Less Is More: Relief Pitching Has Dominated This Postseason

Though the short and ugly starts of Bailey Falter and Mike Clevinger in Game 4 of the NLCS put a small dent in the numbers, thus far in the postseason we’ve seen a welcome rebound when it comes to starters preventing runs and pitching deeper into games relative to recent years. After two postseasons in which relievers threw more than half of all innings, starters have reclaimed their rightful spot atop the marquee. Even so, when it comes to October, the relievers we’ve seen have been stingier and more dominant than at any time in recent years. Thankfully, we’re seeing fewer of them — meaning fewer pitching changes and less dead time — but that’s of cold comfort to both the Mariners and Yankees.

As you can probably guess, that’s because Houston’s relief corps has been so overwhelming thus far. The Astros are 7-0 in this postseason, and while their starting pitching has been a major part of the story, their relievers have held opponents to just three runs in 33 innings (a 0.82 ERA) and a .127/.207/.227 line, with a 34.7% strikeout rate. As old friend Mike Petriello wrote recently, the staff — not just the bullpen — doesn’t have a weak link, even without having a late-inning lefty.

The Phillies’ bullpen (3.19 ERA, 30.3% strikeout rate) has performed well, too, and in fact Seranthony Domínguez has been the single most dominant reliever in the postseason according to FIP. In fact, he’s one of three relievers with at least five innings — we are indeed in small sample theater here — and a negative FIP, and one of four with a strikeout rate of at least 50% (you can sort the table if you like):

There’s another reminder that Hader, not Suarez, should have been in NLCS Game 5 to face Bryce Harper, but that’s water under the bridge now.

Truth be told, for the purposes of this analysis I’m less interested in the work of either World Series team’s relievers than I am the bigger picture, as the flip side of my annual tracking of regular season and postseason starter usage. As you’d expect, what we’re seeing is the amplification of trends that we observed in a regular season where per-game scoring fell 5.5% (from 4.53 runs per game to 4.28) and the league-wide OPS decreased for the third straight season (from .758 in 2019 to .740 in ’20, .728 in ’21, and .706 this year). The lower scoring environment and the new roster rules have led to a welcome cutback in reliever usage.

Due to workload concerns and awareness of the times through the order penalty, teams have generally been asking less of their starters than they did a decade or two ago, and that’s in part because relievers are generally more effective than starters, thanks in part to their ability to exert maximum effort in their short stints. By ERA and FIP, the reliever advantage during the 2022 regular season was the largest it’s been in half a decade:

The improvement wasn’t dramatic, and I’m not sure why the advantage was larger than in recent years, but that doesn’t mean I don’t have a potential explanation to offer. After two seasons of decreased starter workloads due to the pandemic — in 2020 due to the lack of a full-length preseason after the layoff and in ’21 due to year-over-year buildup concerns — a combination of the increase in starter innings (from 5.02 per start last year to 5.21 this year) and the shortening of staffs to 13 meant fewer innings for the chaff, so to speak. The roster rule adopted for 2020 but delayed first by pandemic conditions, and then because of the lockout, restricts teams to carrying 13 pitchers on a 26-man roster; it was finally implemented on June 20. A couple of additional rules changes, namely the return of 15-day stints for the injured list (up from 10) for pitchers and a limit on the number of times a player can be optioned to the minors in a given season (five) helped to reduce the roster churn. That means that innings that previously had gone to pitchers lower in the pecking order were reabsorbed by better relievers, as each organization’s ninth-, 10th-, or eleventy-third-best reliever summered in Sacramento, Scranton, Syracuse and so forth; the remaining pitchers were better at run prevention. I’m open to other interpretations, but that one certainly seems plausible.

In the postseason, the reliever advantage generally grows as fifth starters get cut from rotations and the bullpen’s innings are more concentrated among better pitchers. This year’s reliever advantage has been more pronounced, not only higher than in the regular season but also in either of the last two postseasons:

For the eight-season period, relievers have a 0.21 runs per nine advantage in terms of ERA during the regular season and a 0.34 runs per nine advantage during the postseason, but this year the postseason gap is about two-and-a-half times that. Generally, the gap is because starters give up more homers while striking out fewer hitters than relievers, and that’s true this time around (1.32 to 0.95 per nine for homers, 24.2% to 29.6% for strikeouts).

While relievers in general pitch more effectively than starters, this year’s combination of the latter’s increased footprint, the lower-scoring environment, and the new roster rules led to the first year-over-year reduction in reliever usage per game since 2012-13, when scoring fell by about 3.5%. That’s obviously not that long ago, but it’s further back than the window of time I’ve been tracking for this ongoing project (since 2015, a cutoff originally chosen with Statcast in mind). Managers are still using more relievers in October than they did during the rest of the season, but they haven’t needed as many as last year due to the lower offensive level:

In the postseason, the number of relievers used per game has dropped substantially relative to last year, and in fact has fallen to its lowest level since 2017:

That means fewer pitching changes, though those relievers are still occupying a greater share of innings than in the regular season. In the 2021 postseason, relievers threw an additional inning-and-change per game relative to the regular season (4.86 vs. 3.75), but this year, that gap is down to about an additional half-inning (4.18 vs. 3.66). The good news is that shift tipped the balance to the point where starters have returned to throwing more than half of all postseason innings for the first time since 2019 (though it was close in ’18). Here’s how the percentage trend looks from the standpoint of the relievers:

Again, I think that during the regular season, the new restrictions on bullpen size have something to do with what we’ve seen. Most teams’ best relievers are generally better than their tiring starters going through the order a third time, but that doesn’t necessarily hold true when the alternative to those starters is the organization’s 10th-best reliever who just flew in from Columbus (boy, are his arms tired) and is pitching tonight because the manager needed five innings from his bullpen in a tight game the day before. In the postseason, with rotations shortened and extra days of rest built in (some of which fell by the wayside this year due to the lockout-compressed schedule), managers are better able to keep the innings in the hands of their best relievers.

One thing that has historically driven up the reliever count is the number of pitchers facing just one or two batters. The three-batter rule, which has now been in effect for three seasons, has helped to curb that, though managers can still pull pitchers who finish an inning even if they haven’t met the minimum. Both before the rule’s implementation and since, short stints have made up a greater share of postseason appearances than regular season ones — though this year, the gap has been much smaller:

As you can see from the table, there’s another piece of good news to report: Game times in both the regular season and postseason have gone down. I don’t mean to lay it all at the feet of the three-batter minimum, as the other factors discussed here — mainly the reductions in offense and in the number of relievers used for stints of any length — are probably having bigger impacts, as is the use of the PitchCom system; we’ve seen times where crowd noise has forced teams to ditch the newfangled transmissions, but the manufacturer has come up with some workarounds that have gained acceptance. Seven minutes shaved off regular season games, making them the shortest since 2018, doesn’t feel like much, but 17 minutes off postseason games, making them the shortest since ’15, seems like a bigger deal. Note that all of these times are adjusted to a per-nine inning basis, so that 18-inning Astros-Mariners marathon isn’t skewing the numbers, nor are the seven-inning doubleheader games (barf) of 2020.

The one thing about the trend towards shorter postseason games that gives me pause when it comes to celebrating is the extent to which higher reliever strikeout rates may be a factor. Relievers are striking out 29.6% of all hitters this October, up nearly four points off last year’s postseason rate (25.8%) and six points above this year’s regular season rate (23.6%):

Strikeouts are exciting, but too many of them mean too few balls in play and some fairly monochromatic late innings where hitter after hitter is blown away or goes down flailing. The Astros’ bullpen doesn’t even have the highest strikeout rate from among the teams that reached the Division Series:

Having said that, reliever dominance hasn’t prevented a surprising number of unlikely comebacks. Per Baseball Reference, this year five teams have had a win expectancy below 10% at some point during a playoff game and rallied to win, compared to just one last year, three in 2020, one in ’19, and two in ’18. The last time there were four in any postseason was 2014, and the last time there were five such games was 1986, one of the most memorable Octobers in baseball history. Here’s the table for this year:

As we look to the World Series starting Friday, here’s hoping we get at least one game to add to the list, just to shake things up.