Sunday, April 25, 2010

Bullpen Redux.

First off, I want to address the main criticism I’ve heard about my last article. As reader Dan wrote in the comment section:

“I’m really surprised that Rany missed Craig’s point, which many have already pointed out. The best use of the best reliever in your bullpen is not necessarily as a closer. Bring him in anytime in the last three innings when the scoring threat is greatest (men are on base, middle of the order, etc). Let him get out of that jam, and whether he gets three outs or stay[s] in for six, he got you out of the jam and you still have the lead. Worry about keeping a lead in the ninth when there is actually a lead in the ninth to worry about.”

I’ve seen a lot of comments like these, saying that it’s okay for Soria to come in to pitch the 7th inning in a key situation, because he doesn’t necessarily have to finish the game – once the crisis has passed, someone else can pitch the ninth.

I get this point, I really do. And I agree with it to some extent – in a perfect world, we wouldn’t have “closers”, we’d have “stoppers”, who are brought in to stop an opponent’s rally cold.

But the whole point of my article was that the way teams use their bullpens has evolved so drastically over the last 30 years that it’s a foolish utopian dream to argue that Trey Hillman should immediately jump back into the Dan Quisenberry model. Before we turn Joakim Soria into Quisenberry, I argued, let’s first turn him into Jeff Montgomery. Once the Royals see that Soria can handle that kind of role without the world ending, then we can talk about extending things further.

But to idea that the Royals should bring Soria in to pitch in the 7th and then come out of the game, letting someone else to pitch the 9th (and earn the save), is even more of a pipe dream than having him go 7 or 8 outs for the save. It’s simply never been done before. Never mind that the fans and media will absolutely revolt the first time Robinson Tejeda or Josh Rupe blows the 2-run lead in the 9th that Soria hands off to him. (This to me is actually a minor concern, because – for reasons I may get into in a future post – the media and fan base in Kansas City is probably more willing to accept sabermetric and other “unconventional” baseball ideas than anywhere else in America.)

Here’s a rough description of the four ways a team can use its closer, from the least efficient to the most efficient:

1) The Eckersley model: 9th-inning save situations only.

2) The Montgomery model: pitch the 9th, and sometimes the 8th, usually save situations but also in tie games.

3) The Quisenberry model: come in whenever the game is on the line, from the 6th inning on, and stay in to close out the game.

4) The Utopian model: come in to put out fires at any point in the game, then exit the game and let a lesser reliever pitch once the fire is out.

Currently, pretty much every team in baseball follows the first method; a few teams (including the Royals) will dabble with the second method. In my last article I argued for the Royals to whole-heartedly adopt Method #2, and worry about Method #3 later. The response has been “what about Method #4?” To which I can only say, “let’s get the Royals to embrace Method #2 first.” Sure, I’d love to see the Royals adopt Method #4 – but I’d also love to see them embrace the 4-man rotation, and carry just 10 pitchers, and build a platoon at a couple of positions, and clone George Brett. But step by step, guys.

Otto Von Bismarck once said, “Politics is the art of the possible.” I am trying to be political here, because I want to see the Royals (and every other team) use their best reliever more efficiently, and I’d rather push for a small change that’s possible than a large change that isn’t. I’m not saying that the people pushing for the Royals to use Soria in the Utopian model are wrong. They’re not. I’m not even saying that they’re wasting their time, because they’re not. As I’ve learned from reading Nate Silver, if nothing else, suggesting the Utopian model may help move the Overton Window, which may make a less extreme but still radical idea like using Soria for 6 outs become more palatable.

So by all means, let’s keep reminding the Royals that there’s a better way to use Soria. But let’s also temper our expectations. It took 30 years for teams to make a complete mess of the way they use their closers. It’s going to take time – and a lot of intermediate steps – to reverse the damage.

- Last night, Hillman brought Soria in to pitch the 9th in a tie game. This isn’t that notable – once the game goes into the 9th inning tied, it becomes impossible for the home team to have a save situation, so you have no choice but to use your closer in a tie game. Still, there are a lot of managers who would have waited at least until extra innings before bringing in their closer. Then, after Soria worked a rather efficient 9th inning with just 12 pitches, Hillman let him throw the 10th as well. If Yuniesky Betancourt had ever bothered to learn that you don’t actually have to swing at every pitch, Hillman might have been rewarded for his willingness to defy conventional wisdom with a victory. (In the bottom of the 10th, with the winning run on third and one out, Betancourt saw four pitches. None of them were in the strike zone, yet he swung at three of them, and struck out.)

The bullpen remains a cluster**** of the highest order, but I stand by my position that Hillman is mostly an innocent bystander in all this. It’s hard for any manager to look smart when he has exactly one reliable reliever. Hillman deserves better than to be scapegoated because his GM has made such a mess of things.

- Speaking of Dayton Moore and his mess, Moore turned the impressive trick of dumping Luis Mendoza without actually improving the bullpen. Mendoza was replaced with Bruce Chen, and while Chen pitched poorly for the Royals last season, his career ERA (4.71) is barely half that of Mendoza (8.43). Moreover, in three starts for Omaha he had allowed just 13 hits in 21 innings, with 20 strikeouts against just 5 walks. Chen is one of those Quadruple-A guys who always dominates in Triple-A but struggles in the majors, so I’m not expecting much from him, but he’s better than Mendoza.

Unfortunately, Moore also released Juan Cruz and replaced him with Brad Thompson. This is mystifying. I understand that Cruz has been an unbridled failure with the Royals, and it’s impressive that Moore was willing to accept the sunk cost of his contract (and that David Glass let him). But at least Cruz has a history of success, something Thompson really doesn’t. Cruz went on the DL late last year, and based on the results I suspect he was pitching hurt for most of the season. After striking out 33% of the batters he faced in 2007-2008, he struck out just 17% of opposing batters last year.

This year, he was ridiculously ineffective with men on base (he allowed all 6 runners he inherited to score), but he also struck out 7 of the 28 batters he faced. There was at least a glimmer of hope there. I don’t see how, in just 5.1 innings of work, he showed the Royals enough to release him – particularly when they’re keeping Kyle Farnsworth, who’s been almost exactly as ineffective. (Farnsworth has thrown 5.2 innings. Both pitchers have allowed 9 hits. Cruz walked four; Farnsworth has walked only two, but also hit two batters, and he’s given up a homer. And Farnsworth has only three strikeouts.)

In his place, the Royals have brought up Thompson, who had a 4.84 ERA for the Cardinals last year, and is an extreme finesse pitcher who struck out just 34 batters in 80 innings. Given how bad the team’s defense still is – as put on vivid display each of the last few nights – I’m not sure a pitcher who’s primary skill is pitching to contact is going to be any better than Cruz.

The Royals have a wealth of relief options in the minors, from the just-waiting-for-his-turn Carlos Rosa to the nearly-ready Blake Wood to the always-forgotten Chris Hayes to some intriguing Double-A options like Blaine Hardy and Federico Castaneda. None of these guys are sure things; they’re all still better than recycling yet another failed major league pitcher.

- Just to make it clear how much the bullpen has hurt the Royals, consider this: if games ended after six innings, the Royals would be 8-6 with 3 ties. Instead, they’re 6-11. Get this: the Royals have a losing record when leading after six innings. In the eight games they’ve led after six, they’re 3-5.

The 7th inning is the culprit. Thanks to Soria, the Royals have won all 5 games they’ve had a lead going into the 9th. And they’ve only lost 1 game where they had a lead entering the 8th. But between the beginning and the end of the 7th, the Royals go from 8-6 with 3 ties to 6-10 with 1 tie.

- What is most amazing to me about the pitching staff this season is that, despite a middle relief corps that is historically inept, no one has even suggested that one possible solution might be to let the starting pitchers work a little deeper into the game.

The Royals have played 17 games so far, and not once has their starter thrown even 110 pitches. Only six times has the starter thrown 100 pitches. Neither Kyle Davies nor Brian Bannister has thrown 100 pitches yet.

I’m not saying this is a bad thing, necessarily – the usage of starting pitchers today is much more sane than it was 10 or 15 years ago. I’m just saying that if you had told me back in 1998, when I first invented Pitcher Abuse Points because Jim Leyland was letting Jesus Sanchez throw 146 pitches and Livan Hernandez throw 153 pitches and Tony Muser was grinding Jose Rosado’s arm into dust, that a scant 12 years later a team would rather lose game after game in the 7th inning than let its starting pitchers (none of whom are younger than 26, mind you) throw even 110 pitches…well, let’s just say I would have been surprised.

The starting pitchers have averaged 5.82 innings per start this year; even if you take Gil Meche out of the equation, it’s 6.17 innings per start. The Royals’ starter has completed the 7th inning just twice all season. The Royals could easily let their starters get another out or two – allowing them to complete the 7th more regularly – without letting them exceed 115 or 120 pitches. For pitchers with mature arms (i.e. older than 25), I don’t really get concerned with their pitch counts until they approach 120.

It’s clear the Royals don’t have a reliever capable of protecting leads in the 7th. Maybe the solution is to see if their starters are capable of protecting leads in the 7th.


Unknown said...

There is a realistic way to maximize Soria. 2 issues get in the way of maximizing a good pitchers usage. 1. Contracts that reward saves 2. Modern relievers preferences for structure in their roles.
For #1 you simply adjust his contract to reward holds equal to saves or develop a matrix that pays Soria for being a good pitcher not just getting a certain stat.
For #2 as a manager the 1 thing I want to avoid is using him in blow outs because he hasn't pitched in 6 days, those are the worst.

Before the season I approach him and tell him the goal is never to use him in a blowout and never have him not pitch in several games in a row. He will likely respond well to this. I will use him as a traditonal closer mostly 9th inning saves, a few 8th inning saves and in some tie games late.

BUT if he has not pitched in 3 days I want to be able to use him in a close game 6th inning on for a maximum of 30 pitches. He would then go back to being a traditional closer.

The best of both worlds- maximize the numbers and pay homage to modern usage and player preferences.

Anonymous said...

Otto Von Bismarck once said, “Politics is the art of the possible.”


I prefer Voltaire's "The perfect is the enemy of the good." Which is exactly what you're saying.

We need to get rid of saves. Its a miserable statistic that rewards extremely varying levels of performance in non-obvious ways, like pitcher wins and loses.

Marmot said...

Rany, Rany, Rany. Can we let the whole Chris Hayes thing go? He's given up 11 hits in 5.2 AAA innings this year. Small sample size be damned, that's a .458 batting average against. With no (that's zero, zip, nada) strikeouts. Add it to last year and we're looking at 78 hits in 54 AAA innings and about a .340 average against.

Hayes turns every hitter he faces in Joe Mauer and you're seriously suggesting that's he's the answer in KC? There's just no evidnece to support that matter how much you want it to be true.

Mitch said...

The idea of using Soria and then bringing in another pitcher to pitch the ninth will blow up in KC's face. Not anyone, not Mendoza, Cruz, Farnsworth, or Tejeda have proved they can hold onto a lead consistently. Soria can. Parrish can, sometimes. Rupe seems like he can.

If you put in Soria in the 7th, and let him pitch the eighth, but put in Tejeda to start the ninth, you're asking for trouble.

There's a rather now-common saying for this kind of system: bullpen-by-committee. That's not a saying fans like to hear when it's tossed around by coaches and managers. Most people think it doesn't work, however, it can work if you have a bullpen that pitches similar to the way Twins pitchers are right now.

In the simplest terms, if you have a bullpen with a 3.00 ERA overall, you are giving up 1 run every three innings. If the starter goes six, then if your bullpen has a 3.00 ERA (which only two AL teams are achieving this year so far), you could reasonably add 1 run to the opponent's total at the beginning of the seventh to get an approximation of where the score will end up.

The Royals bullpen is averaging somewhere around 6.00. If you take out Soria, it's going to be closer to 7.00. That's over two runs per three innings.

A bullpen by committee, right now, would not work in KC. I agree with other Royal fans when they say let Soria pitch more outs, let the starters go a little bit longer, pare down the bullpen and get some players that can pinch-run or be a defensive replacement later in the game (or in extra innings, huh?).

If we want to go "Utopian", we need a much better middle relief staff than we currently have, plain and simple.

Rany said...


Well, I'm certainly not banging the Disco drum quite as hard as I was yesterday. He was doing fine until last night - he had allowed just one run in his first four innings. Then yesterday he allowed 5 hits (all singles) in one inning, which this early in the season obviously kills his numbers.

I obviously don't think he's earned a promotion at this point. But I'm not writing him off either.

Karte said...

Hey, maybe this change to Utopian use of the Closer won't take so long after all.

12 years after you started tracking Pitcher Abuse Points (and one year after getting hammered by fans for overusing pitchers), the Royals are not letting their starters go more than 110 pitches.

The only problem that I see with a Utopian bullpen is we do not have Utopian relievers (apart from Soria).

It doesn't really matter when you bring Soria in if the next reliever can't get outs.

Nathan said...

Rany, I wonder if the limited usage of Royals starters so far this year isn't a function of still being fairly early in the season. I don't know much about it personally, but the conventional wisdom seems to be that 120 pitches in April /= 120 pitches in July. I suspect that, over the next few weeks, the starters will begin to have a longer leash.

Nathan said...


You should look into the concept of relief pitcher leverage. You'll find there's no justification for making it a hard and fast rule to limit your best relief pitcher at the end of the game only. It is sometimes wise to use your best reliever in the 9th. But not always.

I agree with Rany, though, that this won't change overnight, and it's nothing to call for Hillman's head over.

Nathan said...

Jim wrote,

"It doesn't really matter when you bring Soria in if the next reliever can't get outs."

Actually, that is precisely why it does matter when you bring Soria in! It is the large disparity between Soria and the other relievers that makes his usage pattern important. Hillman can't avoid having to sometimes use relatively ineffective relievers. But he could use them in situations where they're least likely to cost us the game. Sometimes, those low-leverage situations are save situations.

Fast Eddie said...

I really can't understand why they keep bypassing Carlos Rosa.

Shinatoo said...

for the royals to bring Soria in the 6th or 7th would be like grabbing a branch before falling off a cliff. You are just delaying the inevitable. Or like that moment when Wiley Coyote hovers in the air before he realizes he's standing in the middle of the thin air. Like that cartoon the Royals chase the Road Runner only to wind up a little circle of smoke at the bottom of the canyon.

Bob McWilliams said...

On another topic, I read in the Star today that Gordon is likely headed back to Omaha once the immortal Chris Getz comes off the DL. I am so baffled. Assuming Gordon is healthy, why are the Royals going to screw around with one of the few potential impact players they have????? Yes, Gordon has been a disappointment, but the time to find out what he could still be is this season. And you won't find out with him on the bench or in Omaha. This one is really making me mad. But hey, dayton traded for Getz and Fields, so gotta have one of them playing, right?

Mike said...

Two things come to mind after having read your past two columns (well, I read ALL your posts, but I'm commenting on the past two):

--First, as far as the Quisenberry Model for a closer, is it really fair to use Dan Quisenberry as an example of what could be employed by someone like Joakim Soria? The reason I ask this is the simple fact that Quisenberry's delivery was of the extreme submarine variety, which is much closer to that of an underhanded softball delivery, and pitchers who throw that way put much less stress on their arms. Hence, it was much easier for Quisenberry to regularly enter a game anywhere between the sixth and ninth innings and finish a game, whereas asking Soria to drastically increase his workload could lead to more shoulder or elbow problems for a guy who has already undergone Tommy John surgery and has never thrown more than 73 innings in a professional season.

--My second comment has to do with your assertion that it'd be okay for Trey Hillman to allow the Royals' "mature" (26 years old or older) pitchers to go longer in their starts. Right now, you mentioned that the Royals haven't had a pitcher reach 110 pitches this season, but that you think they could easily work up to around 120 pitches without issue. This brings to mind Gil Meche's notable start last year, in which he threw a complete game by reaching 132 pitches, and it's something that many of us, including yourself on several occasions, have been extremely critical of Hillman for.

My question is, if 132 pitches is too much, would it really make much of a difference if Royals pitchers had a green light to regularly work up to about 120 pitches? I mean, over the course of the season, wouldn't the possibility of being allowed to throw 10-20 more pitches than, say, the 100-pitch limit they might now be held to prove just as likely, if not more, to cause certain pitchers arm issues?



Nathan said...

I think Blayne's idea has potential. If a manager set things up like that, I suspect a reliever would respond well, and it would probably be worth several wins over the course of the year. It's an innovative compromise that could work out well for everyone, if ever proposed.

Anonymous said...

I can almost gaurentee that every single fan that want the Royals to use Soria in the 6th or 7th will bitch and moan at Hillman if the next pitcher gives up the game....."Hillman brought Soria in way too early, he's a complete moron.

Anonymous said...

Blayne....why not add the hold incentive to ALL the bullpen guys and not just Soria?

Anonymous said...

Rany, I think you missed a great example of a manager doing exactly what we would like to see Hillman doing (and it's even recent too!). In 2008, Joe Maddon used Troy Percival in traditional "closer" situations (i.e. in the 9th with a 2 or 3 run lead) and he racked up nearly 30 saves. However, his best reliever was either J.P. Howell (thank god we got Gaithright for him!) or Balfour. He regularly used those two in the 6th, 7th or 8th to stop opponents from rallying. As a matter of fact, J.P. Howell had the highest leverage index of any reliever in baseball that year. Maddon was clearly aware that it is more important to have your best reliever in the game in the inning when the opponent is threatening to start a rally than it is to have them in a "save" situation. I would argue that Hillman should do the same thing with the Royals bullpen. Carlos Rosa should be able to nail down an acceptable rate of "easy" saves in the 9th to allow Hillman to use Soria in a more mercenary manner. I'd love to see the look on the opposing managers face when he gets a leadoff walk in the 7th or 8th and the heart of his lineup is coming up... only to have the Mexicutioner come trotting in to slam the door.

Anonymous said...

"By all means, keep reminding the Royals there is a better way to use Soria"

Uh - yeah. Move him into the rotation.

gsmith601 said...

I'm surprised no mention of Bryan Bullington as a possible call up???


Wabbitkiller said...

Marmot, you're missing the point in regards to Chris Hayes. The point is this: Hayes can certainly do no worse than the retreads that we already know SUCK. At least there's a CHANCE that Hayes will be productive. We already know that guys like Mendoza WON'T.

Anonymous said...

Royals fans amaze me. Last season everyone was crying that hillman left starters in too long. This season they are crying they are not being left in long enough. Ugh.

Anonymous said...

Rany, other than on Opening Day, I don't think the starters have been taken out too early at all. Trey's generally been letting them stay in until they get into trouble, and unfortunately, that hasn't been more than 6 innings, for the most part.

28 years and counting said...

The best idea would be if Soria could wear disguises and have something like three different jerseys. Then he could just come in for one or two outs multiple times throughout the game and thus earn holds and saves. It's a given that wherever you use Soria to "get out of jams" there would be a subsequent jam that would pop-up later in the game. However with the multiple uniform strategy he can enter in the 7th after Tejada gives up a quick pair of runs (under disguise) and get a hold then come back as himself in the 9th and get the save.

Anonymous said...

Can we just take all the relievers except for Soria and line them up in front of a firing squad? Or at least demote them all to the Kansas City Little League, where they might get someone out occasionally? said...

Rany, if one reason modern managers use relievers incorrectly is the save statistic, maybe we shoulld invent a new statistic. The "hold" isn't sufficient for our purpose. What we need is something called the "escape". A reliever would get an escape if he is brought in the seventh or later with two or more men on base and gets out of the inning without allowing a run. If all the sabermetric icons then promoted the "escape" and "escape percentage" as the best way to evaluate relievers, then maybe managers might start using them more rationally (and GM's might reward their relievers in a way that reflects their true performance). said...

Also, the "escape" would have to be in a situation where your team is either leading or trailing by 2 runs or less. You might also consider giving an escape if a reliever is brought into a bases loaded situation with 1 out or less, and gives up only one run.

Kansas City Little League said...

To Anon @ 2:21:

The various Little League Baseball associations and leagues of the greater Kansas City area insist on a certain level of quality of play. Kansas City Little League pitchers must show a certain level of competence in order to be allowed to pitch. No pitcher on the staff of the Kansas City Royals other than Joakim Soria meets the standards of Kansas City Little League, as they have not demonstrated a level of pitching skill even suitable for retiring Little League hitters.

May we suggest they be reassigned as peanut vendors at Kauffman Stadium?

Michael said...


I hope you gave Jeff Passan the OK to basically rewrite your post from a Twins perspective!;_ylt=AupvxQK2ci7jRoiRjvEgpeERvLYF?slug=jp-relievers042610

Evan H. said...

With regards to the Eckersley method, what effect would repeatedly warming up have on Soria? Say Luke Hochevar has a one run lead in the 6th, and there are no outs with men on first and second with Damon, Ordonez, and Cabrera coming up. Naturally, you start getting Soria warm. But then pretend Luke gets Damon to pop up and then Ordonez grounds in to a double play. What do you do with Soria, bring him in for the 7th or sit him down and possibly warm him again next time the meat of the order comes back up?

Once teams start going with the Eckersley method (and I think it's inevitable, although it may take a while), these questions will come up. Have there been any studies regarding the effect of excessively warming up pitchers in the bullpen? How much less taxing is a bullpen pitch than a game pitch?

Anonymous said...

"I’m not even saying that they’re wasting their time, because they’re not."

Stop using the comma before the subordinating conjunction "because."

Anonymous said...

I wonder if Dayton MORON still wishes he had Tony Pena Jr. around? All he's done is put up an 0.91 ERA in Triple-A so far. MORON figured out Pena should be a pitcher, and then let him walk as soon as he proved he could.

kpellow said...

In my utopian model, there'd be no distinguishing between starters, relievers, closers, stoppers, etc. In a close game, I'd run out the best pitcher available, period. If it meant occaisionally running guys up from Omaha when needed, so be it. In other words, win today, worry about tomorrow tomorrow!

Jacob G. said...

I know we won't get the Quiz model overnight here in KC. But i was sad to see this morning (after the Seattle debacle blew another Greinke win) that Soria has pitched in 7 games, while Tejada, Farnsworth, Parrish and Hughes have all been in more (10 for Tejada, 11 for Hughes!). Even if we cannot get the way pitchers are treated altered, there must be SOME way we can ensure that our best bullpen arm leads the group in appearances.
Or is that too much to ask for?

Anonymous said...

I believe these numbers are correct regarding Greinke:

Greinke: 31.2 IP, 9 ER, 2.56 ERA

Bullpen when Greinke pitches:
14.1 IP, 20 ER, 12.56 ERA

In 31.2 innings this year, the Royals have had 4 runs or more in exactly 1 inning when Zack has been in the game. That was Opening Day.
It would be interesting to know how Greinke's run support has been this year compared to some other pitchers.

Royals Mafia said...

Dayton Moore needs a pair of cement overshoes, along with his sorry excuse for a bullpen!

Ben F said...

While I want to see situation 4 sooner rather than later, Tuesday's game was situation 2 or 3, yet Soria wasn't used. The Royals' good starters + terrible bullpen + good closer may be the perfect storm for an active sabr fanbase to actually prod the mgmt to buck convention and use Soria as a hybrid fireman/closer.