Weekend chart: starting pitchers innings over MLB history

With Baseball Hall of Fame season just passed, I’ve been thinking a lot of about measuring starting pitchers across different eras. I like to compare players across the two dimensions that really matter: productivity (creating or preventing runs per opportunity), and total playing time. The combination of these is essentially total value (WAR, etc), but it’s useful to break down into its constituent parts especially when it comes to more borderline cases.

For position players, it works just fine across time, since player plate appearances have remained relatively consistent. Though baseball history, everyday players have gone to the plate about 600 to 700 times a year, and careers tend to max out around 10,000 to 12,000 PA.

That’s definitely not the case with pitchers though. The active leader in career innings pitched, C.C. Sabathia, is just shy of 3,500 innings and unlikely to get to 4,000 innings as his career winds down. Mark Buehrle was considered the epitome of durability for a 21st century pitcher, and he didn’t even get to 3,500 innings. A pitcher today will do extremely well to get to 3,000 career innings, whereas in previous decades the long lasting pitchers would get well past 4,000 and towards 5,000 innings. And that’s to say nothing of 19th century pitchers, when there’s pitchers doing that in a decade by throwing 500 innings a year.

With such significant shifts over time, I was curious about what the standard for innings pitched by a front-line starter looked like across MLB history:

The first line in blue the average of the top 10 pitchers in innings pitched. In any era, these would be your aces or at the very least quality inning eaters (think Buehrle). But that’s by definition a very, very high standard, so I also extended out to the top 25 pitchers in a given season. It’s still a very high standard of pitcher seasons, but a little broader. Not surprisingly, the two lines largely track each other.

The final line in red is a more flexible standard. A fixed number means different things at different times as the number of teams increases (alongside the populations from which the pitchers are drawn). So the red line is the average of a number of pitchers equal to two times the number of MLB teams for that season — the top 32 pitchers per season to prior expansion, the top 60 today. It’s also an even broader measure (at least from 1901 onwards), but I think it’s good proxy for what would be expected of a front line pitcher.

Some observations and thoughts:

• Other than the very early days when the game splayed per season was expanding rapidly, the long term trend is towards fewer innings. What I didn’t realize though is that from the end of the dead ball era after World War I when 300 inning seasons became the exception rather than the norm, the average held quite steady for 40 years through the late-60s when there was a jump amidst the backdrop of pitchers dominating the era.
• The dropoff over the last couple years has really been remarkable, regardless of which cohort is used. The averages held pretty from the mid-aughts to 2014, but they’ve been in free-fall, falling by about five innings a year. Top 60 pitchers are down to just about 185 innings a year.
• 2017 and 2018 were the first years since 1979 that the number of qualified pitchers (at least as many innings as team games) was less than two times the number of teams, with just 58 and 57 respectively. The Jays didn’t their part, with none last year and just two in 2017 after five in 2016.
• Even if it was an entirely different game, the idea of pitchers throwing 500 and 600 innings in a season is just mind boggling.

Here’s an example of how this can help contextualize. Roy Halladay pitched 2,749 career innings (with a 76 ERA-) whereas Jim Palmer pitched 3,948 innings to a 79 ERA-. They were similarly good at preventing runs, you just seemingly got a lot, lot more from Palmer (about 45% more innings). But Halladay’s career works out to about 13 full season equivalents of the average frontline pitcher of his era (at about 210 innings/year), whereas Palmer’s works out to about 15-16 full season equivalents. Still a meaningful difference, but a lot closer than at first glance.