Baseball Therapy: The 26th Man

Negotiations for MLB’s collective bargaining agreement (CBA) are heating up, and one item reportedly being discussed is the possible addition of a 26th roster spot. It would be the first expansion of active rosters in a very long time and, if it happens, would bring up the question of what teams will do with that 26th spot.

OK fine, 27 of the 30 new guys will be relievers, right? And 20 of them will be left-handed. Dave Cameron of FanGraphs goes so far as to suggest that MLB have a rule that it can’t be a pitcher. But is that the best use of the spot? BP alumnus Michael Baumann floated a few suggestions for the spot that don’t rhyme with “riddle believer.”

Let’s see if we can pull apart what teams are likely to do, and maybe even what they should do.

Warning! Gory Mathematical Details Ahead!

Let’s start with something important. The 26th man on the roster is someone who is currently, by definition, a replacement-level player. He’s a fringy guy not good enough or important enough to make the roster currently. Otherwise, he’d be on the roster. If teams add a reliever, it’s going to be the guy who was seemingly always on the AAA/MLB shuttle last year.

But let’s start with the obvious natural experiment that gives us an idea of what teams might actually do with the extra spot. September roster expansion has (up until now) come with the promise that teams could–if they wanted to–use all 40 players on their 40-man roster. This isn’t a perfect analog, because teams have different incentives in September. Some teams are out of contention and spend the month getting their younger players a few extra reps at game speed because … well, what else are they gonna do?

There’s also the issue that this is not an increase of one roster spot, but in theory, of 15. Still, perhaps they can offer some insight. Using data from 2015 (the 2016 Retrosheet file isn’t quite ready yet), we can see that in August, teams use an average of 4.06 pitchers (including the starter) per game. In September (and early October), that increased to 4.62 (a difference of 0.56).

For games played in AL parks, the number of pinch-hitters and pinch-runners used jumps from 0.68 in August to 1.17 in September/October, for a difference of 0.49. (In NL parks, the corresponding numbers are 1.70 and 2.25, for a difference of 0.55). In terms of in-game tactical substitutions, when given a buffet of options, managers actually find just as many new ways to use position players (and I’m not able to easily identify players inserted entirely for defense) as they do new ways to use pitchers.

From this data, we don’t know whether managers view those extra moves differently. For example, the pitcher moves might have been ones that they had been dying to make all year, but were hamstrung by the lack of roster space, and the batter moves might have been low-impact, but hey, every little bit helps. Or maybe that’s reversed. But let’s assume for a moment that a team decided to spend that extra roster spot on a reliever, just like we figure. Again, it would be a fringy reliever.

Most teams carry seven relievers now, so I looked at the pitchers who threw the 211th- through 240th-most innings in relief in 2016. They averaged a contribution of 0.1 WARP and had an average DRA of 4.52. We’re talking about a below-average reliever at best. And for all the talk of teams having their starters go fewer innings, the fact that teams have carried such expansive bullpens means that even if the workload is greater than it used to be, many hands make lighter work. In 2015, fewer than a quarter (22.6 percent) of relief appearances were after having pitched on the previous day. Another quarter (27.8 percent) were after one day of rest.

Yes, an extra arm might be able to soak up some of the innings and maybe give someone a bit of rest, but since this would be the worst reliever in the pen, they would have to be innings that are currently being soaked up by the worst guy in the pen now, who would then be more free to soak up an inning here or there rather than the fourth- or fifth-best guy in the pen might be getting now and saving him from having to go back-to-back once in a while. Or y’know, it could just be another lefty.

Using 2015 data, we find that in close games (a team is either ahead by three runs or fewer, tied, or down one as they pitch), teams aren’t always able to claim the platoon matchup. Here are the numbers, by inning, of how often the pitching team had the platoon advantage (relievers only):

Teams appear more likely to play matchup in the sixth through eighth innings but more than two-thirds of lefties were facing a right-handed hitter on the hill in these close games during those innings. In 2015, there were 7,741 plate appearances in which a left-handed hitter came to bat against a reliever in the sixth through eighth innings, or an average of about 258 such at-bats per team per season, and roughly 180 that weren’t covered by a lefty. Seeing that most teams already carry one or two lefties, it suggests that another one might be able to handle another 60 or so of those, plus probably throw some garbage innings.

I’ve previously shown that the effect of being able to match up against one more lefty hitter in close games is around a quarter of a win, if you pro-rate it out over a whole season. But, that was done with the assumption that it was a league-average lefty doing the lifting. Remember that this 26th man/third LOOGY would be a below-average lefty reliever. Sure, he throws from the port side, but if he’s not any good, there comes a time when the starboard side guy who’s actually decent might be a better option. Even an extra lefty isn’t going to get you much.

So, if it’s not going to be an extra reliever, then perhaps an extra position player?

A designated pinch-runner sounds fun, but a few years ago BP alumnus Sam Miller ran a thought experiment in which he gave the 2012 Reds Billy Hamilton (OK, that part was real), but also endowed then-Reds manager Dusty Baker with the power to magically intuit when the highest-leverage point in the game would be and allowed him to break the rules of substitution by pinch-running Hamilton and then re-inserting the pinch-run-for player back into the game. (So if Hamilton were pinch-running for Joey Votto, the Reds didn’t lose Votto’s bat.)

He figured that it would be worth about 5-6 runs over the course of a full season. But that breaks a lot of rules. Maybe teams do have their own Hamilton (Terrence Gore, come on down!) who can’t hit or field, but the rules of baseball will still apply. The net effect would probably be less interesting.

A defensive whiz who would exist to play shortstop or center field in the eighth or ninth inning of games might also come in handy. I’ve estimated that a good defender replacing a bad one is an upgrade worth about .03 runs over the course of an inning, and if we figure that he plays about 100 innings, that’s still three runs per season. Not bad.

Another use of the extra spot might be the ability to have a platoon partner for a guy who really needs one. I’ve estimated that the existence of a well-constructed platoon might add 2-3 runs to a team’s overall performance over the course of a year.

If there’s something that’s worth saying about this 26th spot, it’s that it’s not going to be worth a lot. We’re talking about a marginal upgrade somewhere. Still, the options for using that spot on a position player seem better than the ones for using it on another pitcher. Maybe in certain cases, it makes sense for a team to go with a reliever, but I think a closer look at the evidence suggests that predictions of 30 new LOOGYs entering baseball are a bit premature.

Michael Baumann’s article had an intriguing suggestion. Teams should use their 26th roster spot for a “Team Dad.” Someone like David Ross. A veteran guy who’s amazing in the clubhouse and might be able to still hit .220 and hack it a couple days a week in the field. But he’s there mostly to be a dad. We still have no idea what this sort of team chemistry/veteran leadership piece brings to a team. I think it’s interesting to note that there are teams that have carried these “Team Dads” on the 25-man roster in the past, including Ross. I think this is actually the wild card suggestion. A guy with limited on-the-field usage potential is sometimes hard to roster, but now the roster just got a little easier to handle.

But another one of Baumann’s suggestions might be more interesting: the return of the swingman. Oddly enough, in a world that expects a 26th roster spot to produce a lot of new relievers, maybe the best bet is a semi-starter? In the past, I’ve looked into pitchers who appeared to be better suited to throwing 50 pitches than to throwing 100, and show average to above-average performance in those first 50, but not in the second 50, so they look like boring guys who aren’t really worth the investment.

We don’t really have a name for these three-inning warriors other than “failed starters,” but they used to be known as swingmen. They would occasionally start and occasionally relieve. The closest thing that we have nowadays is the idea of the “long reliever/spot starter” role, except we pretty much expect that he’s not going to be any good. The idea is that there are pitchers who can pitch in a three-inning role and actually be good at it. Having one (or two or three) of them around might allow us to do some interesting things with our starting rotation and our pitching rotation more generally.

I think they are out there, we just don’t realize it yet. In the past, I’ve estimated that if a team could take its fifth-starter spot, fill it with two of these 50-ish pitch guys, and ask each to go three innings, rather than one guy to go six, that could be worth a good chunk of a win over the course of a season.

I’d also recommend one other bit of strategy to go with it. Teams often carry five starters and use them in order, even if they don’t play five games in a row. What if they were to commit to pitching their good starters every fifth day, rather than every fifth game? It would mean that the “fifth starter” might lose 10-15 starts over the course of the year. I’ve shown previously that pitchers don’t pitch better (or worse) with an “extra” day of rest, nor are they healthier because of it. When teams need a “fifth starter” they could have the two swingmen make the “start."

When they could be skipped, the swingmen could do things like give the third and fourth starter a half-day off. Or they could sop up garbage innings and maybe save the bullpen after a blowout day. This gives the manager some flexibility to strategically reduce the workload a bit. The problem with this strategy (and a lot of exotic starting rotation “strategies”) has always been that if you made the no. 5 spot into a tandem or tried to do piggy-back starters, it shortened the rest of the 25-man roster. Well, now it’s longer. Maybe it’s time to give something a little different a try.

In a few years, if teams were to commit to this strategy, they might organize both their development system and their offseason strategy to find them. "Swingman" might even enter our vocabulary the same way that “setup man” has. But since this role doesn’t exist now, it’s not likely that teams have thought much about whether they already have one of these on-hand (or have given up on the thought that they just need another year to turn him into a “real stater”) or where to get one.

In the short term, I don’t think this strategy would work. Because these swingmen don’t officially exist, the players who would be the 26th man/swingman would be fringy starters from last year. To show why it would be a problem, I took all pitchers from 2011-2015 who pitched at least 100 innings as starters, and divided them into five groups, based on their unadjusted RA9. The top 30 in each year were the "no. 1 starters” and the next 30 were the “no. 2 starters” and on down the line to no. 121 to infinity, which were the fifth starter/fringe group.

The swingman that a team is going to pull out of their magician’s hat for 2017 is going to be from this last group. For each group, I looked at the average number of runs that they gave up, paneled by inning. It’s pretty obvious which group is which, because I’ve specifically separated them out by RA9, so the “aces” give up the fewest runs.

Take a look at the shape of that graph. It’s pretty much the same for all five groups. It’s highest in the first inning, when the pitcher is facing the best hitters at the top of the lineup. It dips in the second, and jumps around a little. Look at the fourth to the fifth inning. Scoring goes down slightly for all groups, except the fifth starter/fringe group. This is where that group starts to lose it. The idea of turning this into a tandem makes sense here. Get that guy out of there before things (on average, anyway) turn ugly, rather than expect him to finish six innings.

Of course, notice too that those guys are also pretty awful. Maybe a couple of them play up better in shorter bursts, but it’s going to take a while for them to show themselves. But if those swingmen are out there, teams may want to have a longer-term strategy for how they might use the 26th spot to their benefit, rather than just adding another reliever and coasting.

Don’t Expect Too Much

The 26th man isn’t already on the roster for a good reason. There are things that a team can do with the spot that would benefit their chances, but the actual effect that the player would have is going to be minimal. In all candor, the point of the 26th man is mostly to create 30 new jobs for MLB union members.

I have some faith that the 30 teams won’t just grab a reliever off the shelf and roster him. That might be the best option, given the parts that they have on-hand, but assuming they have an “average” fringe LOOGY and an “average” speed-only guy and an “average” defensive whiz/no-hit guy, a lot of strategic plays that a team could make are remarkably similar in terms of value. I don’t think the reliever is always going to be the best bet.

I think we also need to consider the short-term effects of the change to a 26-man roster and the long-term strategic play separately. If teams have a slightly bigger canvas to work with, they might be able to do something a little more creative, and it may take them a little bit more time to create a masterpiece. So, even if we do end up with 27 new relievers on Opening Day of 2017, it’s still something to watch. Evolution takes time.