One can state that Clayton Kershaw is the most dominant pitcher of his era without receiving much pushback. He’s won three Cy Young awards (and that number could easily be five); he’s led the league in strikeouts three times; in ERA four times; and he’s never posted an ERA higher than 2.91 in a full season. He was said to be the next Sandy Koufax and he has somehow, some way, surpassed that albatross of an expectation.

But Clayton Kershaw has a problem―he hasn’t pitched like Clayton Kershaw in the postseason.

The playoffs are where heroes are born. October is a forge, where your mettle is pushed to the brink, and those who survive the heat come out the other side more solid and with a cemented legacy. Kershaw—for all his regular-season glory—has not survived October, and his legacy has suffered accordingly. The question is, why? Some will tell you he lacks the stones, the stomach, the #rig, the heart, whatever other body part you’d like to summon as a euphemism for fortitude.

They are perhaps right. We cannot know what is in Kershaw’s heart of hearts; it’s possible that it is a flower that wilts under the October moon. Nevermind how he’s survived the pressure of justifying his first-round selection, his lofty (read: insane) comps, his ridiculous contract. And nevermind him repeatedly performing better than anyone else at the highest level. Perhaps this flower ensconced within Kershaw blooms in April and dies in October and that’s all there is to be said. Or perhaps, just this once, we can shave with Occam’s Razor.

Kershaw often gets his October surprises compared to the near-flawless execution of his division rival, Madison Bumgarner. Bumgarner is the gold standard in playoff performance, and it’s imperative that what follows not be seen as any sort of attempt to take that away from him. He is the contemporary bar to which others aspire, and rightfully so. The question we are left asking, though, is a pertinent one: “What if postseason Kershaw had the benefit of postseason Bumgarner’s bullpen*?”

*Consider this less Bumgarner’s specific bullpen, but rather a bullpen his manager trusts enough to employ effectively.

Here’s how we’re going to look at this. We’ll go through every playoff start and determine when Kershaw might reasonably have been removed from the game with a trustworthy bullpen. We’ll also give as much benefit of the doubt to actual history as possible, meaning that while 2016 Terry Francona might pull his best starter aggressively, 2014 Don Mattingly—even with a decent ‘pen—wouldn’t have. We’ll also assume that any runners left on base when we pull Kershaw will score, assuming they scored according to our current reality.


2008 Playoffs

Kershaw wasn’t quite Kershaw yet, and he only appeared in two games, both as a reliever. The record stands.

Difference from reality: None

2009 Playoffs

Only one game to really investigate here. Kershaw pitched into the seventh in the divisional series against St. Louis without blowing up (I swear, it’s true). Despite nine hits, he allowed only two runs. Game 1 of the NLCS was a different story, though, as Philly tagged him for more than a run per inning. The problem was a troublesome fifth inning:

The walks were Kershaw’s third, fourth, and fifth of the game, and I’m not sure I fault the manager for leaving him in against Ryan Howard, though two wild pitches in the same inning might have been a clue he was flagging. Still, the left-handed options in the bullpen were George Sherrill (yikes) and Hong-Chih Kuo. If this game happens today, with a more open-minded manager, we might see them turn to Kuo in this situation, but it’s still a bit of a stretch. Kershaw earned his runs here, and he’ll keep ’em all.

Difference from reality: None

2013 Playoffs

We can skip to NLCS Game 6, but I’d caution everyone who believes that Kershaw is somehow a lesser pitcher in October to at least look at his first three starts here, including a clinching Game 4 on three days’ rest against Atlanta. But let’s go to Game 6, the foundational block of the “Kershaw Isn’t Clutch” movement. We should note that he doesn’t fall apart in the seventh this time, but it’s that fifth inning he lacks harmony with.

Kershaw entered the fifth inning having thrown 92 pitches. That’s a lot of pitches in four innings, and while the Dodgers’ bullpen was bad enough that Carlos Marmol was on the roster, the score at the time was 4-0 in favor of St. Louis, due to a big third inning. With the Cardinals up 3-2 in the series, it’s not a stretch to think this is an all-hands-on-deck situation. As stated up top, the premise here is to give as much benefit as we can to the manager at the time. An aggressive manager would start this inning with someone else—being charitable, let’s say we give Kershaw the inning to start, and if anyone gets on, we pull him. How does that change things? Here’s Playoffs Datum, as Terry Pratchett and Stephen Baxter might call it:

In our alternate universe, Kershaw is pulled after Yadier Molina singles, and Mattingly presumably goes to Ronald Belisario for David Freese. If Mattingly/Kershaw have a different/better bullpen to work with in this situation, there’s little doubt that he doesn’t stay in the game to give up an extra single and double. This means Kershaw allowed five earned runs rather than seven. None of this changes the fact that it’s a bad outing.

Difference from reality: -1 1B, -1 2B, -2 ER, Same IP

Cumulative DFR: -1 1B, -1 2B, -2 ER, Same IP

2014 Playoffs

Now we’re getting to the meat of the knocks against Kershaw. He threw six innings of two-run ball in Game 1 of the 2014 NLDS, but paired it with two-thirds of an inning of six-run ball. Yikes. His postseason ERA went from 3.13 following his 2013 finish to 10.80 after this game. But what exactly happened in that seventh inning? Kershaw threw 110 pitches in the game, and entered the seventh with 81 pitches under his belt. A reasonable total, despite the 96-plus degree heat and the pressure (remember that Jon Lester just removed himself from Game 5 of the World Series due to the intensity of each pitch).

Even with something better than the Dodgers’ bullpen—which includes Kenley Jansen and … well, the necrotic husk of Brian Wilson, Pedro Baez, Brandon League, Jamey Wright, J.P. Howell, and people less notable than the names just mentioned—we’re all rolling with Kershaw.

To quote a movie that’s aged as well as Brian Wilson … that escalated quickly. If we’re the Datum Dodgers, it’s reasonable to stick with Kershaw as things played out. Three straight singles is bad, but you’re leaving him in for the lefty Matt Adams, who strokes a single. It’s still 6-3, and you have a tired Kershaw versus the Cardinals’ 7-8-9 hitters. The earliest you pull Kershaw is probably against Matt Carpenter.

In our stepwise (borrowing from Pratchett and Baxter, again) world though, where Kershaw has a better bullpen (or an actively good one, if you want to dream), it’d be reasonable to pull him after the Adams single. Sure, the score is still in our favor, but any reliever worth his salt should be fine against the bottom of the order, and Kershaw would be over 90 pitches on a brutally hot day at this point. If we make the change there, but assume all inherited runners score (as they did against Datum Kershaw), we remove a bases-loaded single and double from Kershaw’s ledger, but Jhonny Peralta, Molina, and Adams still come in and count against him.

Another start on three days’ rest sets the stage for Game 4. Kershaw again goes over 100 pitches despite the short rest, and again runs into trouble in the seventh inning, when pushed past his limits.

By any measure, Kershaw was cruising in this game, ringing up nine batters and allowing no runs entering that fateful seventh. Still, with the limited rest and a pitch count of 94 entering the inning, any manager with two good options at the back of the bullpen (or one that they thought they could push hard enough) would be readying those options in case of the slightest provocation. In real life, we saw two singles and an Adams homer. In an alternate universe, the hyper-aggressive manager with present-day Andrew Miller might have brought him in after one single, but being charitable, let’s say Kershaw is pulled before giving up the homer to Adams. He’s at 100 pitches and clearly burnt. This removes one home run from his tally, and one earned run.

Difference from reality: -â…” IP, -1 1B, -1 2B, -1 HR, -3 ER

Cumulative DFR: -â…” IP, -2 1B, -2 2B, -1 HR, -5 ER

2015 Playoffs

Game 1 against the Mets saw the seventh inning rear its ugly head again, but it’s hard to see even a moderately aggressive approach to using the bullpen being employed in this instance, as Kershaw was doing well until walking the bases loaded. He wasn’t allowed to face a batter with the bases juiced, and his last batter faced was a lefty. He did end the game with 113 pitches, but entered the seventh inning under 90. There’s little fault to be had with the tact employed by Mattingly in this game.

Game 4, on three days’ rest once more, was an absolutely dominant effort. Nothing to investigate here.

Difference from reality: None

Cumulative DFR: -â…” IP, -2 1B, -2 2B, -1 HR, -5 ER

2016 Playoffs

Game 1 Kershaw wasn’t sharp, but his runs allowed came in the third and fourth innings, and it’s hard to see any manager pulling the best pitcher in baseball that early in the first game of a playoff series. Game 4 brought on the latest bout of accusations regarding Kershaw’s mental fortitude in October or more specifically in the seventh inning. The Nationals touched him for two runs early on, but he was in a groove heading into the seventh (sound familiar?).

On three days; rest once again, Kershaw threw 110 pitches. He entered the seventh at 89 pitches, having thrown 101 pitches over five innings in Game 1. There’s not an obvious point at which you might remove Kershaw once the inning starts. It’s the bottom of the Nationals’ lineup and he was doing well enough. But! In the top of the seventh inning, manager Dave Roberts allowed Kershaw to bat for himself, with 89 pitches banked. If the Dodgers had a bullpen with more than two reliable arms*, it’s hard to escape the idea that Kershaw, batting with one out and Chase Utley and Corey Seager behind him, wouldn’t have been replaced for a pinch-hitter. If we opt for that course of action, we remove two-thirds of an inning, two singles, a walk, and three runs from Kershaw’s ledger.

*It’s simple to argue that this bullpen did have more than two reliable arms in it, with Julio Urias having yet to be used at this point, but Roberts’ bullpen management throughout the series made it nigh on imperative that Urias be available for Game 5.

It’s tempting (it’s not) to go through to Game 6 of the NLCS and find a point that might have made sense to pull Kershaw (it was after the third), and while yes, some damage was racked up after that point, it’s hard to buy into the notion that even 2016 Francona brings in his best reliever then (and one can point to Game 6 of the World Series as evidence).

Difference from reality: -â…” IP, -2 1B, -3 ER, -1 BB

Cumulative DFR: -1 â…“ IP, -4 1B, -2 2B, -8 ER, -1 HR, -1 BB


Datum Kershaw

Stepwise Kershaw












































I’m a little surprised that Kershaw still has an ERA in the upper threes. If we looked at ERA alone, it might not be unreasonable to view Kershaw as a choke artist or someone not mentally strong enough to handle the playoff pressure. Even with beneficial (though hopefully not heavy-handed) adjustments, this is not the same guy as during the regular season. Still, look at the component stats. Look at his slash line, even before the adjustments. A .650 OPS allowed means his cumulative playoff opponents hit slightly better than 2016 Jose Iglesias or nearly exactly 2015 Jace Peterson. Stepwise Kershaw is closer to 2016 Jason Heyward.

Does 3.74 ERA Kershaw avoid the label of playoff choker that 4.55 ERA Kershaw has been saddled with? I suspect not. It’s a negative deviation in performance on the bigger stage, under a bigger magnifying glass, and those moments will always draw an imbalanced share of praise or criticism. What we can say is that the issues Kershaw has experienced in the playoffs aren’t to be dropped solely at the feet of his managers or GMs.

Drastically better bullpens would certainly have helped keep the Kershaw veneer around, but drastically better bullpens would help nearly every playoff pitcher, and can’t be used an excuse in this situation. While there has been a slow hook in nearly every Kershaw playoff start, it’s hard to imagine even the less conventional baseball managers being willing to pull him significantly sooner than he was pulled in the above instances.

So, is this proof of a star that dims on the clearest nights? Call me stubborn, but I don’t think it has to be. Kershaw’s regular-season OPS allowed is .574. Considering the better competition in the playoffs, a .650 OPS allowed is a considerable increase, but not unexplained. Additionally, we (as people) are so designed to find significance and patterns in spaces where randomness is the likely culprit, that the likelihood to me (given our bigger sample that includes the regular season) is that this is variance rather than a character flaw or something predictive.

Thank you for reading

This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.

Subscribe now
You need to be logged in to comment. Login or Subscribe
This was interesting, Craig. Thanks for writing it.