The Shortage of Reliable Pitchers Is Worse Than You Think
Pitching! Everyone’s concerned with pitching this postseason, and for good reason. Pitchers are always getting hurt. They don’t throw as many innings as they used to. Even good teams, rich teams like the Mets and Dodgers, are throwing de facto bullpen games deep in the playoffs. And leaving a starting pitcher in past his 18th hitter risks invoking the wrath of the dreaded third-time-through-the-order penalty.
Remember Tanner Bibee? He’s a really good starting pitcher; he had a 3.47 ERA in 31 starts for the Guardians this year. In Game 5 of the ALCS, two trips through the Yankees order got Bibee five scoreless innings. But when manager Stephen Vogt brought Bibee out for a sixth, it was like he’d ordered a punt on fourth-and-short from inside the opponent’s 40-yard line. And sure enough, Bibee allowed three hits to his last four opponents, the last of them a game-tying home run.
And yet, when it came time to evaluate Vogt, Ben Clemens — who was not particularly high on the rookie manager’s tactics throughout the postseason — wrote that extending Bibee was preferable to leaning on a bullpen that had already spit the bit:
“[T]he bullpen was shot. [Cade] Smith was up to eight appearances in nine games. So was [Tim] Herrin. [Emmanuel] Clase had gotten shelled two nights in a row… Bibee came out strong and took a 2-0 lead into the sixth. Vogt had used his bullpen so aggressively and frequently that he really had no choice but to let Bibee face the big boppers a third time.”
On Monday, Jay Jaffe wrote about the state of postseason starting pitching. In short: There’s less of it than normal. When they’ve stayed in the game, the starters have been decent, but there’s been an early hook.
The reason for this is at least somewhat circumstantial. Among the four teams to reach the LCS, Tyler Glasnow, Emmet Sheehan, Clayton Kershaw, Dustin May (I promise there’s a non-Dodger on here eventually), Gavin Stone, Tony Gonsolin, and Shane Bieber are all out with long-term injuries. Kodai Senga, Yoshinobu Yamamoto, Alex Cobb, and (I’d argue) Walker Buehler were all limited to some extent by previous injuries. The same goes for Nestor Cortes if he makes the Yankees’ World Series roster. The Royals, Phillies, and Astros — all teams that stockpiled starters — got knocked out early. The Mariners, Diamondbacks, and Rangers — who did the same — missed the playoffs altogether.
So it makes sense that bullpens would have to pick up more of the slack this postseason. Here, have a graph.
It’s a bit uncomfortable to look at. (In all kinds of relief, you never want to cross the streams.) At the turn of the century, starters carried just under two-thirds of the playoff workload. On a per-game basis, that shakes out to six innings for the starter, and three for the bullpen. I imagine that, like, Warren Spahn is relieved he didn’t live long enough to see the postseason starter’s role shrink as much as it has, but ask any baseball fan born after, say, 1975, and they’ll tell you a 2-to-1 starter-to-reliever workload ratio is how the game is supposed to work.
So, like a relief pitcher, I’ll pick up the torch from Jay and try to get this examination of pitcher usage over the line. And I hope I do a better job than this year’s playoff relievers have, because things are pretty grim.
Back in 2015, the Kansas City Royals figured out that a team with a pretty average offense and rotation could cakewalk to a World Series title if it had four knockout high-leverage relievers. The next year, Terry Francona discovered Andrew Miller, and ever since then it seems like playoff teams have leaned on their bullpens more and more.
In a couple instances — most notably Miller in 2016 and Houston’s Will Harris in 2019 — a relief ace carried his team to the verge of a title. But a max-effort reliever can only pitch every day for so long before he runs out of gas. Bad news: Both Miller and Harris ran out of gas in Game 7 of the World Series.
Last October, Ben Lindbergh of The Ringer wrote an article publicizing research by David Gordon and Cameron Grove; that research pointed to the existence of a repeat reliever penalty. The idea behind this theory is that, just as hitters tend to figure out a starting pitcher given multiple looks in a game, they’ll also figure out a one-inning reliever if they face him three or more times in a short series.
Barring something outlandish happening in the World Series, my enduring memory of this postseason is going to be of the best relievers in the league getting pantsed in high-leverage situations. Just one All-Star after another giving up dingers upon nukes upon tanks with the season on the line. I’m going to quantify this effect later, but I wanted to signpost it up here because, on the surface, it makes no sense.
Like, nobody hit Jeff Hoffman or Devin Williams or Emmanuel Clase this regular season, and yet all of them turned into pumpkins in the playoffs. It beggars belief. There must be some rational explanation, and a combination of fatigue and the repeat reliever penalty must be to blame.
Mike Petriello, MLB.com’s resident brain-in-a-jar, is not so certain. Mike’s a smart guy; I know this because he and I once won SABR research awards on the same day. He not only figured out that the repeat reliever effect is historically finicky, but that it’s reliant on a small sample of data. So small, in fact, that it’s not too difficult to go through each individual datum to look for confounding variables.
And he found one: High-leverage relievers don’t actually see the same hitters three times in a series very often. In the NLCS, there were 21 instances of a Mets hitter getting a third look at a Dodgers reliever; 14 of those came with Brent Honeywell on the mound. Honeywell is Dave Roberts’ mopup guy, and in a six-game series composed entirely of blowouts, he got a ton of work. More to the point, nobody gives a damn what happens when Honeywell is on the mound, because by the time he enters the game, the outcome has already been decided.
Either way, the repeat reliever penalty can’t explain… well, any, really, of this postseason’s biggest reliever meltdowns. There have been six relief appearances this postseason in which a pitcher posted a WPA of -.500 or worse. And they tended to come early in the series, not later.
Player | Team | Opponent | Round | Game | WPA | Appearance |
---|---|---|---|---|---|---|
Devin Williams | MIL | NYM | NLWC | 3 | -0.843 | Second |
Phil Maton | NYM | MIL | NLWC | 2 | -0.615 | First |
Emmanuel Clase | CLE | NYY | ALCS | 3 | -0.608 | First |
Jeff Hoffman | PHI | NYM | NLDS | 1 | -0.514 | First |
Edwin Díaz | NYM | PHI | NLDS | 2 | -0.507 | First |
Jackson Jobe | DET | HOU | ALWC | 2 | -0.500 | First |
I’m entirely open to the idea of a repeat reliever penalty in the abstract. I just don’t think that’s what’s happening here.
Mike points out in his article that overall reliever performance this postseason has been within historical bounds; the leaguewide reliever ERA, opponent SLG, and opponent OPS are all within a few points of the averages from 2016 to 2023. And on the whole, starters and relievers have been similarly effective this postseason.
Role | K-BB% | AVG | WHIP | ERA | FIP |
---|---|---|---|---|---|
Starters | 13.7% | .242 | 1.30 | 4.12 | 4.10 |
Relievers | 12.6% | .218 | 1.30 | 3.73 | 4.34 |
That’s worth remembering, but I’m interested in a slightly different question. You might think relievers ought to be way out in front of starters, but again, not all relief situations are created equal, and nobody cares how many runs Brent Honeywell gives up.
On the other hand, they care deeply how many runs the high-leverage guys give up.
We’ve still got one round to go, so this could still change. But as it stands, 2024 is the worst postseason for relievers, by WPA, since 1986. What’s so special about that postseason? Well, in the NLCS, the Astros’ bullpen got tagged for two walk-off losses and ended the season by getting beat in 16 innings in Game 6. Nobody remembers this because it came in the same season as two of the most famous bed-encrappenings by any bullpen in the history of baseball: Dave Henderson’s three-run home run off Donnie Moore in Game 5 of the ALCS and Boston’s five-alarm meltdown in Game 6 of the World Series, which was likely lost irrespective of Bill Buckner’s lackluster hip mobility.
So yeah, that’s how bad things are right now.
As the importance of elite relief performance — and elite relief depth — became important in the mid-2010s, teams have rejected, or at least updated, the doctrine that reliever quality is volatile and that chasing it is a waste of resources. You might be able to get away with throwing some unpolished fastball-slider guy in a high-leverage situation, but it’s much better to have some studs back there. Remember how dominant the Astros bullpen was in 2022, and how easy their World Series title looked as a result?
Here, we return to the idea of the best relievers in baseball getting completely torched. I thought this was just a case of me remembering the outliers. I was at the first two games of the Mets-Phillies NLDS, which included three blown saves by relievers who have made an All-Star team in the past three seasons. Plus the worst week of Clase’s life, plus the Pete Alonso dinger off Williams that flipped the Wild Card series.
Turns out, that’s a pretty common occurrence this postseason. Here are the top 20 relievers who appeared in at least one game this postseason, sorted by regular season WAR. The names highlighted in gold appeared in at least two-thirds of their teams’ games. The cells highlighted in red are postseason WPA marks of -0.30 or worse. Half of these elite relievers posted a negative WPA this postseason, and 14 of them have a K-BB% ratio worse than what they posted in the regular season.
Regular Season | Postseason | |||||||||
---|---|---|---|---|---|---|---|---|---|---|
Name | Team | WAR | WPA | K-BB% | ERA | G | IP | WPA | K-BB% | ERA |
Cade Smith | CLE | 2.7 | 3.04 | 29.8% | 1.91 | 9 | 10 | 0.13 | 37.8% | 3.60 |
Emmanuel Clase | CLE | 2.3 | 6.40 | 20.7% | 0.61 | 7 | 8 | -0.85 | 18.4% | 9.00 |
Matt Strahm | PHI | 2.1 | 1.20 | 28.7% | 1.87 | 3 | 2 | -0.44 | 10.0% | 18.00 |
Jeff Hoffman | PHI | 2.1 | 2.47 | 27.5% | 2.17 | 3 | 1 1/3 | -0.52 | -10.0% | 40.50 |
Joe Jiménez | ATL | 1.8 | 1.49 | 21.3% | 2.62 | 1 | 1 1/3 | 0.03 | 0.0% | 0.00 |
Lucas Erceg | KCR | 1.7 | 1.16 | 22.1% | 3.36 | 5 | 6 | 0.58 | 12.5% | 3.00 |
Tanner Scott | SDP | 1.6 | 3.27 | 16.3% | 1.75 | 5 | 4 1/3 | 0.18 | 29.4% | 0.00 |
Jeremiah Estrada | SDP | 1.6 | 0.38 | 28.2% | 2.95 | 4 | 4 1/3 | -0.19 | 21.1% | 0.00 |
Hunter Gaddis | CLE | 1.6 | 2.44 | 18.7% | 1.57 | 8 | 7 1/3 | -0.51 | 19.4% | 4.91 |
Orion Kerkering | PHI | 1.5 | 0.96 | 23.2% | 2.36 | 4 | 3 2/3 | -0.22 | 17.6% | 2.45 |
Keegan Akin | BAL | 1.4 | -0.35 | 24.9% | 3.32 | 1 | 2/3 | 0.03 | -33.3% | 0.00 |
Jason Adam | SDP | 1.3 | 2.62 | 20.6% | 1.95 | 5 | 5 1/3 | 0.12 | 30.0% | 3.38 |
Clay Holmes | NYY | 1.2 | -1.55 | 17.0% | 3.14 | 8 | 7 2/3 | -0.10 | 9.4% | 3.52 |
Carlos Estévez | PHI | 1.2 | 2.21 | 17.9% | 2.45 | 3 | 2 2/3 | -0.31 | 25.0% | 3.38 |
Mark Leiter Jr. | NYY | 1.1 | 0.58 | 25.0% | 4.50 | 2 | 2 1/3 | 0.02 | -8.3% | 3.86 |
Adrian Morejon | SDP | 1.1 | 1.00 | 18.4% | 2.83 | 2 | 2 1/3 | -0.08 | -10.0% | 3.86 |
Will Vest | DET | 1.1 | 0.23 | 15.6% | 2.82 | 6 | 6 | 0.23 | 40.9% | 1.50 |
John Schreiber | KCR | 1.1 | 1.54 | 11.6% | 3.66 | 5 | 3 2/3 | 0.27 | 0.0% | 0.00 |
Aaron Bummer | ATL | 1.1 | -0.44 | 20.9% | 3.58 | 1 | 2 2/3 | 0.07 | 0.0% | 0.00 |
Edwin Díaz | NYM | 1.1 | -0.07 | 29.6% | 3.52 | 6 | 8 2/3 | -0.33 | 16.2% | 3.12 |
Also, notice who’s not on this list: Williams, who was hurt half the year but would be near the top of any list of the best closers in the game. Or Josh Hader and Ryan Pressly, who conspired to allow four runs in the eighth inning of Game 1 of the AL Wild Card series, putting Detroit on the path to an upset win.
Overall, relievers have been fine. Except when it counts most.
This postseason, there have been 47 relief appearances with an average leverage index (aLI) of 1.75 or higher. And generally, the guys who have pitched in those situations have been the players you’d want there; 38 of those 47 appearances have been executed by a pitcher who produced at least 1.0 WAR this season, had an ERA under 3.00 or both.
In those high-leverage situations, the best pitchers in the league have been absolutely horrendous. Here’s how those relievers compared to relief pitchers in comparable circumstances in two previous seasons: 1986, the nadir of playoff relief pitching, and 2015, the last season in which relievers accounted for less than 40% of the postseason innings total.
Year | G | IP | WPA | aLI | ERA | Opp. OBP | K% | BB% |
---|---|---|---|---|---|---|---|---|
1986 | 17 | 29 | -2.61 | 3.26 | 4.66 | .348 | 17.4 | 12.1 |
2015 | 45 | 47 | 0.117 | 2.48 | 3.83 | .346 | 24.5 | 10.6 |
2024 | 47 | 41 2/3 | -2.03 | 2.61 | 7.13 | .440 | 21.6 | 16.1 |
The best teams in the league are built to get the ball to world-class relief pitchers when the game is on the line. And time and time again, those elite relievers are going to pieces. Except for the Dodgers and Yankees, which I guess is why those two teams are in the World Series.
Even if it’s not directly attributable to fatigue or the repeat reliever effect, I think we’ve reached yet another historical inflection point. The paucity of healthy and effective starting pitchers had led to a massive overreliance on relief pitchers; turns out we’re running low on those too. A decade ago, the Royals won a World Series because they built the entire bullpen out of unhittable relievers. Now there’s no such thing.
Michael is a writer at FanGraphs. Previously, he was a staff writer at The Ringer and D1Baseball, and his work has appeared at Grantland, Baseball Prospectus, The Atlantic, ESPN.com, and various ill-remembered Phillies blogs. Follow him on Twitter, if you must, @MichaelBaumann.
Relievers are fickle and the playoffs are a crapshoot. But hey, atleast we get Yanks vs Dodgers.