:format(webp)/cdn.vox-cdn.com/uploads/chorus_image/image/70026090/1332601232.0.jpg)
In seven games of the 1985 World Series, Royals starting pitchers tossed three complete games and accounted for 93 percent of the innings for Kansas City. By the time the 2015 Royals were in the Fall Classic, only one starting pitcher tossed a complete game (thanks Johnny Cueto!) and starters accounted for just 60 percent of the innings, with Ned Yost relying on his intimidating bullpen to finish off the Mets.
Even that kind of performance from a starting pitching staff would be unheard of this post-season. For the four teams remaining, only one has had a pitcher go as many as six innings in a start - a 6 2⁄3 inning start by Houston’s Lance McCullers, Jr. At Fangraphs, Jay Jaffe notes that while shorter starts for pitchers has been a trend in baseball the last few post-seasons, it has become particularly prevalent this year.
But what we’ve seen lately are things we haven’t seen before, namely the length of the average start dipping below five innings, and starters’ share of innings collectively dipping below 56% (i.e., five innings per nine).
Some of that is due to the COVID-19-shortened season, with managers wanting to protect their pitchers from ramping up their workloads. But this has been part of a long-term trend well before the pandemic.
You can see some of the regular season starting pitcher usage trends over time here. MLB starting pitchers once averaged 6.33 innings-per-start in 1980, but have averaged less than 6 innings in every season since 2012. The percentage of starts where a pitcher lasted at least six innings has fallen dramatically, from 61 percent as recently as 2014, to just 36 percent this season. And complete games have almost become completely extinct. In 1980, pitchers went all nine in 17 percent of all starts, but less than 1 percent of starts were completed in every season since 2018. Oakland’s Ray Langford had 28 complete games in 1980. There were 31 complete games in all of baseball this season.
Starting pitcher usage over time
| Season | IP/GS | K/9 | BB/9 | ERA | Pct of 6+ IP GS | Pct of 9 IP CG |
|---|---|---|---|---|---|---|
| Season | IP/GS | K/9 | BB/9 | ERA | Pct of 6+ IP GS | Pct of 9 IP CG |
| 1980 | 6.33 | 4.58 | 2.94 | 3.97 | 65.7% | 17.5% |
| 1991 | 6.08 | 5.56 | 3.17 | 4.02 | 63.6% | 7.2% |
| 2001 | 5.92 | 6.37 | 3.10 | 4.57 | 62.1% | 3.2% |
| 2011 | 6.03 | 6.75 | 2.85 | 4.06 | 65.7% | 2.7% |
| 2012 | 5.89 | 7.14 | 2.84 | 4.19 | 63.5% | 2.3% |
| 2013 | 5.90 | 7.19 | 2.83 | 4.01 | 61.8% | 2.0% |
| 2014 | 5.97 | 7.36 | 2.69 | 3.82 | 61.3% | 1.9% |
| 2015 | 5.81 | 7.40 | 2.72 | 4.10 | 58.8% | 1.6% |
| 2016 | 5.65 | 7.75 | 2.96 | 4.34 | 54.3% | 1.4% |
| 2017 | 5.51 | 7.96 | 3.13 | 4.49 | 50.2% | 1.0% |
| 2018 | 5.36 | 8.25 | 3.03 | 4.19 | 46.8% | 0.7% |
| 2019 | 5.18 | 8.58 | 2.96 | 4.54 | 42.9% | 0.8% |
| 2020 | 4.78 | 8.78 | 3.17 | 4.46 | 31.8% | 0.8% |
| 2021 | 5.02 | 8.62 | 2.99 | 4.34 | 36.6% | 0.6% |
Why are managers employing such a quick hook? It’s all about a pitcher facing the lineup for a third time. Hitters facing the average starting pitcher the third time through the lineup hit .262/.327/.453 this season. Compare that to how they fared against the average reliever - .238/.321/.395. You may run into situations where a reliever doesn’t have it that day, but the odds are a fresh-armed reliever will do a better job than your tired starter showing hitters the same pitches they saw in the first few innings.
Is this a problem in baseball? As a longtime fan of baseball, I find it rather annoying to see a really quick hook. As Jack Elliott put it in the film Mr. Baseball, when addressing his Japanese manager:
Try letting your pitchers pitch their way out of trouble sometime. You know they might just reach down, and show you a little something. They surprise the hell out of you.
But this annoyance is largely because I grew up watching guys like Bret Saberhagen and Kevin Appier go deep in games. If you were to introduce someone to the game now, the quick hook strategy makes a lot of sense. Who cares which pitcher gets the last nine outs, so long as the other team doesn’t score?
However, baseball could have a problem if it makes pitchers anonymous, interchangeable pieces. Why pay top dollar for a Max Scherzer if he only goes four innings? How can fans ever get excited about a Mark Fidrych, Fernando Valenzuela, or Nolan Ryan if they know he’ll be gone by the fifth inning?
If baseball does want to elevate the role of the starting pitcher, there are a few things they could do. First, they could just lower roster limits. Right now rosters are at 26, up from 25 a few years ago, because managers use so many relievers. Baseball could bring that number back down, or require teams to have an “active roster” limit of say, five pitchers per game, that can be lifted if the game goes into extra innings or the opposition scores a certain number of runs.
Another interesting idea floated by a few writers is to tie the designated hitter rule to the starting pitcher. The DH only stays in the game as long as the starting pitcher stays in the game. Once you lift your starter, the reliever takes the DH spot in the lineup. Teams will have to consider if lifting the starter is worth losing one of their best hitters.
The game is changing thanks to analytics, with some of the changes long overdue and needed, while others may be creating a less aesthetically-pleasing fan experience. Baseball may want to consider the role of starting pitchers in this new game, and whether they should allow the role to continue to be de-emphasized, or whether it's time to tweak things to bring back the starter.