There are apparently certain things in life that you don’t want to last for more than four hours. They might be fun for the first half-hour, but after a while it just gets dangerous for your health and you should call a doctor. Of course, I’m talking about baseball games, and for the past few years, MLB has talked about ways in which they could make the game go a little bit faster and wrap up in a shorter amount of time.

The ideas have ranged from "apocalyptic when proposed, mundane when implemented" (the “guidance” around batters keeping one foot in the batter’s box), to the “why didn’t I think of that?” category (the reliever just spent 10 minutes throwing warm up pitches in the bullpen… why does he need eight more from the mound if we’re all just gonna stare at him?). Then there are the ones that are somewhat more radical. Perhaps a less gracious adjective would be “crazy.”

For example, every once in a while, a proposal is floated that the solution to long games is to simply make them shorter. As in, to reduce the number of innings in a standard game to seven, instead of nine. In fairness, no one really expects that one to take hold, but I think it makes for an interesting study in really going deep into an idea before declaring it the solution to all of life’s problems. I think the idea is borne from some crude math that says that if the average baseball game is nine innings long and three hours long, then seven-inning games would take two hours and 20 minutes on average.

If baseball were different, how different would it be? It would be a seven-inning game. It seems like chopping off the last two (scheduled) innings would be such an easy fix, but baseball is a system and if you change something about a system, you change everything about a system.

So let’s jump in and see what else would be different in a world of seven-inning baseball. Other than getting used to the fifth-inning stretch.

Warning! Gory Mathematical Details Ahead!

For this exercise, we’re going to assume that after the last out of the 2016 World Series (presumably in the ninth inning), Commissioner Rob Manfred announces that for 2017, Major League Baseball is switching over to a seven-inning game. Nothing else will change. All previously signed contracts are still valid. Still doing the 162-game schedule. Still having 25-man rosters. Still having people dressed up as over-sized food products racing against each other between innings.

The game would be thrown into a bit of chaos at first, but over a couple of years an equilibrium would develop. But what would develop? The (overly) simplistic answer is that most of the game would remain the same. Starters would still go 6-7 innings (which would now be a complete game) and there would be less of that whole bringing in relievers thing and having seven or eight different pitchers take the mound during the game. But past that, Rich Hill would still be Rich Hill and life will continue.


Let’s back up for a moment. Baseball is a game of resource management under uncertainty. Once a player is used in a game, he can’t re-enter, and you have no idea whether this is the best place to deploy him tactically or whether there will be a better time to use him later in the game. On top of that—at least with pitching—using a player now might have implications for whether you can use him in the game that you will play tomorrow. It’s instructive to note that despite the fact that teams need to have eight position players in the field at once and only one pitcher, they generally devote half of the available roster spaces to pitchers.

Managers aren’t completely without information. When managing a game in-line, they know what the inning is and how much more of the game there is to go. At least they have some mental probability map. A manager can say “We’re in the seventh, so I know I need to get through the eighth and ninth as well.” He can know that there is a non-zero chance that the game will end up in extra innings. He also knows how close this game is and whether or not it’s worth using the good relievers versus saving them for tomorrow, which may or may not be a close one. This is all well-trodden terrain, but there’s something that we need to say that sounds pedantic, but is crucial to remember and fully reckon with: All of those decisions are made within the context of an ecosystem of a nine-inning game. The calculations change when you begin to assume seven innings.

Let’s look at an example. Consider a manager deciding whether to pull his starter after the fifth inning or sending him back out for the sixth (of nine). Let’s assume that his pitcher has been doing pretty well so far, but there comes a point where even if he’s having a good game, you need to say “the relievers I have available are just better.” Call it a tactical hook, rather than a forced hook. For a moment, let’s leave aside any consideration of tomorrow’s bullpen needs. Let’s just do some #GoryMath on this decision. In the sixth inning (of nine), it’s basically a question of whether the starter in his sixth inning of work is more likely to get through this inning unscathed or whether the guy who is probably the fourth or fifth best reliever in the pen is the better bet.

In 2015, starters (as a whole) had a combined RA9 of 4.39, while relievers had an RA9 of 4.07. Of course, the average starter isn’t a true-talent 4.39 RA9 guy in the first inning. We know that starters perform worse as their pitch counts go up, and in the first inning he’s still fresh. We also know that the reliever that’s going to come into the sixth inning is probably a below-average reliever. But at some point, that scale tilts the other way. When does that happen?

Let’s take a look at a subset of starts from 2015, games in which the starter was pitching well enough to go five innings—as evidenced by the fact that, retrospectively, we know that he managed to go at least five innings. There were 3,990 such starts last year.

Here’s the RA9 by inning for each inning of those starts:













Now, this is a very selective sample. These are guys we know in retrospect were having at least an okay day, if not better. They didn’t flame out in the third inning. But we see that if all we know is that they were good enough to get through the fifth inning (and nothing else), we can assume a 3.89 RA9 against them if they continue into the sixth, and that’s a sample drawn specifically from pitchers who got the stamp of approval to go out for the sixth. If we believe that managers and pitching coaches have some clue, they would have kindly removed a lot of the guys who would have been disasters in the sixth from our sample.

Now, let’s compare those numbers to the aggregate bullpen RA9 that we saw in the sixth through ninth innings in 2015, first for all games, and then for situations in which the game was within three runs either way.


All Games (relievers only)

Game within 3 runs (relievers only)













Let’s compare those two sets of numbers. In the nine-inning game, the type of reliever who would come into the sixth inning in 2015 is much more likely to exit the sixth inning having allowed a crooked number on the scoreboard. Ditto even for the seventh inning. Of course, it will vary for each manager based on his roster, but that “league-average” sixth-inning reliever isn’t looking like such a good option right now. Of course, in the seven-inning game, our manager doesn’t really need to go to what has previously been his sixth-inning guy. After the fifth inning, there are only two more left, so he can go to what had been his eighth-inning guy. And that starts to look like a decent tradeoff. We need to win this game.

Suddenly, there’s a lot more pressure for the manager to tactically hook the pitcher more quickly, even though under the old rules, that pitcher probably would have gone out for the sixth. Add in the fact that he doesn’t have to worry about filling as many innings tomorrow (there will only be seven). And while we’re at it, if a starter is getting in trouble in the third inning, rather than agonize over whether to let him try to fight through and save some of the carnage that comes from having the pen pick up 6 2/3 in a night, it’s suddenly only 4 2/3 that he has to worry about. So why not extend the hook a little earlier?

It’s tempting to just stop there, but I think that has some knock-on effects too. It seems that a fait accompli that starters would be asked to do less (even less than they are doing now), even in situations where they could conceivably do more. I could see this leading in a couple of different directions.

· Starters, knowing that they only need to go five innings, rather than aim for seven, might adopt an “air it out” sort of mentality. If I don’t have to make it to 100 pitches, why not put a little extra mustard on the 80 that I will throw? It’s hard to project that out over everyone, but it seems likely that it would make some of them better.

· Pitchers who don’t really have 100 good pitches in them on a consistent basis, but who might have 60 good ones might find new life in the new system. We normally call these gents “failed starters” now.

· On the flip side, the guys who are good for 100 “pretty good” pitches every five nights, but whose stuff wouldn’t play up in shorter bursts… not so much. We generally call these guys “innings eaters.” We’ve taken away a lot of the innings that will need eaten.

· Freed from the chains of the model that basically relies on starters to (most of the time) pitch six or seven innings, teams might start experimenting with some alternate starter models, including piggy-back starters or four-man rotations or using a mix of “traditional starters” and swingmen. I’ve previously found that the main impediment to many of these models is that the current five-man, six- or seven-inning rotation model is the best at delivering bulk innings from a limited number of roster spaces.

· If teams decide that they only need four starters or that they can jettison the traditional “long man” and “sixth-inning guy” job descriptions, they can repurpose those spots into swing-men and piggy-backers. That could be fun. You could have different teams using radically different pitching roster construction strategies, rather than having all 30 using the 5-6-7 model that they do now.

· But of course, there’s a nightmare scenario in all of this. Teams could also just use those extra roster spots on more ROOGYs and LOOGYs. Especially since the goal of the exercise was to shorten the game, it would be depressing if the end result was simply that teams started playing matchup in the fourth inning.

But if I may point out that all of those strategies have one thing in common. Whether it’s starters throwing a bit harder over 60 or 80 pitches because they don’t need to throw 100 or just having more platoon matchups, all of these favor pitchers. That means lower-scoring games, even accounting for the fact that there will be fewer innings. Part of the magic of the nine-inning game is that eventually, you have to let the replacement-level guy pitch.

Hitters would be affected too. The days of grinding out 10 pitch at-bats in the hope of chasing the starter in the fifth and exposing the soft underbelly of the pen would be gone. You’d just end up with the guy who used to be the eighth-inning guy to face. While that sort of model was to blame for some of the 12-hour Yankee-Red Sox games of yesteryear, it would all but vanish.

It’s possible that teams might also use the extra roster spots to have a few extra bench bats and play a few more platoons. Given that teams are forever trying to find a decent second LOOGY even now, it might be easier to simply grab a couple more bats and play platoons as an antidote to teams using models where they are loaded up to try to grab the platoon advantage from the mound. Suddenly, guys with platoon splits are a little more useful.

But one other thing that might just happen in a seven-inning ecosystem is a drive toward more one-dimensional players, either all-glove or all-bat guys. Using 2015 data, let’s look at the percentage of plate appearances that are taken by each of the nine spots within the first nine innings, and within the first seven innings.

Lineup Spot

1st nine innings

1st seven inning




























Right now, in the nine-inning model the leadoff hitter takes about 25 percent more of his team’s plate appearances than does the nine-hole hitter. There’s an advantage to signing a good hitter, because he can hit high in the lineup and get more plate appearances. You can hide an all-glove guy in the nine spot using the same logic. The seven-inning model just makes this spread more extreme. (And if our theory is right that the changes in pitching structures would lead to a lower run-scoring environment, then the lineup will turn over even less and the advantage will be even more pronounced.) This likely wouldn’t be a seismic shift, but suddenly the evolutionary pressures would be a bit different.

Suddenly, we’re talking about a very different game. No one’s put a tree between home plate and the pitcher’s mound, but now we’re probably in a land where rosters are constructed differently, and pitchers are used in different ways. But on top of that, the very rise and fall of a baseball game would be different. Consider that from a leverage perspective, a baseball game would essentially start in the top of the third (of nine) innings tied, meaning that from the get-go, each at-bat is a little more important than it used to be. Plus, the leverage of an individual at-bat increases in a lower-scoring environment, which we would likely see. And that means more incentive to do all of the time wasting glorious chess maneuvers that make baseball so long fun.

But there’s one little wild card that might mess everything up. There would—almost by mathematical necessity—be a lot more extra-inning games. Consider that in 2015, 8.7 percent of games went into extra frames. However, the two teams were tied 12.8 percent of the time after seven innings. Even if one team is notably better than the other (in a true-talent sense), if there are fewer chances for them to separate themselves, then they will be tied more often at the end of the chances that they have. (In 2015, after one inning, the game was tied 52.3 percent of the time!) And in a lower run-scoring environment, the chances of extra innings go even higher! Managers might think twice about going crazy with bullpen moves and teams might think twice about carrying only nine pitchers.

There are several archetypal baseball games. Each has its own emotional cadence. There are slugfests. There are old-fashioned pitchers’ duels. There are back-and-forth games. There are 11-3 snoozers. There are games that take two hours to fully blossom and games where the important part comes in the fifth inning when half the crowd is buying some peanuts and Cracker Jack. There would still be dramatic comebacks, although they might look a little different than they used to. None of these would fully disappear, but they would appear in different proportions.

I’m not entirely sure that the games would actually be notably shorter. It’s a matter of aesthetics as to whether this sort of mix of cadences would be better or worse than what we have now. But by now, reader, I hope you see that by making a small change, we’ve altered the game considerably. (And there would be even more ramifications than this, ones that we can’t even predict.)

How different would baseball be if it were different? Very different.

The Lure of the Simple Solution
We can see that—at least with a little bit of systemic thinking and #GoryMath—what started out as a simple solution to shorten games has had ended up in a very different place than what we might have expected. Baseball is an intricate system, one that defies simple solutions. Now, you could make the case that the new baseball reality might be better or worse, but you couldn’t make the case that it would be largely unchanged.

If I may reveal myself a little bit. This exercise was partly just a fun baseball thought experiment for me, but I have some ulterior motives. Writing it was goaded along by the recent political conventions that you may have heard a thing or two about. It’s the political season, and there is no shortage of people offering easy-sounding solutions to all the tough problems out there. Frankly, I don’t care whether your political leanings are to the left, right, center, or you’re just part of the Official Monster Raving Loony Party of the USA. Beware of solutions that don’t go beyond the surface and delve deeper into systemic thinking. There will be plenty of them, from all parts of the political spectrum, on offer. You will end up with a complicated mess once you look into it, but life is messy. If nothing else during this political season, remember that.

But if you don’t want to get political after all this, then perhaps just appreciate once again the gorgeous and sometimes maddening complexity of the game of baseball.

Thank you for reading

This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.

Subscribe now
You need to be logged in to comment. Login or Subscribe
Roster sizes would shrink or change. Why do you need 11-12 pitchers to pitch 7 innings? A team could go to carrying more position players. Baseball could reduce the roster size by two also.
Hell will freeze over before the MLBPA agrees to a roster size reduction.
Four inning starters, multiple R/LOOGYs in each team's bullpen, MORE pitching changes, longer innings, longer times between half-innings for those lost commercial breaks... And those are just the things right off the top of my head - not a good idea.
One other reason offense might go down is you'll see more one run strategies such as sac bunting. Imagine each team having a pinch bunter!