The Sweet Spot for Pulling a Starter in 2023, and How the Brewers Found It

Most of the time, a starter makes it at least twice through the batting order, even if they're struggling. Except in blowup starts or cases in which managers intend a strategy like the opener or a bullpen game, teams need their starter to get into the third trip through the opposing lineup. That prevents the bullpen from being overtaxed, since facing at least 18 batters often gets a starter at least into the fifth inning. Ideally, of course, the starter goes more like six or seven innings, even in the modern game. 

As a pitcher works deeper into a given game, though, they tire, and as each hitter in the order gets more and more looks at the hurler's stuff, they slowly gain an advantage over them. A manager ignorant of the times-through-the-order penalty (or just insufficiently proactive in mitigating it) can lose a lot of games before even getting the chance to have their bullpen save them.

As we all know, starters pitch much less deep than they used to, and it's not only because of times-through-the-order effects. Still, that's a key factor to consider, and it's also different from one year to the next. When, on average, should a manager let the progress of the game nudge their starter off the mound? It depends as much on the dynamics of the opposing order as on the pitcher himself.

Start, then, by considering the average relative production from each place in the batting order this season. I also tried to place that production in historical context. To do so, I confined my search to the American League, over the last 51 years. That way, we're looking only at lineups that go nine players deep, without the complicating factor of the pitcher in the ninth spot.

League OPS+ By Lineup Spot, 2023

As you can see, the third through seventh spots in the order were about as bad (relative to the rest of the league) as they have been at any time in the last five decades. By contrast, the top and bottom of the order were more dangerous than they have traditionally been. Modern lineups are flatter than their predecessors, and the bad hitters in them have more power than they used to. The danger in a lineup starts earlier, but in the middle, you don't have titanic and well-rounded sluggers to fear.

Managers have reacted accordingly. By looking at the frequency of each number of batters faced within a start throughout the league, we can see trends and changes in when the skipper is going to get his starter and turn to his relief corps.

Again, we already knew about the macro trend toward starters facing fewer batters and working less deep into games. What this illustrates, though, is that it's not a uniform change. There's a distribution shift taking place, too. The bell curve is taller and thinner than it was in the past. Notice how much more often pitchers are lifted before seeing the first or second hitter a third time. The biggest spikes are in the relative frequency of taking pitchers out after 22, 23, or 24 batters, though. That means that skippers are disproportionately removing their guys as they wade through that soft(er than the historical norm) spot in the middle of the batting order.

The sweet spot is right in there. Assuming a good bullpen and a healthy starter, managers do best when they take their starters out after somewhere between 21 and 25 batters, at least under 2023 conditions. That worked like gangbusters for the Brewers this season. Only the Orioles lifted their starters in the 21-25 sweet spot more times (108) than Milwaukee's 105. Craig Counsell only left his starters in past that sweet spot 21 times. Given the shape of their pitching staffs, no team better timed the key substitution of any ball game--taking out the starter--than did the Crew. It was part of why they were so sensational at run prevention.

Many of these dynamics could subtly shift again in 2024. Many things could change on the Brewers' end, too, including Brandon Woodruff being out for the year and Counsell (perhaps) not being the one pressing the buttons and pulling the levers. For this season, though, everything worked gorgeously, because the team was ahead of the curve of league-wide change.