Some inconvenient truths about pitching
From opening day to September 1st, no major league team can have more than 13 pitchers, half of the total number of players on the roster. On September 1, the numbers go up to 14 of 28. On average, teams end up using between 25-30 different pitchers over the course of season, split evenly between starters and relievers. Last year, the Mets used 46 different pitchers, thereby setting the major league record, roughly 3.5 times the number of pitchers allowed on a major league roster. Impressive, but not in a good way.
Math problem of the day: How many different combinations of 13 pitchers would 46 different candidates yield? And how many of the 46 can you name? How many do you think even Carlos Mendoza or Jeremy Heffner (without a cheat sheet) could have named once the season ended?
Interestingly, the Braves, Angels, Diamondbacks and Orioles managed to use 40 or more pitchers as well, and the Dodgers fell just short of doing so after having hit the 40 mark in 2024.
Pitchers experience more injuries than ever in spite of pitching fewer innings and appearing in fewer games. Tommy John surgery has become a right of passage. There seem to be more and more relievers and fewer and fewer effective starting pitchers. Fans invariably find themselves praying their team's staff can get the game in the hands of the closer, only a handful of whom are capable of reliably closing down the opposition, anyway. No doubt, position players on any given day find themselves in quiet prayer as well
It's unimaginable that a team would go through position players at anything approaching a comparable clip, even though position players play many more innings and appear in many more games.
The explanation of the difference lay in the nature of the workload involved and especially the toll pitching takes. Pitchers are involved in every pitch in every inning in which they appear which means that their workload is not best reflected in the number of innings or appearances they make, but in the number of pitches they throw and the conditions under which they do so. The effort and toll on the body each pitch requires is enormous. From the point of view of impact on the body, there is no such thing as low impact pitching All pitches are high impact, on the body if not necessarily on the outcome of the game.
There are no doubt a number of ways of answering this question. Managers invariably think of their pitching staff in terms of innings. They need their pitchers to cover innings. Asking pitchers to cover too many innings risks injury. Having pitchers cover too few innings runs the risk of overusing other pitchers and thereby also risking injury. When they are not discussing covering innings, managers typically speak about the need for the staff to keep the team in the games as often as they possibly can.
As constructed, far too many pitching staffs are finding it difficult to meet these goals. The Mets needed 46 pitchers (in addition to a couple of position players) to cover the required innings. It is not be surprising that digging as deeply as they did into the pool of pitchers (and beyond) that they were unable to stay in enough games to catch the break that would have led them to a playoff appearance.
This problem is not unique to the Mets. Something is amiss with how teams in general construct their pitching staffs. In my view, the problem begins with an inapt idea of what teams are building a staff for: what problem they design the staff to solve.
Another way of thinking: A pitching staff is designed to create strategic advantages
I am playing around with a different way of thinking about a baseball team and its roster construction. That overall idea is incompletely developed, but goes like this. Arbitrarily, perhaps, let's divide the roster into offense, defense, and pitching. At some level, I think you want to think about each of these three areas as having the same goal, but with different strategies for achieving that goal.
The overall goal is to efficiently create exploitable strategic advantages over the course of a game, a series, and ultimately a season. Obviously, Stearns looked at the Mets' overall capacity to prevent runs and determined that it was a strategic liability. Run prevention is on the defense and the pitching. So the question is, how to turn a strategic disadvantage around, and ultimately into a strategic advantage.
I think Stearns is thinking in effect along the lines I am proposing. I'm just focusing here on what would be the natural questions he would ask himself about pitching and a pitching staff. They are (1) How can a pitching staff, taken as a whole create a strategic advantage; (2) How would one organize a pitching staff around the goal of securing strategic advantages.
Because pitchers throw pitches to batters, one at a time, one batter at a time, the focus on innings is inapt to this way of formulating what the goal of a pitching staff is.
The goal must have more to do with pitches than with innings. And it must have something to do with creating a competitive advantage wherever possible of the pitcher to the hitter he is facing, and turning that competitive advantage into a strategic one, that is, one that has positive outcomes in (1) at bats, (2), innings, (3) games, (4) series of games, and (5) over the course of a season.
Making sense of the idea that pitching staffs are designed to create strategic advantages
The basic idea is pretty simple and straightforward. Pitchers throw pitches. And what we want from them as a staff is for them to throw the maximum number of quality pitches over the course of a season, distributed over games and innings within games. Ultimately we are looking to the data to determine which pitches, when, by which pitchers leads to the most favorable outcomes.
What do I mean by 'quality pitches'? Quality pitches are those that in sequence give the pitcher the competitive edge in an at bat. The competitive edge if realized would result in more at bat pitching wins, which translates into an overall strategic advantage. Ideally, you would want every pitch to be the best version of the pitch that the best strategy would call for at that point in the game.
Practically speaking, the best any team can hope for is that taken as a whole its pitching staff will maximize the opportunities to have a strategic advantage as often as possible -- in at bats, across innings, games and seasons.
That's the goal, now the question is, what plans can one plausibly implement to move in the direction of achieving it?
Here's my idea, my defense of it, and why I think it may prove to be a reasonable alternative approach to the ones that have evolved within the existing pitching paradigm. It's easier to make the idea concrete by using an illustration.
Since last year's Mets serve as the most vivid and painful example of what can go wrong under the existing paradigm, let's use the current Mets projected staff (I'm obviously making some assumptions about its make-up) as an example of how thinking in terms of maximizing strategic advantage would influence the construction of the staff and the use of the pitchers on it.
There are currently 6 starting pitchers on the staff all of whom have shown a capacity to pitch very effectively for 3 or 4 innings, My first suggestion is that we match 2 of them in any one game, strategically, and that we do this for 4 of the pitchers on the team.
Now how should we match our pitchers with one another? What criteria should we use? For starters, I suggest pairing righties with lefties, as well as paring different pitching styles. pitch arsenals, or arm angles, etc. Anything that reduces comfort in the batter's box. These are suggested criteria. Think of them as a first pass, revisable upon new information. So how about:
* Senga and Manaea
* Peterson and Holmes
That would leave Peralta and McLean to pitch without an assigned mate, i.e as standard 'starters.'
Each member of a matched pair is expected to pitch 3-4 innings in a game. When they are successful at doing so, on average they would pitch through the 7th inning leaving only the 8th and 9th. Then one would have McLean and Peralta as traditional pitchers expected to go 6 innings each, which when they are successful, would leave the 7th, 8th and 9th to be covered by relief pitchers.
One obvious problem is that the best we can do with 6 pitchers is cover 7 innings of 4 games. We are one game short of the modern standard of a full rotation's collective workload. As an initial thought, I suggest that we increase the number of 4 inning pitchers to 8, which still leaves us with 5 so-called relievers on a pitching staff of 13. For the purposes of illustration only, let's add Tong as the righty and Thornton as the lefty to fill out our initial rotation of 4-inning pitchers. Now we have five games covered on regular rest for everyone. The relievers are: Williams, Weaver, Raley, Minter when he's back, Myers and Garcia.
The presumed workload of everyone is not only reduced, but much better defined. Every 4 innings pitcher is asked to pitch through the opposing line-up no more than twice and more likely less than that. In general, pitchers are most effective under these circumstances, which in my interpretation, means they win far more than they lose battles with hitters, thus exhibiting a competitive advantage and creating a strategic one.
In most games the reliever's tasks are extremely well defined and their workload appropriately distributed. There are fewer pitchers in the bullpen because there are fewer innings in fewer games in which they are likely to be called upon.
At the very least you have your best pitchers throwing in the circumstances in which they are most likely to perform well. Of course, they won't always do so.
All you can reasonably do is put them in the best position to do so.
I would add one 'in game' suggestion to the usage strategy set out above. If, say, the first pitcher on a day in which there is a tandem of pitchers, is having a rough outing and needs to be taken out of the game in the third inning, he should not be replaced by his mate, but by a reliever who gets the team through the 3rd or the 4th inning and is then replaced by the second pitcher in the tandem.
The second pitcher in the tandem is a former traditional starter and at least at the beginning will need more time and go through a different ritual to prepare for his outings than would someone who had been and continues to be a reliever.
So if Senga is having a rough go and needs to be replaced, he should be replaced by a reliever who one might even think of as a high leverage pitcher, or something of a competent if not elite 'closer' under the old paradigm. That pitcher, say Weaver or Garcia, closes, in effect, the first four innings, a mini game within the game. Then Manaea enters.
One objection and many unresolved issues
Let me anticipate an objection. I am claiming that I am offering an alternative way of thinking about the goal of a pitching staff and the problems it is designed to solve. Yet here I am talking about innings, and isn't that just the old way of thinking that I am criticizing?
Fair question.
I am re-introducing 'innings' but not as part of the goal of the staff, but as a proxy for being able to throw pitches creating a strategic advantage. The paradigms are different because the goals are different. Innings matter only to the extent that they reflect a pitcher's actual effectiveness in creating a competitive advantage and turning it into an exploitable strategic one.
Innings are a proxy because the data shows that pitchers are typically as good as they can be on a particular day the first and second time through the opposing line-up, while being considerably less effective thereafter. which is just another way of saying that they can no longer have a competitive advantage often enough, and their ability therefore to create an overall strategic advantage has decreased materially, so much so that it may, and often, does become a liability. Their ineffectiveness shows up in increased walks, hits and runs given up after the second time through a line-up.
There is so much more left unresolved of course: from metrics for compensation to the relative attractiveness of being a pitcher on a staff organized in this way.
One suggestion: the Spring Training and early season beta test
Of course, the problem with novel ideas, both grand and small, is that reality can get in their way, sometimes in predictable way, other times in surprising ways; some ways that are manageable, and maybe others that are not.
I would like to see whether this kind of thinking actually holds water in practice. And there is an appropriate time to give it a try in the next few months. In the later stages of Spring Training as the pitching staff takes shape, and pitchers are just beginning to stretch out a bit, why not devote a few games to 'pitcher matching' experiments. Indeed, if it is at all promising on initial testing, why not use it early in the season while pitchers are still being stretched. Not only would doing so provide useful information and perhaps result in more than the normal number of wins, it would also prevent overuse of the bullpen.
Next, treat the minor leagues as a testing ground.
At the end of the day, the current approach to constructing a pitching staff isn't working, not just for the Mets, but certainly for the Mets. There is always risk in thinking genuinely outside the box, but when it comes to pitching, and given the assorted problems pitching staffs face and their apparent intractability, perhaps the time for taking risks is upon us!
What do you think?
