Reimagining the modern pitching staff

Our understanding of baseball has changed a lot in the last couple of decades. The sabermetric revolution kicked off a more analytical approach to the game, where the data that had existed for years was suddenly revealing new truths. And in the last few years, just as the pace of new revelations seemed to slow, Statcast has provided additional grist for the mill, with its wealth of detailed data measuring previously unquantified aspects of the game.

And yet, through all that, the way baseball is played has remained remarkably stable. The range of acceptable players has perhaps broadened somewhat — the value of a high-strikeout, high-OBP player is better understood, for example — but the players still do all the same things they did before, and in mostly the same way. The advent of shifting has probably been the single largest change, and in the big picture, it’s minuscule: FanGraphs reports a shift of some kind on less than 30 percent of balls in play in the 2016 season, and even when there is a shift, it tends to consist of nothing more than a single player moving from one side of second base to the other.

The kind of creative, aggressive strategic changes that all this analysis could have enabled simply haven’t materialized. We know all about platoons and spray charts, but while sticking a lefty specialist in the outfield for a single righty batter has happened before, it’s exceptionally rare. We can figure out what hitters are most vulnerable to a five-man infield or a four-man outfield, but it never actually happens. Baseball is fundamentally the same as it was 40 years ago.

It’s too bad, even from nothing more than an entertainment perspective, because a lot of these alternatives would be pretty interesting. Last week, I wrote about the ability of managers to maximize great relievers in a way that they can’t capitalize on great position players or great starters, because they have total control over when relievers enter the game and can save their talent for the situations where it makes the biggest difference. All of Andrew Miller’s innings can take place in close games; the same is not true of Francisco Lindor’s plate appearances, or Corey Kluber’s starts. But there’s no reason that pitching staffs have to consist of five starters, each responsible for 150–200 innings, and some number of relievers to take over after them.

We can imagine, for example, a pitching staff made up of some number of hybrid starter-relievers and some number of “pure” relievers, without any traditional starting pitchers. At the beginning of a game, the score is tied and the game is guaranteed to be close. These are innings that you definitely want to give to talented pitchers. Sometimes, however, it’s 5–0 by the fourth, and you’d rather take the four innings Clayton Kershaw is about to pitch and use them to start the next game (when the score will again be 0–0) instead of using them in a game that’s probably out of reach for the opponent.

With the current way pitching staffs are structured, you can’t, of course. Kershaw is a starter, and starters pitch their innings in single chunks once every five days. But if you could fragment those innings into smaller chunks, they could be distributed to align more closely with the parts of games that are actually important.

Would the math work out? The median pitching staff in 2016 threw 1,443 innings, and the standard roster saves 12 or 13 slots for pitchers — we’ll go with 13. That’s an average of about 111 innings per roster spot. The reliever who threw the most innings in 2016 was Brad Hand, with 86 ¹⁄₃, which is roughly in line with the maximum seen in most years, so it’s plausible to assume that former starters converted to a quasi-relief role could break easily break 100 innings if they were used two or three times per week for 2–4 innings at the beginning of games.

So imagine the following pitching staff, which doesn’t seem to be fundamentally out-of-line with reality. The innings listed are the number of innings they’ll pitch in our hypothetical no-starter system, and the ERA is how they’ll perform in those innings. (The numbers are totally made up; the point of this is just to show how this system might work, not that it would be a great idea.)

The more durable pitchers are the ones who would generally start games, with the group behind them being chosen based on how close the game remained after those initial innings. Ideally, however, any one of the pitchers could come in at any time, and while rest would be an issue, defined roles like “starter” or “closer” wouldn’t be. The innings thrown by the best pitchers could be concentrated in the most important situations; no more aces pitching in the fifth inning of a blowout.

But as with every other aspect of baseball, pitching staffs have been almost totally stable over the last several decades, with the only changes being minor tweaks like adding another roster spot. This suggestion, like a five-man infield, will remain strictly a hypothetical.

It’s probably obvious what’s preventing this kind of innovation: inertia, and the difficulty of getting players (and coaches) to buy into a system that has no resemblance to the baseball they’ve watched or played in the past. Nor is it just mental; there would probably need to be some serious physical adjustment that starters and relievers went through to move to this set-up.

Both of those are totally legitimate reasons for pitching staffs to stay stable over the last several decades. But it’s a shame, because it’s keeping us from the kind of exciting, experimental baseball that I would love, and that might inject some excitement into a sport that could use it.