I was worried about this. Early this year Tyler Thornburg made a few multi-inning appearances and I got my hopes up but it looks increasingly likely that he is going to assume a traditional* bullpen role sometime soon.** This opinion comes with the usual caveats that I hate the concept of  closers and assigning relievers  to specific innings, but this goes beyond general hatred of defined roles.

*”Traditional” in this instance meaning “circa 1989″. 

**I did like Ron using him specifically against McCutchen the other day in a high-leverage situation. It didn’t work out because McCutchen is simply a great hitter, but it was the right call at the time.

tyler-thornburg-baseball-headshot-photo Tyler Thornburg currently still has that fun and super-useful ability among relief pitchers to throw multiple innings, and thus far he’s been absolutely dominant at doing so. He has, in fact been dominant in general, sporting a .82 ERA and a 1.17 WHIP. While many relief pitchers do benefit (in terms of effectiveness) from only pitching one inning at a time, there is good evidence that using Thornburg in this fashion is wasting his value.

Most relievers benefit from the knowledge that they will only go one inning and can let loose with their best stuff, knowing they’ll be out of the game in short order. Because we got to see Thornburg start a good amount last season we can see some evidence that there is more to his dominance. Almost every pitcher gets worse each time he cycles through a lineup. Tyler Thornburg, in his young career, has been dominant the first time he’s faced anyone and absolutely ordinary after that. Let’s compare him to Marco Estrada just for kicks using BA/OBP/SLG/OPS

As a starter:



1st time

2nd time

3rd time

Tyler Thornburg




Marco Estrada





Marco Estrada has always been a wizard the first time he faces anyone and Thornburg has much been better. The ability to hold nine batters to a .560 ERA is extremely valuable if used properly.  Using Thornburg as a single-inning puts a severe artificial cap on the number of batters he will face.

Last season for all relief pitcher with over 25 saves Kenley Jansen pitched the most innings with 76.2. Next highest was Steve Cishek at 69.2. This is the fundamental problem with closers/one-inning relievers. Tyler Thornburg has shown that he can be dominant over two innings (or more if he’s efficient enough). If he were to throw something like 2 innings every 3 days he would end up throwing something like 130 innings, or about double what you see from your typical closer. I think you can make an argument that the age of one-inning relievers has led to more dominant bullpens in general, however if you have a pitcher who can be dominant over more than one inning there is no reason to force him to throw less.

Looking to prove my point I went to the baseball-reference.com play index and searched for pitchers who worked in relief 80% of the time or more and threw 130 innings or more, which led me to what I think is probably (by bWAR) the most valuable relief season in major league history. It is also worth mentioning as current Brewer closer Francisco Rodriguez just passed this person on the all-time saves list.

In 1975 Goose Gossage pitched 141.2 innings for the Chicago White Sox. He finished the season with a 1.84 ERA, a 1.193 WHIP, 130 Ks and 70 BBs. His FIP* was 2.62, but his true results were extremely good. His ERA+ was 212 and his bWAR was 8.2 . Yes, in 1975 Goose Gossage, a relief pitcher, was basically Mike Trout.

*from a cursory glance at his numbers it looks like Gossage’s ERA was typically better than his FIP. If you want to use fWAR instead he had 4.2 in 1975, 2.3 in 1976, and 4.2 in 1977, but starting in 1974 (which is the first time he was really really good) his ERA was lower than his FIP in 15 of 20 seasons. Make of that what you will.

Goose-Gossage The following season the White Sox turned him into a starter where he was much less successful posting a 3.94 ERA over 224 innings and 2.9 bWAR. The following season Gossage joined the Pirates who put him back in the pen where he was once again dominant, throwing 133 innings with a 1.62 ERA, 135 strikeouts and 90 BBs, good for 6.0 bWAR.*

*In his 12-year career K-Rod has produced 21.5 bWAR, which is pretty good. Gossage produced 35.8 over an eight year period from 1975 to 1983.

Comparing Tyler Thornburg to Rich Gossage isn’t fair to either of them as Gossage is a Hall of Famer and Thornburg is just barely not a rookie, but the point is not that Thornburg can be the next Gossage, the point is that if you have a player who can be dominant over multiple innings, you piss away a lot of his value by not using him over multiple innings.

I would argue there are ancillary benefits to having a player pitch more innings less frequently. I suspect but cannot prove that it causes less strain on your arm. It will serve to spell two of your other pitchers on a regular basis, which should also help overall bullpen health.

Finally, if you have a guy throwing longer outings every three days, it should be easier to transition him into  a starter role if you need to. If Thornburg were to become the closer they would need to stretch him out again to start, and it’s worth keeping in mind that several of the Brewer starters have a checkered injury past. I would quite frankly be surprised if Tyler Thornburg does not make at least a few starts this year out of necessity.

Mariano Rivera is by almost all accounts the greatest closer in major league history. The most bWAR he ever had in a season was 5.0 in 1996. That is basically the ceiling when you turn a player into a closer. If you have a guy capable of throwing 130+ innings of dominant relief and turn him into a closer, you’re almost definitely making your team worse.



Tags: ,

  • Ryan Topp

    This is really great, Paul. It is really frustrating that managers don’t seem willing to adjust away from the model that has been established since Eckersley and the late 80′s A’s put it in place. After all, they used to do things differently, but seem to have more or less abandoned the Gossage/Fingers model and have gone to this one.

    The problem is that we’re basically fighting against a thing that is the industry standard for setup men. To me, there seems to be two main reasons this is now the standard: managers want to be able to use these guys as often as possible from day to day and they also want to be able to play matchups in certain situations. That makes it harder for managers to let a guy who they like to use for setup face 6-9 consecutive batters on a consistent basis for a couple of obvious reasons.

    On the plus side, I think that once Kintzler returns they should be able to dedicate either him or Thornburg to perhaps a 4-6 out role in tie or slightly trailing situations because they’ll have KROD, Smith and Kintzler/Thornburg to pitch the 7/8/9th of games with the lead. It would be nice if they were more flexible than that, but again, this does seem to be the way the whole industry does things (more or less) now.

    • BadgerNoonan

      Thanks. And yeah, specialization obviously has a lot to do with it.

      Maybe when Kitnzler comes back that’ll happen. Probably just as likely they end up as the 7th and 8th inning guys in some form, which makes me sad.

  • akschaaf

    Didn’t Thornburg see a decent sized velocity drop last year as he got into the 4th-5th inning last year? I think that was even some of the reasons the prospect guys saw him as a bullpen guy coming up through the minors.

    Going back to back w/1 inning might have the same effect as going 2 innings and then taking two days off (or whatever you want to work it), but worth wondering about with him.

    If they’re going to go with a huge 7 man bullpen and all the options are good (well, maybe besides Wang), I don’t think its a huge issue if you’re unsure how Thornburg would handle it. The 1975 comparisons are fine, but that’s so dated. Starting Pitchers threw like 300 innings and 150 pitcher in a game were commonplace, there’s no way we’d want that now.

    It’s really weird, many NBA teams are doing all sorts of experimentation with their D-League teams, the Rockets basically have a policy on their DL team that guys only shoot shots inside 5 feet or 3 pointers, its crazy that no MLB minor league team do much different in terms of strategy or innovation.

    • BadgerNoonan

      That is nuts. The minors are the perfect place to try out different bullpen management. And yeah, it shouldn’t be a problem with a 7-man pen, and yet I’m sure it will be at some point. I mean, it’s dumb that K-Rod pitched 4 days in a row too. Obviously this is a losing battle and nothing will change and yada yada, but it’s still something you could do, and make life easier on everyone.

      • akschaaf

        One thing I think would be interesting, use your 5th starter as the occasional “long man.”

        If you have an off day, use that as a reason to skip the 5th starter’s start and have him available for long relief around the days where he would start. Maximize your best SP and save your bullpen some.

        Even though the Brewers are pretty equal 1-5, they could skip Marco’s spot and have him availble for multiple innings after a Wily start or something. Seems preferable when the back end is overworked or when they would be going to a subpar option like Figaro or Wang.