Starter's Workload: Stop Blaming Analytics for 122 Year Trend
One often hears baseball fans and pundits alike lament the shortened workload that MLB starting pitchers are asked to bear in the modern game. Often in the course of such discussion the argument that "analytics are ruining the game" is trotted out and held aloft as some great insight.
The next time you hear that, please note: That's complete and utter nonsense. In fact ever since 1901 the percentage of innings pitched by starting pitchers has continually decreased from previous years and decades. There is some ebb and flow of course, but the 122 year trend is undeniable, and did not just start since the popularization of modern analytics over the last 20 years.
Jack Sommers
One can plainly see that the decrease began immediately after 1901, and despite some ebb and flow within smaller time frames, there has been a continual downward trend. The average rate of decrease for the entire 122 seasons is 0.3%. Notably however there are several points in history where that decrease was accelerated for a seven or eight year stretch.
- In 1918 SP IP% was 96.1%, dropping to 87.9% by 1925, a decrease of 8.5% in just 7 years
- In 1950 SP IP% was 85.8%. dropping to 76.0% by 1957, a decrease of 11.4% in just 7 years
- In 2014 SP IP % was 66.5%. dropping to 58.7% by 2022, a decrease of 11.7% in 8 years
- Note that the percentage actually dipped to 55.5% during the 2020 Pandemic season, but bounced up 3% over the last couple of years.
There is a lot to unpack when looking at this, and we are not going to get to all of it today. But we can probably guess at a few reasons behind this continual change and what might have been behind these acceleration points.
Offense began to increase in 1919, and then exploded in the 1920's with the advent of the "live ball" era. More offense means more baserunners and more pitches to get outs of course. That coupled with an improvement in economic conditions helped improve the depth behind the starting pitchers.
By the mid 1950's integration and even better economic conditions improved the depth behind the starting pitching a great deal. Those two factors most likely led to the accelerating decline in starting pitcher innings during that decade.
The accelerated pace in the decrease of starter's innings over the last 7-8 years can partly be tied to teams making conscious choices as a result of the analytic work being done. And certainly a lot of focus has been put on pitcher's health. But in fact even these accelerated decreases are right in line with the other periods of acceleration and the overall 122 year trend.
Why has the game evolved this way? Simply because the quality of play on both sides of the ball have improved incrementally. Now before anyone starts yelling at me, let me say this:
Players that were great in decades past would be great today and players that are great today would have been great in decades past. They just would be doing things differently depending on their environment, context and placement on the evolutionary timeline.
Pitchers throw harder and with sharper movement than ever. While there were always a few guys across history that maxed out velocity and movement potential, they were outliers, and because of that they completely dominated relative to their peers. Walter Johnson, Bob Feller, Sandy Koufax, Nolan Ryan, et al....all likely threw as hard or close to hard as pitcher's today. But the vast majority of pitchers could not approach today's speed and movement of even the average pitcher in 2022.
Meanwhile, go watch You Tube videos of games from the 60's and 70's. Pay special attention to the hitting mechanics, and the footwork. You see a lot of guys spinning out and off balance with long loopy swings and strange batting stances. Those just wouldn't work in the modern game. Hitters have cleaned up and tightened up those mechanics out of necessity in a gradual, evolutionary process.
This evolutionary process on the part of both the pitchers and hitters has resulted in gradual trends in the statistics we use to measure on field performance and events. So the numbers simply look different. Another example of this is what is known as "three true outcomes" . This was a phrase coined by Bill James to describe walks, strikeouts, and home runs, or outcomes that do not involve the fielders. That rate has continually gone up over the years in a non stop fashion. I include Hit by Pitch in my table below, so consider it "four true outcomes". As one can plainly see, this chart is not the result of an analytics revolution. The trend here also existed long before the book Money ball came out. This is not the "fault" of analytics, it's simply a continuation of the natural evolution of the game. Recent rule changes may retard this upward trend or even temporarily reverse it, but it won't likely last.
Jack Sommers
So what's next? If the rate of starting pitcher's innings workload continued at the 122 year trend of decreasing by 0.3% per year, then the percentage will be below 55% by 2034 and below 50% by 2050. There could obviously be another acceleration point in the meantime as there has been in the past to thrust this average below 50% well before 2050.
But there may also be an efficiency threshold at some point that will cause this line to flatten out. We may already be seeing that with the slight increase the last couple of years, or this may just be a temporary increase before resumption of the trend. What's almost 100% certain is there will be no going back to the days of even 70% of innings going to starting pitchers as last seen in 1988.
From an aesthetic point of view, that's sad perhaps. I personally miss the days of starters throwing complete games and dominating for all nine innings. It's still the most exciting thing for me. But lets stop blaming "nerds in basements", shall we? It's not their fault, (nor should they take credit). It's natural evolution.
