Why a Starter-by-Committee Approach May Work for the D-backs

If we weren’t already doing it this way, is this the way we would start? Before taking over as General Manager of the Los Angeles Dodgers, Paul DePodesta explained in a presentation (which, sadly, was taken down once he became GM) that like with many other endeavors, continually asking that question could help a team win. DePo credited “management efficiency” guru Peter Drucker with the question, and I think applying it to the D-backs’ rotation situation brings us to a smart — if radical — idea.

The idea? Do away with the starting pitcher as we know him, entirely. I understand that the change I’m espousing today would be difficult to implement, especially mid-season. There’s an intermediate step the D-backs could take right now, however: replacing just two spots in the rotation with teams. But first, let me try to win you over to my premise.

In baseball right now, there are two kinds of pitchers: starters and relievers. Starters tend to be better. They tend to have at least three pitches they can throw with confidence, even if they have difficulty throwing strikes with one of them. And if everything goes to plan, they tend to throw 180-220 innings in a season. Relievers, on the other hand, tend to rely primarily on two pitches, and they may pitch to strikeouts a bit more frequently. From matchup guys to multi-inning guys, they tend to throw between 50 and 80 innings in a season.

Are there really no pitchers in the middle, pitchers that could excel in 100-120 innings of work, even if they’d become exposed at 180 innings? I think there are some pitchers who would struggle in 6-7 inning starts because of a limited repertoire or because they rely on deception, but who could get more outs per season than they would if used as a classic reliever (about one inning every two or three games). I think there are some pitchers that are underutilized as regular relievers, but overutilized as full starters, and I think Josh Collmenter is one of those pitchers. I also think that diminishing the workload of most starting pitchers would get rid of more bad innings than good, and that if relievers are relievers because they shouldn’t see the same hitter twice, they can still face 9 hitters.

The Proposal

What if, instead of pitching either 5-8 innings every fifth game or 1-2 innings every second or third, all pitchers on a staff threw about 3 innings every third game? Picture an eleven-man staff (you could still have a closer if you really wanted). For Game 1, pitchers 1, 2, and 3 each pitch about three innings. Game 2, pitchers 4-6 would go. Game 3 would feature pitchers 7-9, and if any of those games went to extra innings or were especially stressful, or if any pitcher really struggled on his appointed day, you could go to the next guy right away. You could steer your best pitchers toward your opponent’s best hitters by deploying them first: the “starting” pitcher would tend to face the top of the order twice, more often than not.

Adopting this approach would mean that ten pitchers would get approximately 140 innings each, per season, a manageable number. Pitchers would pitch frequently, but they’d occasionally get three days rest (because of off days, or if a team got through three games in a row with just three pitchers). They’d also tend not to pitch much more than once through the batting order, and they’d only pitch against the same team in consecutive appearances very rarely.

I think the benefits of this approach outweigh the drawbacks, and I think the benefits outnumber the drawbacks.

The Benefits

1. The starting pitchers would be in a better position to succeed. Pitchers face a “Times Through the Order Penalty” (TTOP) that is significant, if not enormous. All other things being equal (i.e., if competition was the same), a pitcher would simply do better in innings 1-3 of two starts than in innings 1-6 of one start. We can speculate as to why: for a hitter to hit certain pitches that may move in unique ways, it’s helpful for a hitter to have seen (or been recently reminded of) what they look like. This passes the smell test for me. In the Science of Hitting, Ted Williams explained that he very rarely swung at the first pitch in a game (game, not at bat). Sure, he swung once in a blue moon — but that was to keep pitchers honest (seeing the first pitch is less helpful if it’s not a “real” pitch). The called strike was worth it, in his eyes, just to have that intelligence.  Stands to reason more intelligence would be worth something, at least to some hitters.

2. The starting pitchers would be less likely to fail. Pitchers generally have more good innings than bad, and so by having pitchers pitch until they fail, they’ll going to fail a higher proportion of the time. I hate suspending reality to go with pure numbers in a persuasion piece like this, but indulge me just for this paragraph. I call this “the Strasburg Effect,” after Strasburg’s 2012, when he pitched with a hard limit of 6 innings per start after surgery. Red Sox manager John Farrell‘s rule is that, no matter how early, a pitcher is done if he’s had to pitch out of three jams in one start.  I think we also know from experience that if a pitcher really runs into trouble in the fifth or sixth innings, he’s likely to get replaced in most games regardless of how well he did before that. But if you treat each inning as having a 2/3 chance of being good and a 1/3 chance of being bad, a pitcher who gets removed early more frequently for failing will have fewer chances for good innings. This is not a very important point, and it’s very minor, but it’s real.

3. The starting pitchers might pitch better. Forget the TTOP for a second. One reason why pitchers get pigeon-holed as relievers is that they don’t have a useful third pitch. If you’re a “starter” because you’ve got three or four, often you don’t emphasize one or two of your pitches in early plate appearances, so you have something new to show hitters the next time around. If you’re really only facing the order once, though, pitchers can use their full repertoire right away, and that’s worth something — there are plenty of stories of Eric Gagne types who dominated as relievers with a starter’s repertoire. And what about velocity? If you’re only pitching in shorter stints, you might pitch harder, like we’re seeing with Randall Delgado‘s velocity spike since moving over to the bullpen, where he’s had success. And while a pitcher’s comfort level is hard to quantify, allowing oneself to let loose could also lead to a benefit; supposedly Trevor Cahill has stopped worrying and just pitched since moving over to the bullpen, and that’s worked out well, too.

4. The team could get more out of its best relievers. Again, if relief pitchers are relief pitchers for repertoire reasons, that should only affect the number of times they can successfully pitch through a batting order. Well, the average number of batters faced in an inning is about four, and a batting order has nine players in it; there’s really no reason why a relief pitcher couldn’t pitch two innings instead of one, especially if he was pitching less frequently. And as for the TTOP — there is a smallish spike the second time through the order, but another spike from the second time to the third. None of the pitchers in a starter-by-committee rotation would pitch a third time through the order, and so even if a relief pitcher might sometimes pitch to some hitters twice, there isn’t much to worry about. Simply put, most relief pitchers are underutilized, and utilizing the good ones more fully is a significant benefit.

5. Relief pitchers may benefit from pitching less frequently and at more predictable times. This is another thing that’s hard to quantify, but I could see how there would be a benefit to this, and I can’t see how it would be a drawback. Instead of trying to convince you, I’ll let Bryan Price of the Reds do that. Price plans to go “against the current thinking” by playing matchups much less with his bullpen arms, in part because playing matchups inevitably means getting relievers to warm up more than necessary. Excessive warmups have a cost, according to Price, a former pitcher and former pitching coach. Well, in the starter-by-committee system, that problem of taxing relievers unnecessarily is solved: everyone knows who the next pitcher out will be, or, at the very least, matchups will be used a lot less.

6. There are more pitchers who might only excel in these circumstances. Grant, for a moment, that a pitcher like Bronson Arroyo has real value. He may not be above average, but he excels in some ways because he’s above replacement level, and he doesn’t cause other pitching staff problems by being alternatively great or mediocre, or by really struggling with pitch counts. I think there are some analogous pitchers that would excel similarly in starter-by-committee. It may be for mostly TTOP reasons, but I’d put Josh Collmenter in this category. It’s tough to fully use a pitcher like Collmenter in the normal starter/reliever setup, and I think there are others out there (Delgado), or in-between prospects (Carlos Martinez and Marcus Stroman), who will just have more value to a starter-by-committee staff.

7. Pitchers can be fixed more easily in side sessions. Starting pitchers have side sessions in which they can fix things if they’re broken (mechanics, feel, etc.). Relievers don’t. The fact that it’s hard for relievers to fix themselves outside of game settings is one big reason why young bullpens can be more effective than veteran bullpens just by virtue of being young. The key is flexibility, and I think teams leave a lot of potential success untapped by not having the flexibility to take a relief pitcher aside (and out of the bullpen for a few days) and have a coach help them fix what’s broken. In starter-by-committee, this is no longer a problem. It’s one of the big reasons why you’d have ten or eleven pitchers, instead of nine; you can simply pull one pitcher out for a turn to help fix him, and the rest of the staff can close ranks. Think of the enormous change for David Hernandez last season after he was finally optioned to get fixed. In starter-by-committee, Hernandez could have been pulled for a little while to get fixed. And veteran relievers can’t be optioned for fixing the way Hernandez was last season. Do not underestimate this benefit to starter-by-committee.

8. Smaller pitching staffs means better pitching staffs. This one is pretty simple. If you’d normally go with a twelve- or thirteen-man pitching staff and you can instead go with an eleven man staff, you’re not going to drop your best pitcher; you’re going to drop your worst one. Going from some innings to no innings for your worst one or two pitchers is a pretty immense benefit, is it not? That should be pretty self-explanatory.

9. Smaller pitching staffs means more position players. An extra position player means more available pinch hitters, and more flexibility. Having an extra Eric Chavez type can be pretty damned helpful. The number of times one might use a very light hitting backup infielder decreases, which means a sizable upgrade at the plate, if just occasionally. Or the extra bench player could mean more pinch running appearances. There are only benefits to having an extra position player on the roster, no drawbacks.

10. Fewer plate appearances for pitchers means more plate appearances for professional hitters. It’s a good thing a starter-by-committee team would have an extra bench player, because they’ll probably have more occasions to use them. Let’s say you’re the home team, and it’s the bottom of the second inning. Your team has had trouble scoring runs. There are runners on first and third with one out, and the pitcher’s spot in the order comes up. A team with a regular staff will let the pitcher hit, and he’ll have to swing away, as a bunt means the runner from third probably can’t run on contact. At the same time, a ground ball could turn into a double play. It would be nice, in that situation, to plug in a pinch hitter who has a better shot of hitting a deep fly ball, or getting a hit, or running fast enough to make a ground ball a fielder’s choice instead of a double play. Starter-by-committee means that professional hitters can be used to pinch hit for the pitcher in high-leverage plate appearances, even if they come early in the game. Like the previous benefit, it’s a significant upgrade that comes up only infrequently (you can’t pull a pitcher ahead of schedule all the time). But this benefit is magnified because of the high-leverage aspect of the appearance — and being 25% more likely to score a run in a particular situation, maybe once a week, can be worth an extra win per season. In addition to high-leverage appearances, a team would also get the happy coincidence at least occasionally that a pitcher’s spot comes up in the bottom of the third or top of the fourth, etc., a time when a new pitcher is about to come in anyway. The team would have the extra position player for the extra pinch hitting appearance, which is a real advantage.

11. Starter-by-committee costs less money. Josh Collmenter is making $925k this season in what was going to be his first year of arbitration eligibility. He’ll play for $1.825M next season, and the team has an option for the same amount the year after that. The following season (2017), Collmenter could be controlled with a $2M or $2.25M option in what would have been a free agent season. That’s ridiculously cheap, and buying more players like him would also be cheap. The fact that young starters might also get lower arbitration salaries also inures to the team’s benefit (if the arb system is struggling with the value of saves, it’ll struggle with the value of starter-by-committee types, but in the opposite direction). And bench position players are often cheaper than veteran relievers. In short, starter-by-committee would depress the salaries of young players, and at the same time allow the team to sign undervalued pitchers in free agency. If all teams used starter-by-committee, these financial benefits would diminish. The first team to go starter-by-committee will save tons of money on pitchers, money which can be used on position players. And that might be the biggest benefit of all.

That’s a lot of benefits. I think they outweigh the drawbacks, but there are drawbacks, too. Let’s look at them, and you can decide for yourself if starter-by-committee makes sense to you.

The Drawbacks

1. Some pitchers could be underutilized. No one would ever put Clayton Kershaw in a starter-by-committee situation. The list of pitchers on that list is not short. But to me, that dividing line might be right around Wade Miley. Who is more likely to have success: Miley during his third time through the batting order, or the other pitcher who would replace him? I think it’s a close call with Miley (and we’ll return to that below). But the point is: it’s a drawback that some elite pitchers wouldn’t fit into a starter-by-committee staff, because the 40-50 innings they’d miss that way would be better than the innings thrown by replacements. There’s a way to make starter-by-committee work with just one full starter, but not with two. And if you think of an elite prospect like Archie Bradley, you wouldn’t want to cap his potential (although a starter-by-committee backdrop would be a potentially effective backdrop for increasing innings workload season to season).

2. Pitchers are not trained to pitch every three days. In a starter-by-committee staff similar to the one I’m proposing, pitchers would throw with two days rest most of the time (and three days rest the rest of the time). Starters typically pitch on four days rest. Yes, there’d be less of a reason to throw side sessions (although if one is ever needed, the pitcher can get temporarily pulled from the committee). But it’s just not for me to say how big a transition this would be. It’s not really much different from, say, what Randall Delgado is expected to be ready to do right now, except that appearances become more foreseeable. But I do understand that transitioning to starter-by-committee is a significant adjustment, one that could even affect performance. I don’t think this concern means flushing the entire proposal, but I want to be up front that I don’t know how to quantify this.

3. It could be tougher to recruit good pitchers in free agency. This is another one that I find hard to quantify. If you were Matt Garza this last offseason, would pitching for a starter-by-committee team be attractive? Maybe not so much. You certainly wouldn’t sign at a discount to do that, and no team, even with a lot of homegrown talent, can afford to pay enough Matt Garzas to fill out an entire staff. But quality and dollars go hand in hand, and even if the starter-by-committee team ends up getting lesser talents, it’ll pay lesser money (and can spend the money on other parts of the team). I think pitchers will start to see their peripherals look better in starter-by-committee, and a team with such a staff may end up signing a lot of “pillow” contracts with pitchers coming back from injury. There also may be two or three relief pitchers every offseason that are looking to become starters, and see pitching for a starter-by-committee team as a step in that direction. A starter-by-committee team doesn’t need a dozen free agent pitchers to buy in every year — they just need a handful. And the pitchers best-suited for a starter-by-committee approach could only get paid like true relief pitchers by other teams.

4. Harder to play matchups in late innings. I think matchups are overrated in general (for the Bryan Price reasons above), and many managers who think they’re doing them well are not matching guys up as productively as they think. But pinch hits in late innings are important. Benefit #10 above refers to pinch hitting appearances in the early and middle innings, and there will be more pinch hitters at those times. But there may be fewer pinch hitting appearances in late innings for a team that switches to starter-by-committee, because if you wouldn’t be able to use four pitchers in every game, and because the pitcher’s spot may come up in the top of the eighth after the current pitcher has pitched just one inning. I think this drawback is definitely smaller than Benefit #10, but it still exists. As a partial fix, a team could still use a closer in a more traditional relief role, someone who only pitches in short stints as part of another pitcher’s shift, but who might end up pitching in back to back days. I think closers are used strangely, but a competent reliever in a flexible role like that would help with late-game batting situations.

Another issue: getting a team to buy in. Even if a GM were to want to try this novel approach, he’d need the support of his manager and a significant number of other people. The system would need to be implemented in the minors. But those aren’t drawbacks; they’re impediments. We started this exercise with one question: if we weren’t already doing it this way, is this the way we would start? The benefits and drawbacks of starter by committee should be evaluated independently of those start-up issues. It might be that starter-by-committee is better, but only slightly better, and not enough to justify all of those start-up costs. It’s just that that is a separate question.

How the D-backs can Implement Starter-by-Committee

I’m not suggesting that the D-backs do a total tear down right now; that’s just not realistic, although I would love it if the D-backs tried something like this down the road. As a mid-market team, the D-backs can’t play the same game as the Red Sox and Yankees and Dodgers and consistently win. They don’t need to play Athletics baseball, either, although a bigger dose of that wouldn’t hurt. But they do need to stop trying to swim upstream, and to start trying to get more edges wherever they can be found. It may be that Kevin Towers is not the guy to do this, considering he’s how we ended up in this mess. But, at the same time, he’s done innovative things before, if on a smaller scale.

The D-backs have an enormous opportunity right now to make a half-measure. Making two rotation spots “starter-by-committee” spots would reap many of the benefits of Benefits 1, 3, 4, and 6, while tapping into Benefits 10 and 11, as well. The team has Collmenter and Delgado, two pitchers who seem perfect for this kind of role. It has already pulled Trevor Cahill from the rotation, who will be underutilized if used as a normal reliever. And while I don’t know exactly what to make of Michael Bolsinger right now, I’ll throw him in the same category, too. What if these four pitchers split two starts per turn? The result would be something like:

Game 1: Bronson Arroyo (Cahill, Bolsinger available for short relief)

Game 2: Brandon McCarthy (Delgado, Collmenter available for short relief)

Game 3: Trevor Cahill (4 inning cap), Mike Bolsinger (4 inning cap) (Delgado, Collmenter available for short relief)

Game 4: Wade Miley

Game 5: Randall Delgado (4 inning cap), Josh Collmenter (4 inning cap) (Cahill, Bolsinger available for short relief)

Two doesn’t go into five, and so in devoting two turns in the rotation to split starters, they’ll only be one start in between on one end, and two on the other. Right now, the rotation is lined up Arroyo/McCarthy/Miley/Bolsinger/Collmenter, and of the three pitchers that would remain as “true” starters, Wade Miley is probably the best bet to stay in the middle of the two committee days, as the least likely to require a bail-out from long relievers (even if he did, short relievers would not be as needed for Games 3 and 5 above, so that would end up being fine).

Game 2 would normally be the side-session day for Collmenter and Delgado; same goes for Game 5 for Cahill and Bolsinger. If they’re not used after one day rest, they could be used after two days rest. After that, the opportunity would be gone. I think history has taught us, though, that a pitcher like Collmenter might do best by pitching in two appearances than in pitching the same number of innings in just one — and as someone pointed out on Twitter after I wrote that, giving opposing hitters a different “look” might be maximally effective if Collmenter is a change of pace from a pitcher who’s just pitched before him, in addition to forming a contrast with the pitcher who relieves him.

I like this approach, and I think it’s a reasonable shift, given the circumstances. If Mike Bolsinger doesn’t cut it, there are other pitchers in Reno who could get plugged in for a similar role (if not Charles Brewer, even Will Harris might work out, considering he’s not vulnerable to lefties). What do you think?