How MLB teams use their bullpens

[Paradigm]

Do you think the Royals success might influence the way teams construct their bullpens going forward? 
 
Here are a few thoughts. I look forward to reading what SoSH has to say.

• The Royals, with Herrera, Davis, and Holland have changed the demands from their starters. They need starters to get through 6 innings and keep the game close, then turn the ball over to an elite bullpen that can keep the game tied, hold a lead, and/or let the offense deliver more runs. Last night, Ventura was able to go 5.2 and keep the game tied before handing the ball to Herrera, who shut down the Giants offense.
• JJ Cooper at Baseball America keeps a running list of minor league pitchers who touched 100 MPH this year. Here's the list for your reference. I'm speculating, but I'll bet some of these pitchers are plus velocity guys without elite secondary and tertiary pitches, who may not have the command to go 6-7 innings. (Herrera is a great example of this -- he can be really wild, you could never see him going 6 innings.) 
• Herrera delivered 1.4 WAR to the Royals this year. Wade Davis delivered 3.1 WAR. These are massive numbers -- Herrera is on the Jered Weaver level, and Davis is in the Sonny Gray/Hisashi Iwakuma level.
• This year, the average SP went 5.53 IP per start. You have to do something with those innings to keep your team in the game. You can't just hand them to a soft-tossing guy with borderline stuff and hope that he can get through a big league offense. 
• And, of course, the most interesting debate: the notion of building a shutdown bullpen and intentionally shifting prospects away from a starting role -- where, if they put it all together, they will deliver the most value -- versus moving them to the bullpen, accelerating their arrival in the big leagues, and relying on velocity to build a better relief corps in the middle innings. Do you give every kid 3-4 seasons in the minors to see if he can start? Or, is the pressure to win now so great that you're better off calling these kids up and turning them into relievers? 

Looking at the Red Sox, this question applies so well to Rubby de la Rosa. He's been starting in the minors since 2009 (though he actually doesn't have many starts under his belt, only 58 in the minors and 28 in the majors. He's had Tommy John, he throws hard, he has enough pitches to start but falls into that average number of innings per start (5.43 IP per start this year). 
 
He's 26. You cringe at the thought of wasting his arm in the minors. But you also can't stomach letting him develop at the major league level when he hasn't proven that he can get the team through 6 innings. So do you try to groom him for a Herrera/Davis role, try to get 1.5-2.5 WAR out of him, and keep him healthy by giving him a defined role so that he knows when he's going to get up every game?
 
Bullpens.

 

[ehaz]

Looking at the Red Sox, this question applies so well to Rubby de la Rosa. He's been starting in the minors since 2009 (though he actually doesn't have many starts under his belt, only 58 in the minors and 28 in the majors. He's had Tommy John, he throws hard, he has enough pitches to start but falls into that average number of innings per start (5.43 IP per start this year). 

 
I think it applies even more to Matt Barnes because of his lack of a third pitch.  I agree with the overall sentiment, it's just tough to implement when relievers are so volatile.  Perhaps next season hitters figure out Davis, Herrera pulls a Dan Bard, and Holland blows out his elbow on April 2nd.

 

[Paradigm]

 
I agree with the overall sentiment, it's just tough to implement when relievers are so volatile. 

 
Is this still true when it applies to players with elite stuff? I can understand this kind of volatility when we were applying it to guys who threw 92 MPH, lacked a wipeout slider, and may be perceived to be good pitchers because of good luck when in reality they were giving up a lot of hard-hit balls, getting lucky with stranded runners, etc.
 
But if velocities are up, then relievers might be better pitchers overall, not just failed starters. Could there be less volatility then?

 

[geoduck no quahog]

This kind of compliments a discussion we had here a couple of years ago about getting rid of "starters" and throwing 4-5 pitchers into every game. If I recall, the valid argument was - how many times a year can you throw a pitcher into a game on short rest?
 
In this case, having a "Big 3" (tm) only works if their usage during the season blends in with the starters' performance. You can't use the same 3 guys every day (meaning you use the chaff), except in the post-season.
 
The other argument against is the salaries of middle relievers versus starters. As long as the Save pollutes baseball, middle inning guys are going to be underpaid as a group. Every pitcher wants to be a starter.

 

[Paradigm]

In this case, having a "Big 3" (tm) only works if their usage during the season blends in with the starters' performance. You can't use the same 3 guys every day (meaning you use the chaff), except in the post-season.
 

 
Sure. It's also not going to be necessary -- some days your starter will be able to go eight, other days he's going to get blown up in the third inning and the manager will just use the long-man to mop up the game. 
 
Let's look at the Cubs, another team led by a fairly progressive front office. The Cubs have a pretty formidable bullpen going into next year. They took Hector Rondon (26) in the Rule 5 draft last year and after a year of adjusting to relief, he turned in a very good year as their closer. They acquired Neil Ramirez (25) from the Rangers, immediately converted him from starter to reliever and he had an exceptional season. They acquired Justin Grimm (26) from the Rangers where he was starting and immediately shifted him to the bullpen. (Both pitchers were likely available because the Rangers didn't project them as starters). They also have Felix Doubront (27), Jacob Turner (23), and Dan Straily (25) on the team, all of whom could be options for the bullpen. That's six possible relievers, most with premium stuff and starting pitcher pedigrees, none over the age of 28, none requiring a FA contract. (Cringe at the thought of Joba Chamberlain's contract this offseason.)
 
Meanwhile, they let Kyle Hendricks start 13 games this year. His average fastball velocity was under 90, but he showed great command (1.68 BB/9). He put up 1.5 WAR in 80 IP. That's the kind of starter who might be able to get through two turns of the batting order and give up 2-3 runs, allowing your offense to do enough work to keep you in the game so that you can turn ball over to a flamethrowing bullpen.

 

[ehaz]

Meanwhile, they let Kyle Hendricks start 13 games this year. His average fastball velocity was under 90, but he showed great command (1.68 BB/9). He put up 1.5 WAR in 80 IP. That's the kind of starter who might be able to get through two turns of the batting order and give up 2-3 runs, allowing your offense to do enough work to keep you in the game so that you can turn ball over to a flamethrowing bullpen.

 
Perhaps part of the bullpen's success is due to the starting rotation.  Vargas/Guthrie/Shields aren't exactly known for their velocity and rely heavily on their off speed.  The change of pace after the 6th inning could be really devastating to hitters.
 
It does seem that KC is a step ahead when it comes to developing their own relievers.  Herrera and Holland never started games past A ball, Aaron Crow was converted pretty quickly and they were wise to pull the trigger on converting Davis after his year of suck.  Looks like they're doing the same with Finnegan.  
 
I don't know if their FO is just faster at identifying when a player is not cut out to start or whether they look for potential relievers through the draft. 

 

[ALiveH]

The 2013 Sox also had an awesome bullpen that shortened games in the postseason.  I don't consider it a failure at all if a Barnes or Ranaudo go to the bullpen and become dominant setup men - there's a lot of value in having 3+ great relievers on your roster.
 
I thought I read that velocity plays up better in the postseason because hitters have lots of nagging wear-and-tear injuries that slows bat speed?  That's why you see certain high-velocity "aces" become absolutely dominant & carry teams in the postseason.

 

[Just a bit outside]

The 2013 Sox also had an awesome bullpen that shortened games in the postseason.  I don't consider it a failure at all if a Barnes or Ranaudo go to the bullpen and become dominant setup men - there's a lot of value in having 3+ great relievers on your roster.
 
I thought I read that velocity plays up better in the postseason because hitters have lots of nagging wear-and-tear injuries that slows bat speed?  That's why you see certain high-velocity "aces" become absolutely dominant & carry teams in the postseason.

I agree.  The Red Sox currently have 6 starting pitching prospects from AAA to the majors between the ages of 24 and 26. I understand that starters have more value but I think it is time to move a couple or three of these guys and see if you can develop a dominant bullpen.  If you could develop two mid rotation starters and two dominant relievers out of RDLR, Webster, Ranaudo, Kelly, Barnes, and Workman I think it would be a great success. 

 

[phenweigh]

 
Perhaps part of the bullpen's success is due to the starting rotation.  Vargas/Guthrie/Shields aren't exactly known for their velocity and rely heavily on their off speed.  The change of pace after the 6th inning could be really devastating to hitters.

This is why I like the idea of Wright in the bullpen as a set-up guy.  Batters have been looking at 90+ for 2-3 at bats, then have to deal with a knuckler ... it has to be a big adjustment. 

 

[ALiveH]

knucklers tend to be so inconsistent & durable that I really think the wakefield role is perfect.  Make him a spot starter, long relief, mopup, extra innings & all-around emergency innings-eater guy.  I think there'd be tremendous value in having Wright on the MLB roster in such a role.
 
But b/c of their inconsistency, I wouldn't want to have to rely on a knuckler to be able to consistently get big outs in close games.

 

[Hee Sox Choi]

Just read some article about how the Yankees should sign Andrew Miller and David Robertson and go with Betances as the Relief Ace, Miller as the 8th inning guy and Robertson closing.  Turn games into 5-6 inning affairs (assuming Betances comes in mid-6th to put out fires).  Writers seem to be bringing it up more and it's a pretty solid strategy I suppose.

 

[Paradigm]

That's an important point. Right now, pitchers are paid significantly more money as starters than as relievers. If that gap closes and teams allocate more money to the bullpen, then good pitchers may be more willing to become relievers because they might be leaving less money on the table than in the past.
 
This is especially true if there's a greater likelihood that they get to that payday. Whatever the likelihood is of making it as a starter, it is probably a little better to make it as a reliever. 

 

[Paradigm]

Lurker "trs" chimes in from Madrid:
 

 
 
I realize you don't know me from any other pseudonymous person on the internet, but since you started what I think to be a very interesting thread concerning pitcher usage, I'd just like to throw in a few cents. As I mostly lurk, and that I don't do that frequently, I'm not sure if this has come up, but I wonder how much of pitcher usage is dictated by the 5IP = win rule. I realize some managers in the post-season have adopted a split the innings strategy and foregone the traditional starting pitcher, but I wonder how effective this might be if used as an institutional strategy. If it truly takes a few/many years of development (arm strength, secondary and tertiary pitch mastery, control, etc) to produce a potential starter, wouldn't it be more cost-effective and perhaps also just plain ol' effective to develop a roster of 11-12 'relievers' capable of 3IP every 2 days?

As I'm sure if mostly obvious, most pitchers for financial reasons would not want to join a team where if they're lucky they would just build up a healthy 'hold' count and not much more for the counting stats, let's just put that aside. Since most pitcher abuse stats focus on pitches thrown well beyond what a pitcher would throw in 3IP might not a roster full of slightly stretched out relievers be more consistent and require less rest? I certainly don't have a medical background and so I'm not sure if it's necessarily healthier to exert more for less time than vice-versa (the MMA is better than boxing argument), but perhaps that too would figure into the situation.

I just wonder if it's a case of the tail wagging the dog here and that stats that are supposed to measure performance are actually controlling performance because there is the unspoken pressure to give pitchers 'wins.' We know this pressure exists (as it does with saves and that I know has been discussed many times with the relief ace argument), but is that pressure reducing pitching staff efficiency?

 

[Minneapolis Millers]

Riffing off lurker TRS's ideas, I'd like to see a team use the 5th spot in its rotation for starter/reliever development. The team could intentionally treat that spot like a spring training game: schedule 2 guys to throw 3 innings and fill the rest with RPs. You could have each of the guys throw additional relief innings during the week. Each three-inning stint could be rotated around 3-4 pitchers. If that works well, you could expand it to the 4th rotation slot. Or you might find that one pitcher develops that third pitch over time and becomes a consistent starter. This might be a better evolutionary approach to pitcher development at the MLB level.
 
Edit to clean up mistyping.

 

[Cesar Crespo]

Didn't Tony LaRussa try the 3ip thing?

 

[geoduck no quahog]

He did, in July 1993. There's a Seattle Times article about it. He did it out of desperation in late July because his starters were all shitting their pants.
 

La Russa, who has 13 pitchers on his active roster, scrapped the traditional five-man rotation for a nine-man rotation divided into three-man platoons. Todd Van Poppel, Ron Darling and Kelly Downs are in the first group; Mike Mohler, Bobby Witt and John Briscoe in the second; and Bob Welch, Shawn Hillegas and Rich Gossage in the third.
 
Dennis Eckersley remains the closer in a four-man bullpen.
 

 
Per LaRussa:
 

"It's not a concession," he told the Oakland Tribune. "A concession is if you cover your(self) by doing nothing. We're in sixth place. . . . What are we jeopardizing?"
 
Don't ask Darling.
 
"I don't even mess with saying I truly understand what's going on," Darling said. "We have to stay ready to go, and then go until he taps us on the shoulder."
 
 

 
Looking at the Baseball Reference page for that team, the bar graph game results after 20 July look pretty horrible. I'd have to go through each game log to see what happened, and I don't feel like doing that.
 
Put it this way - if it worked, Joe Maddon would have claimed credit by now.

 

[Just a bit outside]

I think that throwing 3 innings every third game would be more taxing on an arm and lead to more injuries. Most relievers throw an inning or less per appearance and still only appear in half the games at most. The value in that is that you can air it out on every pitch. A three inning appearance would still force pitchers to conserve energy and defeat this purpose.

The question to me is can you, or should you, try and develop bullpen arms? The Royals have developed 2 of their guys as relievers through the minors and they are dominant. The Sox have the opposite philosophy and leave guys as starters as long as possible. I wonder if turning Barnes, RDLR, Workman, and Webster all into relievers to go with Tazawa, Koji, and Layne would give you six guys that you trust in high leverage situations.
Layne as a loogy.

 

[Oppo]

KC not using wade davis in the middle innings of game 4 with the game on the line blows my mind. KC goes on to get blown out, Davis pitches game 5 already losing.

 

[DennyDoyle'sBoil]

I've often wondered whether it might make sense to use bullpen to get through the line up once, and then bring the "starter" in.  Using a pinch hitter is costly early on, and you need to set your lineup based on the pitcher you'll see for the most innings.  So, if you face a "reliever" early on, you might be less inclined to pinch hit.  So I sort of think it might be a way to maintain a platoon advantage early without worrying about pinch hitters.  There are probably a ton of downsides.  The one that jumps to mind most readily is that you're using bullpen you may never need to use, or with you had saved, based on how the starter performs.  And, whatever platoon advantage you gained in the early innings is negated when they stick a pinch hitter in there against your "starter" who is working his fifth inning in the 8th inning, and you don't have as much to work with in the bullpen.  The idea, though, is to get some outs and innings over with quickly, before turning to your longer use, pitch every fifth day, doesn't get bothered as much by platoon advantage guy.
 
Sometimes I think that managing pitching is the next big thing -- the next Billy Beane OBP type of move that could make a difference over 162 games.  On the other hand, though, if there really were an advantage to be gained by now using non-traditional pitching set ups, it probably would have been discovered by now. 
 
I guess breaking it down, you have the facts that: (1) Once a player is used he cannot come back.  (2) Some pitchers are most effective working a larger number of innings every few days.  (3) Others are most effective working a shorter number of innings and can do that more often.  What's the arrangement that takes best advantage of these three factors and minimizes opponents' runs is the question, I suppose.  My sense is that the answer is "exactly the way it's always been done," but it's hard to say because how can you experiment when all the games count?  Starters seem to need their 5 days whether they pitch 3 innings or 8 innings, and you rarely know what it's going to be, so any system where bullpen usage isn't dictated by starter usage seems destined to fail or result in inefficiencies.  But perhaps when you get really consistent starters, who seem capable of pitching six innings at least 80 percent of the time, there could be some way to manipulate things to get a slight edge.

 

[ivanvamp]

I think this is a great topic, worth a lot more investigation. Not only for bullpen discussion, but for lineup construction.

It used to be if you can get a pitch count up you could get to the bullpen easier, and then start putting up crooked numbers. But what if teams' better pitchers are now in bullpen roles? Putting up zero after zero. Then it doesn't make sense to try to get to them, because you're more likely to have them shut you down.

And if teams are turning the game over to them more quickly, maybe it affects your run scoring strategy, and thus your roster construction. Maybe the relative value of OBP and SLG changes. Or maybe base running becomes more important.

I don't know. I wonder what the effect on offenses vs strategy will be.

 

[HriniakPosterChild]

I've often wondered whether it might make sense to use bullpen to get through the line up once, and then bring the "starter" in.  ...

 
The team that tries that first is likely to face a grievance from the arb-eligible [SIZE=14.3999996185303px]pitcher[/SIZE] who knows that 20-win guys have big bags of cookies waiting for them in free agency.

 

[HriniakPosterChild]

 
Not likely. If the bullpen started the game with the premise of getting through the lineup once, the "starter" would be coming in in the 3rd inning most likely, maybe the 4th if the bullpen is lights out, 1-2-3 for three innings (not likely). In that case, the win would be at the discretion of the scorer, since the nominal starter in this case didn't get through 5 innings. If a "starter" comes in in that early and finishes a game, or even gets to the 8th or 9th, he's likely getting the win. 
 
There's bigger problems with that experiment than worrying about a grievance. 

And if the team loses because the starter didn't really start the game....

 

[williams_482]

I think this is a great topic, worth a lot more investigation. Not only for bullpen discussion, but for lineup construction.

It used to be if you can get a pitch count up you could get to the bullpen easier, and then start putting up crooked numbers. But what if teams' better pitchers are now in bullpen roles? Putting up zero after zero. Then it doesn't make sense to try to get to them, because you're more likely to have them shut you down.

And if teams are turning the game over to them more quickly, maybe it affects your run scoring strategy, and thus your roster construction. Maybe the relative value of OBP and SLG changes. Or maybe base running becomes more important.

I don't know. I wonder what the effect on offenses vs strategy will be.

"Getting to the bullpen quicker" sounds good on paper, but it was never that big of a deal and never the real reason why grinding out long at bats and drawing walks worked so well as a strategy (in reality, that is just the best way of getting yourself on base and giving your teammates as many additional at bats as possible). Relievers consistently perform better than starters as a group, and a pitcher generally reduces his ERA by roughly one run by moving to the bullpen. The fact is (and has been for quite some time) that even the relatively soft parts of a bullpen will generally perform better in a given inning or plate appearance than a typical starting pitcher. 
 
In 2004, for example, Starters as a group averaged out to a 4.62 ERA and a 4.55 FIP. Of 136 qualified relievers (4.5 per team), 104 (76%) had a better ERA and 98 (72%) had a better FIP. On the flip side, the average reliever had a 4.17 ERA and a 4.31 FIP. Of 86 qualified starters, 43 (50%) had a better ERA and a slightly different 43 had a better FIP. If we expand that to all starters with 100+ innings pitched we get 129 starters, 54 (42%) with a better ERA and 56 (43%) with a better FIP than the league average for relievers. 
 
However, if teams did decide that going to the bull pen earlier and more often was worthwhile and wouldn't result in burning out their relievers too quickly or the like, we might see an even lower run environment which would make OBP slightly less valuable relative to SLG than it is now because extra base runners don't score as much, extra PAs don't end with as many good results, etc. 

 

[jscola85]

Apologies if someone already mentioned this, but Fangraphs just posted an article this week somewhat relevant to this topic, in that teams may start to build their rosters away from patience because bullpens are so deep now that you may prefer to keep the starter in there as long as possible:
 

 
“There are a lot of ways to look at how you might augment your offense, but it can’t just be nine guys working a pitching staff over,” said Maddon. “If your goal is to get a starter out of a game, that might be the last thing you want to do. You see a lot of 95-plus out of the pen now, and some of those guys have quality secondary pitches. I think it’s become easier to build bullpens, and it’s rare a team has a bad one.”
The Kansas City Royals are a fit for Maddon’s musings. Not only is their pen dominant, their speed-focused offense posted the lowest walk rate in the game.
“We might possibly need to see a trend away from seeing pitches,” suggested Maddon. “I can see speed – including using it creatively – becoming a more important part of the game. I think the trend might be going back to the way the game had been before the unrealistic home run numbers arrived and walks became prominent. I really don’t know.”

 
http://www.fangraphs.com/blogs/sunday-notes-maddon-cherington-fixing-the-reds-trusting-buck-more/
 
For a team like the 2014 Sox, who struggled tremendously against high-velocity pitchers, getting away from starters and into bullpens the likes of the Royals doesn't seem like such a wise choice.  As such, having grinder lineups that force starting pitchers out in the 5th or 6th inning doesn't have the same benefit it used to, because instead of a John Halama or Curt Leskanic type coming in, you may see Kelvin Herrera as the first guy out of a bullpen.
 
EDIT - the bolded emphasis was mine, not Fangraphs'

 

[HriniakPosterChild]

 
Then the bullpen member that was responsible for the runs would get the loss, just as he would if he came in and lost the game at the end rather than the beginning. 
 
I'm not seeing the scenario you're worried about. Nor do I think a player would have cause to file a grievance. Teams can deploy players as they see fit. If the 2008 Red Sox had decided to acquire a closer, thus demoting a soon to be arbitration eligible Papelbon to set up and depriving him of saves, I don't see how he would be able to file a grievance over that. RDLR is entering his arb years next season, are the Sox somehow forbidden from making him a reliever? 

 
No, not really. The days of the reserve clause are gone. If teams could deploy players as they saw fit, Jay Payton would have been Trot Nixon's platoon partner through all of 2005 and 2006. 
 
More to the point, a relief pitcher late in a game is used in such a way to maximize the chances of getting an out (platoon advantage, personal splits, etc.). That advantage is gone if Abramah Able has the first inning, Bob Baker the 2nd, and so on.

 

[HriniakPosterChild]

You made a very specific point that a player could file a grievance. If you want to change that and say a player would bitch and moan, that's fine, but don't move the goalposts. This has nothing to do with the reserve clause. Teams can, in fact, deploy players as they see fit. The repercussions in the clubhouse or performance wise of that choice is another discussion. There is nothing in the CBA that says a team must use a player in a certain role as he demands or else risk a grievance. To suggest as much is nonsense. 

 
 
Nonsense, huh.
 
In 2005, Lou Piniella threatened to work his pitchers backward and start the games with relievers and end the game with starters. No grievances were filed because the phones started ringing in the Tampa Bay GM's office. Agents made it clear that Lou was welcome do that, and there would be grievances filed.
 
As best as I can recall, Lou never tried it.
 
---
Edit: He never did. Next day's news (but I can't find a link for the grievance threats from agents, which I am absolutely sure I heard about at the time):
 

 
"I didn't think this would draw that much attention," Piniella said.
 
The change of heart for Piniella came later Tuesday night when he sat down to weigh the pros and cons of starting a reliever. When he saw some items on the con side that disturbed him, Piniella decided to stick with his pitching rotation the way he has had it all along.
 
"I've got some young kids in the rotation that I thought about," Piniella said of changing his mind. "We're trying to develop a ball club as well as win baseball games here, so we just decided to stay the conventional way and that's it.
 
"It sounds intriguing. It may work, or it may not work, but we'll let somebody else try it first."

 

[HriniakPosterChild]

 
Please quote the portion of that article - or any other - that says grievances were threatened or were possible. 

 
You didn't read the edit? 

 

[HriniakPosterChild]

 
I posted before you edited, but yes, I read it.
 
Am I now being referred to the part where you "are absolutely sure you heard about grievances at the time" yet can't find a single link; or the lengthy quote from Pinella which states nothing even referring to grievances and proceeds to list the reasons why it was a dumb idea? 
 
I'm not sure what you think that proves. Or that you even understand what grievances are, at this point. 

 
I put no stock in the reasons managers give in public for the reasons the do things, because I've spent my entire career working for managers who do things for the reasons they do them, not for the reasons they say. I saw enough of Piniella managing in Seattle that I know he never gave a rat's ass about "some young kids in the rotation."
 
I certainly understand what grievances are ("Look! It's Messersmith and McNally!") and regret that I cannot find the precise google search terms to produce a link for what I know I read in 2005. Very sorry.

 

[HriniakPosterChild]

So, again, what did the quote in your edit - that you cited me to go back and read - prove exactly?

 
Edit> I can't find a link for the grievance threats from agents, which I am absolutely sure I heard about at the time
 
PP> Please quote the portion of that article - or any other - that says grievances were threatened or were possible.