Winning Ugly: A Look at This Year’s Postseason Starting Pitching

Sunday night’s NLCS Game 6 offered quite a contrast in its starting pitcher matchup. With a chance to push the series to a decisive Game 7, the Mets started Sean Manaea, a 32-year-old lefty who made a full complement of 32 starts during the regular season, set a career high for innings pitched (181 2/3), and had already made three strong postseason starts, allowing five runs across 17 innings. On the other side, with an opportunity to close out the series and claim their fourth pennant in eight seasons, the Dodgers tabbed Michael Kopech, a 28-year-old righty who started 27 games last year but hadn’t done so once this year, instead pitching out of the bullpen 67 times in the regular season and four more in the playoffs. The unorthodox choice owed to the Dodgers’ injury-wracked rotation. Los Angeles has barely been able to muster three workable starters for October, let alone four, and so manager Dave Roberts has resorted to sprinkling in bullpen games, with mixed results.

The ballgame turned out to be a mismatch, but not in the way you might have imagined. Kopech struggled with his control, throwing just 12 strikes out of his 25 pitches, walking two, and allowing one hit and one run. If he set a tone for the rest of the Dodgers staff, it was that this was going to be a grind, the outcome hinging on their ability to navigate out of traffic — which they did, stranding 13 runners while yielding “only” five runs. Meanwhile Manaea, who had limited the Dodgers to two hits and two earned runs over five innings in NLCS Game 2, lasted just two-plus innings and was battered for six hits while walking two. He was charged with five runs, four of which came off the bat of Tommy Edman in the form of a two-run double in the first inning and a two-run homer in the third.

The Dodgers weren’t expecting Kopech to go any deeper, leaving Roberts to follow a script that allowed him to utilize his remaining relievers to best effect (such as it was). The Mets harbored hopes that Manaea could at least pitch into the middle innings so that manager Carlos Mendoza could avoid deploying some of their lesser relievers, but the starter faltered so early that they didn’t have that luxury. As it was, the fifth run charged to Manaea scored when Phil Maton, already carrying an 8.44 ERA this October, was summoned with no outs in the third and didn’t escape before serving up a two-run homer to Will Smith. Faced with a 6-1 deficit, the Mets refused to go quietly, but went down just the same in a 10-5 loss that included 14 pitchers combining to allow 22 hits and 12 walks. It was excruciating viewing, and with a pennant on the line, one couldn’t help but wish instead for starters battling deep into the game. Alas, this was hardly atypical October baseball.

Indeed, the collective work of starting pitchers during this postseason should come as no surprise to anyone who’s tuned in over the past decade. With regular season starting pitcher usage already on the wane due to workload concerns and an understanding of the three-times-through-the-order penalty, the average postseason start slipped below five innings in 2017, and while it has poked its head back above that line a couple of times since, those seasons have been the exceptions. This year’s postseason starters are averaging fewer innings per turn than in any season but 2021, and the drop-off from their regular season average — which was actually the highest it’s been since 2018 — to their postseason average is the second largest in that span:

Despite that postseason average of 4.25 innings per start, the starters have performed respectably, at least relative to recent standards:

This year’s starters have thrown fewer pitches than ever, and lasted at least six innings less often than in any of these seasons save for 2021. Nonetheless, their collective ERA is still the second-lowest mark of the past half-decade, and as you can see, they’ve allowed fewer runs per start — not per nine innings — than at any other point in that span. Managers are using their hooks sooner, for reasons both general and specific to their particular situations.

The Dodgers were well below that average of 4.25 innings, but their starters did tend to work deeper than another LCS participant:

The Guardians got even fewer innings per turn out of their starters than the Dodgers. All season long, manager Stephen Vogt and company struggled to find effective starters, having lost Shane Bieber to Tommy John surgery early in the season. Their rotation’s 4.40 ERA and 4.51 FIP both ranked 24th in the majors; they had the highest ERA among the postseason teams, while in FIP they were just one point better than the Brewers, who nonetheless had the better park-adjusted FIP- (110 to 112). Of the six Guardians who started at least 16 times, four had ERAs of 4.86 or higher. Gavin Williams, he of that 4.86 mark (and also a more respectable 3.67 FIP), made one postseason start, but lasted just 2 1/3 innings against the Yankees in ALCS Game 4, allowing three runs. Tanner Bibee, Cleveland’s best starter during the regular season, reached the five-inning mark only in ALCS Game 5, when he went 5 2/3 innings and allowed just two runs; in his previous turn in Game 2, Vogt pulled him in the second inning after a 27-pitch first.

The two other starters the Guardians used, Alex Cobb and Matthew Boyd, were late additions who respectively made just three and eight starts during the regular season after returning from surgeries (left hip labrum for Cobb, Tommy John for Boyd). With the majors’ most effective bullpen at his disposal, Vogt used a quick hook, with Boyd’s five-inning start in ALCS Game 3 the only other time the team reached that threshold; not once did a Cleveland starter last six innings. Within their limited footprint, Guardians starters pitched well enough to keep the team in games, but their key relievers got lit up in critical spots. Closer Emmanuel Clase blew a save in ALCS Game 3 by allowing back-to-back homers to Aaron Judge and Giancarlo Stanton, and then took the loss in Game 4 when the Yankees scratched out two ninth-inning runs. Setup man Hunter Gaddis served up what proved to be the decisive homer in Game 5, a three-run 10th-inning shot by Juan Soto.

The ongoing struggles of Clase, who turned in a 0.61 ERA and 2.23 FIP during the regular season but was torched for a 9.00 ERA and 6.42 FIP in eight postseason innings, serves as a reminder that a manager can only go to the well so often when it comes to his top relievers. Research by old friend Travis Sawchik shows that by the time a reliever faces a hitter for the third time in the same postseason, his performance has more or less degraded to the point of a third-time-through starter.

The third-time-thru-the-order penalty guides a lot of decision making regarding SPs, but there's a postseason dynamic regarding bullpen arms we should also pay attention to. The more hitters see relievers in a series, they better they hit: pic.twitter.com/tNnvJMT5Qk

— Travis Sawchik (@Travis_Sawchik) October 16, 2024

Gaddis, who faced the Yankees four times, produced an 8.10 ERA. The Yankees’ Luke Weaver, who began the postseason with six scoreless innings and four saves, three of them lasting four outs or more, was touched up by the Guardians in his second and third appearances. Clay Holmes, the pitcher whom Weaver replaced as the Yankees closer in September, reeled off 6 2/3 scoreless innings in four ALDS appearances against the Royals and two more against the Guardians before allowing runs in his third and fourth appearances of the ALCS.

Cleveland’s ALDS opponent entered the postseason with an even thinner rotation. Before their furious late dash to the playoffs, the Tigers traded the resurgent Jack Flaherty to the Dodgers, leaving manager A.J. Hinch to get through the postseason with one true starter, likely AL Cy Young winner Tarik Skubal, and a bunch of openers and bulk guys. In fact, Detroit led the majors with 41 starts of three innings or fewer, 15 more than the second-ranked Giants and 16 more than the third-ranked Dodgers. That count does lump together openers, bullpen games, and starts where rotation regulars were pulled due to ineffectiveness, but it’s good enough to capture the prevailing trend: Such starts have decreased from a high of 12.9% in 2021 to 8.9% this year, the lowest mark since the tactic gained traction in 2018. With Skubal able to take three postseason turns, all lasting at least six innings, Tigers starters somehow averaged more innings than their Guardians counterparts did. Even so, in the Division Series matchup between the two teams, Cleveland got to opener Tyler Holton for four runs in Game 1 and touched up Skubal for five runs in the decisive Game 5.

As for the Dodgers, they were third in short starts during the regular season because they wanted to give Yoshinobu Yamamoto five days of rest, in line with his schedule in Japan. It still wasn’t enough to keep him fully healthy, as he missed nearly three months due to a rotator cuff strain (that after a brilliant June 15 start against the Yankees in the Bronx). The Dodgers have kept him on a short leash since his September 10 return. He completed five innings in just one of his four regular-season starts since then, and once in his three starts this postseason; he’s maxed out at 79 pitches across that seven-start stretch. After yielding five runs in three innings against the Padres in the Division Series opener (which the Dodgers nonetheless won), he turned in a brilliant five shutout innings in Game 5, then followed up with a 4 1/3-inning, eight-strikeout showing against the Mets in NLCS Game 4, another win.

Navigating Yamamoto’s limitations is just one of Roberts’ headaches. Prior to Kopech’s wobbly effort, the team used righty reliever Ryan Brasier to start twice, first in NLDS Game 4, when the Dodgers themselves were facing elimination, and again in NLCS Game 2, when rookie bulk man Landon Knack followed by allowing five second-inning runs en route to a 7-3 loss. The team is in this predicament because it lost three potential postseason starters to season-ending injuries during August, namely Gavin Stone and Tyler Glasnow, its only two starters to reach 100 innings this season, and Clayton Kershaw, its most decorated starter.

Flaherty, the most impactful starter dealt at the trade deadline, turned in the Dodgers’ only postseason start longer than five innings; he threw seven shutout frames against the Mets in the NLCS opener. However, he’s allowed 12 runs over a combined 8 1/3 innings in his other two turns; his three-inning, eight-run dud in NLCS Game 5 produced a game score of seven, the lowest of this postseason. Walker Buehler, their third starter, earned a reputation as a great big-game pitcher with his work in the 2018–21 postseasons, but hasn’t fully recovered his form since returning from late 2022 Tommy John surgery. He pitched to a 5.38 ERA and 5.54 FIP in 16 regular season starts while averaging just over 4 2/3 innings per turn. He was charged with six runs in five innings in NLDS Game 3, but all of those runs came in the second inning, when the Dodgers defense went to hell in a handbasket; though he didn’t strike out a single hitter, he needed just 44 pitches to cover his other four innings, and allowed the Dodgers to make a game of it (they lost 6-5). Buehler returned to flash his old, fiery self while delivering four shutout innings against the Mets in NLCS Game 3, striking out six and generating a season-high 18 swings and misses.

It hasn’t always been pretty for the Dodgers — sometimes it’s been downright ugly — but an offense that has cranked out 6.36 runs per game in October has papered over some of their issues, and a bullpen that’s produced a 3.16 ERA has covered most of the rest. This may not be the most aesthetically pleasing brand of baseball, but their plan, and the larger trend, can’t be divorced from the particularities of the participants or the quirks of the postseason schedule with its frequent off days. You may not like it, but at least in 2024, this is what pennant-winning performance looks like.