The State of Starters in 2024

I won’t sugarcoat it for you, friends. It’s a tough time to be a major league starting pitcher. Their ligaments are under threat like never before. Their workloads aren’t far behind. For a variety of reasons, the old style of starting pitcher is quickly headed toward extinction and we’re transitioning to a new way of doing things.

That all seems like the obvious truth. But I decided to go to the data and make sure. As Malice of the Clipse (and yes, fine, Edgar Allan Poe) memorably said, “Believe half what you see, none of what you heard.” I’m not sure exactly where that leaves you, since I’m going to be telling you what I saw, but that’s an epistemological question for another day. Let me just give you the data.

So far this year, there have been 452 games, and thus 904 starts. Starters have completed 4,735 1/3 innings, or 5.24 innings per start, and they’ve thrown an average of 86.2 pitches to get there. They’ve averaged 94.1 mph with their four-seamers, yet despite all that velocity, they’ve thrown fastballs of any type just 54.9% of the time. This isn’t Opening Day starters, or anything of that nature; it’s just whoever has picked up the ball for the first pitch on each side.

For a lot of people, that’s all the evidence they need. 86 pitches? Gross. 5.24 innings? That’s nothing. Back in my day (2008), 34 different pitchers topped 200 innings. Surely, these are the lowest innings pitched and pitches per start in history, and surely that’s connected to everyone throwing so dang hard and not throwing as many fastballs, right?

I’m not really sure how to prove the second part of that previous sentence, but the first part isn’t quite accurate: Those 86 pitches and 5.24 innings per start are not the fewest in history but some of the fewest. You see, the decline of the starting pitcher is not a 2024 story; if anything, the decreasing workload of starting pitchers is a pandemic story. I isolated the first day in each season where there had been at least 904 starts and compared them to this year to control for the fact that pitcher workloads follow a predictable ramp-up at the beginning of the year. Here’s the data for 2018-2024:

A trend that started in 2019 accelerated thanks to the strange and compressed start to the 2020 season. It continued in 2021 and 2022 (post-pandemic ramp-up was strange, and then the lockout made spring training shorter) before rebounding in 2023, and this year looks a lot like last year. Pitchers are completing slightly more innings per start, though. Why? Fewer homers:

Again, all of these statistics are from similar time periods within each year, with the exception of 2020, when the season started in late July. I don’t think this is a weather story. Balls aren’t leaving the park as frequently, which means that starters are recording slightly more outs per appearance. That just makes good sense to me.

The next part of this story is to compare the piddling workloads of the present to the supposed king-sized portions of the past. There’s just one problem – that’s not how the data looks. In 2008, the first year that Pitchf/x tracked every pitch thrown, starters completed 5.75 innings per game and threw 93.5 pitches to do so (through May 3 and 914 starts). In fact, per-game starter workloads haven’t declined much at all over the last 17 years:

Wait, really? In the halcyon days of the mid-2000s, starters were managing only seven more pitches per start on average than they are now. Even if we take the high water mark of our data set, it’s only 10 more pitches. That’s with the current injury epidemic, the rise of openers, the intense focus on the times-through-the-order penalty, and all that. Ten pitches isn’t nothing, but it’s also not some unbridgeable gap.

Three-quarters of an inning – the gap between the recent high water mark of 6.01 innings per start in 2010 and this year’s data – sounds more significant than 10 pitches. The truth is, though, that the number of pitches it takes to record an out has been shockingly consistent over time. In 2008, it took 5.42 pitches on average to record one out. That number got as low as 5.33 in 2012, a year that featured low rates of each of the three true outcomes. So far this year, starters have needed 5.49 pitches per out. The average across the last 17 years is 5.46.

In other words, 10 pitches and three-quarters of an inning are closely correlated. If starters were posting strikeout rates below 20% and allowing less than one home run per nine innings, they’d probably be pitching slightly deeper into games today.

If there’s one worrisome sign about this year, it’s that a decline in home run rate hasn’t increased innings pitched totals as much as you’d expect. But that’s small potatoes in the grand scheme of things. In fact, most of this is small potatoes in the grand scheme of things. I’m sure that the fans and pundits bemoaning the decline and fall of the starting pitcher empire are referring to something more dramatic than losing fewer than 10 pitches per game.

For me, a bigger question is what will happen as the season wears on. From 2008 through 2019, workloads remained stable between the first month of the season and the rest of the year, whether you care more about innings pitched or pitches per start. But the relationship has broken down since then, and with good reason. Three of the last four years have been oddballs in one way or another, which makes pattern recognition difficult. In 2022, workloads increased markedly as the year wore on, with pitchers throwing nearly five more pitches per game than in the early going. In 2023, workloads declined by two pitches per start as the year progressed. They pitched just as many innings, though. There’s certainly no strong correlation in any direction here.

Where does that leave us? I’m sure this isn’t satisfying, but I don’t see much signal. Pitchers are inarguably working less deep into games than they used to. But things haven’t changed much over the past few years, and it’s just not the kind of sea change that people make it out to be. In 2008, starters faced 25 batters per game on average. In 2024 so far, they’re facing 22. That’s different! But let’s not kid ourselves — starters today aren’t always coming out as soon as they go twice through the order, just as they weren’t all pitching into the eighth inning in the mid-2000s.

Are injuries playing a role? I mean, probably! What role? I’m not sure. Is decreasing fastball usage, or increasing fastball velocity, or increasing strikeout rate the prime driver? None of those are obvious conclusions. Are the very worst pitchers better than they used to be, and thus are the shortest starts longer than they used to be, which in turn would hide some of the decline in top pitcher workloads? Quite possibly so – thanks to Meg Rowley for pointing this one out, and you can expect a follow-up that digs into that very question next week. The number of possible interpretations here is huge because the data looks so similar from one year to the next.

The only thing I’m comfortable saying is that for all the justified worry about pitcher workloads declining and pitcher injuries increasing, this year’s replacements have been quite reasonable when it comes to soaking up innings and retiring batters. Is a deader ball helping them? Very likely. But whatever you hear about the pitching crisis is about as true as it was in 2019. Pitching simply doesn’t work like it used to – but it still works pretty well at the end of the day.