The Five-Inning Start is On The Rise

This post piggybacks on a really interesting piece from Jay Jaffe of FanGraphs, who, a couple days ago, examined in depth the changing workloads of starting pitchers in 2020. Between a truncated “Summer Camp”, the expanded rosters and taxi squads, in-game rule changes, and a drastically shortened schedule (let alone a literal pandemic), it is near impossible to determine how exactly things will play out in 2020.

Early on though, as Jaffe highlighted, starting pitchers are very simply pitching less. Toward the end of his piece, Jaffe highlights the number of 4 inning, or 12-out, starts provided by starters on each team in 2020. As an extension of that view, what follows are just a couple figures to further explore. First, what length starts are being made most often in 2020 and second, what length starts were being made most often to “this point” in 2019, in order to draw a comparison across seasons.

Below, you’ll find a histogram of start lengths as measured by outs recorded. Note the ratio of 5 inning (15 outs) vs. 6 inning (18 outs) starts.

The various spikes in the histogram above generally represent full innings. The most common start length in 2020 is a 6-inning outing, but only just barely. Through August 10th, just one fewer start was made of 5 innings than of 6 innings. Of significance too is the relatively low total of 7-inning outings: in 460 starts, just 24 starts (or ~5.2%) have gone exactly 7 innings. To this point in 2020, the 4-inning start is a far more regular occurrence than a starts of the 7-inning variety.

There have, on the other hand, been two complete games early on in 2020: they came by way of Kyle Hendricks on Opening Day and later on from a resurgent Dylan Bundy. As a final note, managers in 2020 seem to be increasingly pulling starters mid-inning (more on this later). This might be a result of new rules relating to relief pitchers or due to increasingly strict pitch count limitations. Meanwhile, Shohei Ohtani owns the ignominious distinction of having started the only game without recording an out.

Below is a list of the most common start lengths through this point in 2020.

To provide some semblance of context for those 2020 numbers, similar figures populated with 2019 data are covered next. Data from Opening Day 2019 through April 14th 2019 is included. This timeframe was picked primarily to closely match sample sizes across seasons. For 2019 data, a sample of 462 starts is included.

A couple points tend to jump out upon seeing this histogram. First, 6-inning starts are considerably more prominent than 5-inning starts, relative to their ratios here in 2020. Out of the first 460 or so starts of each year, 80 starters went exactly 6 innings in 2020 (17.4%), versus 109 in 2019 (23.6%). Additionally, 7-inning starts were more than twice as common as 4-inning starts in 2019, representing a drastic reversal of those totals in 2020.

One might also notice what appears to be fewer starts that end mid-inning in the outset of 2019. There had been 95 mid-inning “pulls” between 3 and 6 complete innings in 2019 versus 118 in 2020, a 24+% increase. Given 28-man rosters (and correspondingly deeper bullpens), in tandem with the 3 batters rule, managers might be playing for favorable matchups while simultaneously circumnavigating that minimum rule by pushing relievers to finish, rather than start, some innings. Interestingly, there was just one complete game through the first 462 starts of 2019, and it was delivered by German Márquez on exactly April 14th, 2019.

Below is a list of the most common start lengths of 2019 through April 14th.

Given the collapsed nature of MLB’s “Summer Camp” as well as the long layover between that camp and Spring Training, it makes sense that starting pitchers might be treated with extra caution to this point. Still, given all the other variables at play, it remains to be seen whether they are given more opportunities to pitch deeper as August transitions to September.