Fun idea, but now you're giving 140+ innings to the worst pitchers on the staff and taking them away from your best SPs. What's more, you're losing the ability to leverage your better relievers for key situations.

Granted, if you built your roster on this principle, you could save a lot of money on your pitching staff by rostering a dozen Carlos Villanueva types and spending your payroll on awesome position players.

Here's a radical idea. Teach a lefty and a righty to play a position (and one of them to bat) and have them rotate out based on lefty-righty batter.

The only problem with that is that you are not allowed to shuffle in and out the pitchers an infinite amount of times per inning.

Anyone messes up or gets hurt and it's Darnell McDonald on the mound for the save!

The basic premise is very similar to Ken Funck's piece for Baseball Prospectus back in 2009. The problem with Cameron's proposal is that if you're committing 12 pitchers to particular days, you're left with no ability to hook a guy when he's flailing, and you're screwed if your game goes into extra innings.

The tandem-starter idea I floated in the Rockies thread solves that problem. Four tandems = 8 pitchers committed to particular days. If you carry 13 pitchers, that leaves 5 relievers to fill the remaining innings. It's a little tight with 12 pitchers. If the tandems can generally get through 7 innings per game, a 5-man pen provides plenty of arms to manage around bad outings, get situational advantages, and handle the occasional extra-inning game.

You can't avoid your tandem starters facing most batters twice, but if they are reasonably efficient, they shouldn't have to face hitters three times in a game, giving you some of the advantage Funck and Cameron seek. If a tandem starter isn't efficient, lift him after no more than 18 batters and go to the pen.

The short bench is obviously an issue, but if a couple of the tandem starters can hit enough for the occasional PH appearance and the reserves are versatile, a team could pull it off. A 26-man roster would make it much easier, though.

Fun idea, but now you're giving 140+ innings to the worst pitchers on the staff and taking them away from your best SPs. What's more, you're losing the ability to leverage your better relievers for key situations.

Or you could do this less drastically, and leave your best starters out of it.

Ex: The Mets could start Santana and Dickey regularly, and have three tandems of two starters for the other three games per week. That's only eight pitchers -- leaving four more to be a conventional style bullpen.

there is a right way to do this, but the excerpt is not it. as pointed out in 1, this proposal increases the usage of a team's weakest pitchers, and that is shifting the workload in the wrong direction.

the next great leap forward in pitcher usage will be to concentrate a team's innings in their top 6 pitchers, and trim the size of a pitching staff from 12 or 13 down to 9 or 10.

Seems to me the perfect answer to this scenario is 3 pitchers scheduled each day with 2 "floaters".

Anyway perhaps it is in the article but how do they address the NL having no DH? It seems three 4 man groups will still need a 13th man in the pen and that leaves a 4 man bench with one of those guys being a backup catcher. So that leaves 3 bats on the bench to deal with matchups, rest days, and niggling injuries. That isn't a lot now and I can't imagine the situation improving when you've blown through two PH before the 7th inning.

I think the best answer is getting 3 guys who are pretty good through 6 or 7 innings and then using the next 8 guys to cobble together a pretty good pen that will finish those three guys' starts and also start a game or two in the rotation depending on the schedule.

Fun idea, but now you're giving 140+ innings to the worst pitchers on the staff and taking them away from your best SPs. What's more, you're losing the ability to leverage your better relievers for key situations.

Makes a lot of sense for a team like the Royals that have a great bullpen, but a pretty suspect rotation with guys that have trouble going 5 innings at it is.

A really minor change that someone might actually have the guts to try would be to abandon the 5th starter and replace him with a tandem or something. Start your top 4 guys on a regular rotation and skip the 5th slot whenever there's an off day. Then when the 5th slot comes up you either use a tandem or three guys or Cameron's idea of four guys or whatever. This would work best if the team had a very good #1 starter who could be trusted not to get knocked out early the day after you have 4 pitchers go multiple innings each.

I'm not saying that this whole thing is a good idea or that it's practical, but couldn't you get around some of the issues of having a 12 man pitching staff of relievers if a handful of them were not out of options? You could have, say, 9 guys always on the roster and then rotate through the other 3 spots with a bunch of guys*. So Pitcher A pitches three innings and then gets sent down the next day because he's not going to pitch in the next 3 days, so Pitcher Z, who is fresh, gets called up. Then Pitcher B pitches two and a third innings and gets sent down to make way for Pitcher X and so on.

Granted, you would need a lot of MLB-caliber relievers and some of the money you are saving will go to inflated travel costs, but this is all academic anyway.

EDIT: *You have to wait ten days after being sent down to call someone up, right? Then you're going to need a lot of relievers.

A really minor change that someone might actually have the guts to try would be to abandon the 5th starter and replace him with a tandem or something. Start your top 4 guys on a regular rotation and skip the 5th slot whenever there's an off day. Then when the 5th slot comes up you either use a tandem or three guys or Cameron's idea of four guys or whatever.

Isn't this essentially what teams did in the late 80s? I seem to remember the Royals doing this at least with Bud Black, later Steve Farr and I think Ted Power being the "swingman" who would start occasionally but long-relieve when a fifth starter wasn't needed.

I'm not saying that this whole thing is a good idea or that it's practical, but couldn't you get around some of the issues of having a 12 man pitching staff of relievers if a handful of them were not out of options? You could have, say, 9 guys always on the roster and then rotate through the other 3 spots with a bunch of guys*.

Your AAA team would be kinda screwed.

7: I don't see any way that any manager in the foreseeable future would be fine with a 9 or 10 man pitching staff.

12: The Braves continued that practice to some degree well into the 90s.

I've been pointing out for years (I' sure a lot of other people have, too) that the vast majority of starters are worse the third time through the order than just about any reliever who's good enough to stick in the majors, so having starters face the lineup for a third time should nearly always be avoided.

At the same time, most teams are able to have at least one or two starters for whom that's not true. Why not a 12-man pitching staff that uses 8 pitchers as starters/tandem starters and 4 as dedicated relievers? Some teams, those lucky enough to have 3 good starters, could use 7 pitchers as starters/tandem starters and 5 as full-time relievers. That's enough relievers to hold the fort when "starters" have bad outings.

So take, for example, the Orioles, the team I've been following most closely this year (because my Mets-fan girlfriend took it into her head to start watching every Orioles game for some reason!):

Full-time starters, handled like starters are handled normally:
Hammel
Chen

Three tandems:
Matusz/Eveland
Arrieta/Hunter
Berken/Patton (yes, it's a stretch)

Four relievers:
Johnson
Stropp
Ayala
O'Day

If the starting member of a tandem allows 3.5 out of 9 batters to reach base, he can get 11 outs in two trips through the order--he can be allowed to go far enough into a third trip to finish 4 innings. Then the second member does the same thing, but you can afford to have a shorter leash with him, since you do have four relievers to work with. It seems like that could work well. The only problem is that the "win" rule needs to be changed so that the starter only has to pitch 3 innings instead of 5.

If he's going really well, the tandem starter will sometimes make it through 5 innings (and be allowed to do so), which would save the bullpen a little.

How about a modification to the tandem proposal:

since the starters throw on their off days, why not use the starters as the second half of the tandem?
hear me out:

sp1/sp4
sp2/sp5
sp3/sp1
sp4/sp2
sp5/sp3

The first pitcher will throw 3-4 innings, the second pitcher, on his side day, will throw 2-3 innings. now you've gotten 5-7 innings out of your "starters" and you can still have a full bullpen to finish out the rest of the game?

this also solves the win "problem" because while the actual starter will not throw enough innings to get the win, chances are they'll get the win on their side start.

Isn't this essentially what teams did in the late 80s? I seem to remember the Royals doing this at least with Bud Black, later Steve Farr and I think Ted Power being the "swingman" who would start occasionally but long-relieve when a fifth starter wasn't needed.

I wasn't talking about using the 5th starter as a swingman, I was talking about implementing one of the ideas for having tandem pitchers or a triad or starting a reliever and pulling him after 10 batters or whatever, but only for the fifth spot in the rotation and only when you can't skip that spot after an off day. This would allow a manager to experiment with an unusual pitching arrangement rather than committing the whole staff to it, and is close enough to the swingman concept that the press wouldn't go ape #### over it.

this also solves the win "problem" because while the actual starter will not throw enough innings to get the win, chances are they'll get the win on their side start.

It doesn't solve it, though it can mitigate it. SP1 gives a good start and the team goes on to win, then SP3 gives a shitty start and puts SP1 in an unwinnable game. If that happens often enough, you'll have a problem.

I do think the idea would be interesting to try. I wonder why minor league clubs don't test drive this, or other concepts.

Your AAA team would be kinda screwed.

Sure, but those concerns are secondary to the MLB team.

Sure, but those concerns are secondary to the MLB team.

Yea, but they still need to play those games. I suppose if maybe you had a couple of 33 year old starting pitchers in AAA, you could tell them "hey everyone once in awhile we need you to go 7-8 innings, no matter how crappy you are pitching." Or have a knuckler down there. Or teach a knuckler to some position players.

The basic premise is very similar to Ken Funck's piece for Baseball Prospectus back in 2009

This basic premise is very similar to Chuck Tanner ca. 1979. Well, really, the whole thing. He suggested 3 pitchers for 3 innings each. He might have even tried it for a game or two.

Yea, but they still need to play those games.

Of course, but a MLB team doesn't care what the record of their AAA team is or how non-prospects are doing. All they care about is that the AAA team is providing fresh arms for the pool of pitchers and some position players for injury replacements. The team could win 40 games or it could win 70 and it makes no difference to them.

Again, this is all academic anyway, so those considerations are secondary.

Didn't the book have a chapter in this? I think they decided on something like a rotation of

Day 1: Starter 1
Day 2: starter 4/5 (4 getting more innings than 5)
Day 3: Starter 2
Day 4: Starter 3
Day 5: Starter 5/4 (5 getting more innings than 4)

Don't know if that's exactly right but it was something like that.

you are not allowed to shuffle in and out the pitchers an infinite amount of times per inning

This is a rule that exists for no reason, hampers strategies, and takes fun out of the game. So there's no hope of ever revising it, I would imagine :)