as the 2.77 xFIP and 2.43 SIERA would argue. No matter how you analyze it, Kershaw has been excellent in his 25 innings, even including one relatively rough start in Game 4 of the NLCS.

Though the causation certainly cannot be established here, some were quick to point out that, if things hold, 2020 would mark the first postseason in which Kershaw did not start on short rest or get used out of the bullpen. Jon Weisman, likewise, updated his Kershaw Postseason Start Chart, and showed that, including his performance in Game 1, the lefty has now made 12 playoff starts with 1 or fewer earned runs allowed through his first six innings. If you bump the threshold up to 2 runs, it becomes 17 playoff starts.

The key here, though, is “the first six innings.” A big critique of the Dodgers’ postseason Kershaw usage is that, too often, he is kept in games for too long. Is there a way to measure this? Of course, there are a ton of factors that go into “too long,” including those related to the short rest and the bullpen usage. “Too long” for a Kershaw coming in on short rest certainly looks a lot different than “too long” for a Kershaw on regular rest, for example.

I dove into this further, focusing mainly on pitch count. Is there a difference in the distribution of Kershaw’s typical pitch count in the regular season versus the postseason? The answer is pretty clearly yes, but perhaps not in the way that you’d expect. (Note that this figure only shows Kershaw’s starts.)

Indeed, Kershaw throws more pitches in the regular season than he does in the postseason. He averages 99.4 pitches per start during his regular season outings to just 93.5 pitches per start in the postseason. That’s demonstrated in the different peaks of those distributions.

But there’s something funky going on in the left tail. The fewest number of pitches Kershaw has thrown in any postseason start is 72. Let’s compare this to the regular season, where Kershaw has 16 different starts where he’s thrown fewer than 72 pitches. Some of those could’ve been due to injury, some due to poor performance, but if Kershaw had maintained his rate of less-than-72 pitch starts from the regular season to the postseason, we should’ve seen at least one by now and could conceivably have two. The super short Kershaw starts, while rare in the regular season, just don’t exist in the postseason.

Why does this matter? Consider this: During the regular season, there’s a weak-but-notable correlation between Kershaw pitches and Kershaw game score. In other words, in games where he throws more pitches, he’s generally more effective:

The correlation isn’t too strong — with an r-value of just 0.37 — but it exists. In games where Kershaw goes longer, he’s generally better. And that makes sense: The Dodgers are only going to keep him in the game if he’s pitching well.

In the postseason, however, no such correlation exists. In fact, there’s even a slight negative correlation, with an r-value of -0.02:

Kershaw has had great starts where he’s thrown few pitches — like last night — and he’s also had poor starts where he’s thrown a ton of pitches, like in Game 6 of the 2013 NLCS, when Kershaw allowed seven runs in just four innings against the Cardinals. It’s hard to know if the Dodgers should’ve pulled him sooner, but sending him back out for the fifth — after already allowing four runs in the third — seemed disastrous. Single, single, double, and Kershaw was out of the game, having allowed a fifth run and leaving two more runners on base, both of whom eventually scored. Kershaw threw that game on regular rest, but it was his fourth postseason start in just 15 days that year. (In the first three, he posted a 0.47 ERA.)

I want to hedge this by saying that firm conclusions can’t be drawn here: This doesn’t prove anything. All of this is to point to a few things: 1) Kershaw almost always throws at least 75 pitches in the postseason, while not always doing so in the regular season, 2) he is oftentimes doing this on short rest and/or after having come out of relief in a prior outing and 3) there no longer exists a relationship between pitches thrown and performance. Kershaw-is-a-little-bad outings can get much worse if Kershaw is not having his best day and the Dodgers push him further, like in the case of the 2013 NLCS Game 6.

Statistically, there may be some legs to the thinking that the Dodgers generally overuse Kershaw in the postseason. This doesn’t necessarily prove that, but it gives us something to consider about the relationship between pitch count and performance. Usually, pitchers only go deep when pitching well. For Kershaw in the postseason, the link no longer seems to be there, and that might make a bad day worse.

On Tuesday, though, Kershaw pitched well, and threw little. The Dodgers lead 1-0 in the World Series, and if they’re going to come home with the title, they’re going to want a well-rested Kershaw to carry the freight on the mound.


Devan Fink is a sophomore at Dartmouth College and a Contributor at Beyond The Box Score. Previous work of his can be found at FanGraphs and his own personal blog, Cover Those Bases. You can follow him on Twitter @DevanFink.