How Paying Established Closers Saves Teams Money

Ryan Cook's arbitration earnings may not escalate as quickly as he would hope (via Dinur Blum).

Ryan Cook’s arbitration earnings may not escalate as quickly as he would hope (via Dinur Blum).

Teams that pay steep price tags for “proven closers” often get ridiculed by statistically-inclined baseball writers and readers, often for good reason. This tends to be especially true when these teams already have a viable replacement in their bullpen who is cheaper and maybe even a superior pitcher.

However, as Jeff Sullivan wrote on FanGraphs, when those teams include the Athletics and Rays, we tend to step back and reevaluate our positions since these teams typically know what they’re doing. When the A’s traded for Jim Johnson in early December it was a bit of a head-scratcher, as the team had two young relievers perfectly capable of handling the ninth inning in Ryan Cook and Sean Doolittle. While Sullivan pointed out at the time that the A’s roster was deep and it would be difficult for them to upgrade elsewhere, there was another reason that didn’t receive as much attention: signing a veteran reliever could keep the young relievers out of the closer role, thus reducing their cost.

The Athletics were the most prominent team to make a move like this, but plenty of other situations where teams with excellent young relievers and departing closers opted to sign a veteran to take over the ninth inning. The Rays acquired Heath Bell (and later signed Grant Balfour) instead of giving strikeout machine Jake McGee a chance to close, and the Indians followed suit by signing John Axford (coming off a -0.5 WAR season) to handle the ninth inning despite Cody Allen’s strong 2013 campaign. Most recently, the Mariners committed $14 million so that Fernando Rodney could close out games while Danny Farquhar, with his 34.7 percent strikeout rate and 1.86 FIP in 2013, got bumped to a setup role. Earlier in the offseason, the Rockies signed journeyman LaTroy Hawkins to close despite the presence of Rex Brothers as well.

So what’s really happening here? In three of these cases, the veteran closer had actually performed significantly worse than the young set-up man in 2013, yet the teams went out and committed significant money to these free agents. While these moves may be frustrating for baseball fans, a deeper analysis reveals that these deals might make quite a bit of economic sense for these teams. The idea of saving money on young relievers is occasionally thrown around, but how much money can teams really save?

Let’s try to keep things relatively simple by focusing on the Athletics, given the fact that their in-house closer candidates have amassed multiple excellent seasons but have yet to reach arbitration. You can make a case that the other teams might not have been comfortable with their in-house options, but that argument simply doesn’t hold water for the A’s.

For the purpose of this research, we’ll assume that had the A’s kept their bullpen intact after Balfour’s departure, Ryan Cook would have handled the ninth inning. Cook has a 2.55 ERA over 148 innings in two seasons with the A’s and 7.2 innings with Diamondbacks and has earned 16 saves in that time. So far, Cook looks like an elite set-up man, pitching mostly in the seventh and eighth innings and earning 44 holds with the A’s. Luckily for us, the cost for an elite set-up man in his first year of arbitration is pretty well defined as you can see from the pre-arbitration stats and first-year arbitration salaries in the table below.

Elite Set-Up Men, First Time Through Arbitration
Player IP ERA WAR SV Hold Arb 1 ($M)
David Robertson 172 2.62 4.0 3 53 1.6
Jonny Venters 171 1.89 2.9 6 59 1.6
Luke Gregerson 209 3.10 3.2 3 83 1.6
Tyler Clippard 239 2.52 3.0 1 64 1.6
Bobby Parnell 163 2.98 2.0 13 38 1.6
Sergio Romo 144 2.38 4.3 3 54 1.6
Average 183 2.58 3.23 5 59 1.6

If we assume Cook maintains his current level of excellence in 2014, he’d project to beat most of this group in innings pitched (222), WAR (4.5) and holds (68).  However, even players with lots of innings (Clippard) or holds (Gregerson) received the same $1.6 million figure. Cook also already has more saves than anyone in the group at 16, but once again we see that Parnell’s 13 saves didn’t help him out-earn the others.

I think it’s safe to assume that if Cook remains an elite set-up man in 2014, he should project to earn $1.6 million in arbitration. Now, how much might Cook earn in his first year of arbitration if he were to succeed as the A’s closer in 2014? Luckily, there are a pair of decent comps from this past season for relievers who didn’t become full-time closers until their final pre-arb year.

One-Year Closers, First Time Through Arbitration
Player IP ERA WAR SV Hold Arb 1 ($M)
Steve Cishek 188 2.54 3.2 52 16 3.8
Ernesto Frieri 198 2.96 1.8 60 13 3.8
Average 193 2.75 2.50 56 15 3.8

You can make a strong case that many of the relievers in the first group had actually pitched better than Cishek and Frieri. No matter, the 50+ saves from each of them pushed their earnings to $3.8 million. You could argue that Cook is better than these guys, but since we’re being conservative let’s just pencil in Ryan Cook the hypothetical closer for the same $3.8 million.

Looks like Billy Beane has already saved $2.2 million from his 2015 payroll, not too shabby. But since we know that arbitration raises are built off of the previous year, saving $2.2 million in year one will also lead to savings in years two and three. This is where the numbers get a bit more difficult. Let’s go back to the first group and look at what a few of the relievers earned in the rest of arbitration.

Elite Set-Up Men, All Times Through Arbitration
Player Arb 1 ($M) Arb 2 SV Arb 2 ($M) Arb 3 SV Arb 3 ($M) Arb Total ($M)
David Robertson 1.6 5 3.1 8 5.2 9.9
Tyler Clippard 1.6 33 4.0 33 5.9 11.5
Bobby Parnell 1.6 35 3.7 (2015) (2015) ?

Note: The saves (Sv) listed for each arbitration year are the cumulative career figures for each player at that point in his career.

Robertson appears to be the best comp for an elite set-up man who never gets a chance to close, earning just under $10 million over the course of his contract. Clippard and Parnell fit what we might expect if Cook gets a shot to close in 2015 after Johnson’s contract expires, although Clippard went back to being a set-up man in his penultimate year and we won’t know Parnell’s final arbitration contract for another year.

Since Clippard’s ERA (3.72) and Parnell’s low innings total (50 IP) leading into their second arbitration year may have cost them, let’s assume Cook can beat the pair and earn $4.2 million in 2015 in his first year as a full-time closer. To fill in the rest of the blanks, let’s take a look at a few recent relievers who followed similar paths to see how their arbitration costs escalated.

Recent Closers, All Times Through Arbitration
Player Super 2 Arb ($M) Arb 1 ($M) Arb 2 ($M) Arb 3 ($M)
Chris Perez 2.2 4.5 7.3 Non-Tender
Jim Johnson 1.0 2.6 6.5 10.0
Joel Hanrahan N/A 1.4 4.1 7.0
Matt Capps N/A 2.4 3.5 7.2

It’s not easy to find pitchers who didn’t start closing until their third or fourth year of team control, so unfortunately, three of the four recent comps were super-twos, two weren’t all that great, two were non-tendered, and one is Johnson himself. Perez got a $2.8 million raise after a season with 39 saves and a 3.59 ERA, while Hanrahan and Capps both earned $3.7 million raises after eclipsing the 40-save mark with ERAs under 2.5. The key takeaway here is that an effective pitcher who is earning saves as a full-time closer will see raises of $2-4 million in arbitration. While a reliever like Cook is certainly capable of putting up a monster season and earning a $3.5-4 million raise, let’s keep things simple and use $3 million.

The “R” in WAR
How a person can be a hero by being a zero.

Using this knowledge, we can put together a table estimating Ryan Cook’s expected arbitration cost if he had become a full-time closer in 2014 versus waiting until 2015 to close out games.

Ryan Cook, Potential vs. Probable Earnings
Player Arb 1 ($M) Arb 2 ($M) Arb 3 ($M) Arb Total ($M)
Ryan Cook w/ Johnson 1.6 4.2 7.2 13.0
Ryan Cook, closer 3.8 6.8 9.8 20.4

There you have it. By keeping Cook out of the ninth inning for just one year, the A’s appear to be saving around $7.4 million in arbitration costs. This makes the net cost of having Jim Johnson close for the A’s in 2014 around $3M. You could argue that these numbers are a bit generous, but assuming that Cook continues to be an effective reliever, the A’s appear to be saving at least $5 million with this move. A $10 million Jim Johnson doesn’t look too great, but at $3-5 million he has to be considered a steal.

This analysis also shows us the importance of the first year of arbitration. Since the salaries build on one another, an inflated figure in the first year will have a big impact on the subsequent years. Therefore, the difference in cost between a pitcher who starts closing during his first year of arbitration and one who starts closing during his final pre-arbitration year is much greater than the difference between the former pitcher and one who doesn’t close at all.

We also see the degree to which saves trump skill in the arbitration process. A pitcher like Chris Perez (0.8 career WAR in 333 innings) pitching in the ninth will end up being significantly more expensive than David Robertson (7.6 career WAR in 329 innings).

So where does this leave us? Teams that have effective young relievers in their pre-arbitration years when a vacancy in the ninth inning opens up are put in a tough position. If they hand over the keys to the ninth, they’re almost certain to pay the price in arbitration.

However, if teams sign a veteran closer to keep their young guy from racking up saves until after his first arbitration hearing, they stand to save a lot of money, possibly as much as $7-8 million. At the start of 2011, the Atlanta Braves chose to hand the keys to rookie flamethrower Craig Kimbrel. Now, he’s about to earn more than any other first-time arbitration-eligible reliever and may become so expensive that the Braves have to consider trading him.

Given the amount of money on the line, it’s not surprising that nearly every team that lost its closer to free agency and had a pre-arbitration reliever posed to take over opened up its wallet and signed a veteran. Effectively, these teams have a coupon that gives them a huge rebate on a free-agent closer, but they have to use it this offseason. For a team with deep pockets, it might make sense to simply let the best reliever close. However, smaller-market teams that need to squeeze more value of their cost-controlled players might be more likely to take advantage of this discount. In this context, these types of acquisitions become not only defensible, but actually quite brilliant.

Earlier in the offseason, Eno Sarris explored some of the complexities of the arbitration process on the FanGraphs website and podcast, noting how savvy teams could manipulate the system to save money. Nowhere is there a greater gap between the arbitration process and the open market than in the premium placed on saves.

As front offices get wiser and it becomes more difficult to find values on the free-agent market, teams (especially those with tighter budgets) are always searching for new market inefficiencies to get an edge. Judging by some of these transactions, it appears that teams are realizing the economic advantage that they can get by exploiting the arbitration process to suppress the costs of their own players.


Print This Post
Matt is a PhD student researching cancer and stem cell biology, and spends some of his free time writing about the analytics of beer at BeerGraphs and contributing to The Hardball Times. Follow him on twitter at @murphym45.
Sort by:   newest | oldest | most voted
Aaron (UK)
Guest
Aaron (UK)

Superb work. It will be interesting to see how Parnell fills out that table, given the somewhat unusual usage of Clippard.

And if you can’t afford a closer, or [more likely] have no pretensions to contend this year, you’re better off getting a AAAA reliever to do the closing whilst keeping your best young relief talent in the 7th & 8th.

Patrick Golden
Guest
I’ve been beating this drum for years. It also gives the team the ability to pair the better relievers with the higher leverage situations against the highest wOBA batters on the opposing squad, as opposed to pigeonholing them to the ninth inning where they may end up facing some combination of the 5,6,7,8,9 hitters in a lower leverage environment. If I were the rays, I would use my best relievers in set up until they reached arbitration. After one arbitration year, which sets his future value lower due to the raise schedule, you can then promote them to the closer… Read more »
Matt
Guest
Matt

The first arbitration year really is the most important. The difference between a reliever who starts closing the year before and the year after arbitration might be around $7M, but after that first arbitration hearing / settlement, you can move your guy to the 9th and have him close for 3 years and he’ll only be about $3M more expensive than he remained a set-up man.

Alex S.
Guest
Alex S.

This is really great work, Matt. I also agree with what Patrick said in that this also enables teams to use a better reliever (such as Ryan Cook, for instance) with more flexibility. Bob Melvin now has the luxury of using Cook in potentially more high leverage spots and to come in and mop up any potential messes that arise prior to the ninth inning. This would allow the A’s to get to the ninth with the lead more often when (hopefully) Johnson can shut the door.

Atreyu Jones
Guest
Atreyu Jones

Even with the current usage, closers face higher leverage situations than set-up men.

Jeff
Guest
Jeff

I think teams would be better off adopting what I call the Two run save rule.

Use a closer in one and two run save situations and if the lead is three runs (97%+ conversion rate) then use whomever you like out of the pen.

That will spread around 15 or so saves amongst the various members of the bullpen and keep the closers’ save total at a reasonable amount, as well as making him more available for high leverage situations that are not saves.

Jfree
Guest
Jfree

We’re thinking along the same lines

Paul G.
Guest
Paul G.
The strategy you propose makes sense, or at least more sense than how elite relievers are used now. The problem is the current “9th inning, saves situations only” usage that is universally adopted is probably one of the worst options for exploiting an elite reliever, but for some reason major league baseball teams seem insistent on sticking to it. The current closer usage started from a rational concern (we are overusing the relief ace and he is getting less effective/hurt late in the season) but it went too far in the opposite direction (we’ll use our relief ace for 50-60… Read more »
Matt
Guest

Having elite relievers for high-leverage situations outside of the 9th inning is definitely helpful, but it’s important to remember that today’s closers are still pitching in high-leverage situations. The top 16 relievers by average leverage index in 2013 were all closers (14 had at least 25 saves), and they averaged 63 innings (all but three cracked 60). While it doesn’t make sense to avoid using your best reliever in a high-pressure situation in the 7th or 8th, the 9th inning is generally the highest-leverage place to pitch.

Paul G.
Guest
Paul G.
True, the 9th will often be the best spot for the relief ace. When I said “worst” I should have qualified that as among at least semi-rational strategies. “Worst” would probably be something like saving the relief ace for mop-up duty or using him as a pinch hitter or loaning him to the hot dog selling concession. The problem is the goal should be to win games. The current strategy’s goal is two-fold: to maximize saves for one pitcher and to keep the manager from being second guessed. Because leverage tends to be higher in the 9th there is significant… Read more »
Jfree
Guest
Jfree
Seems to me that the better option would be to forgo the ‘closer’ slot completely – and just spread the save opportunities around the bullpen. Just like in the old pre-closer days. There is probably one legitimate skill set in that 9th inning role – lack of platoon splits – but other than that, the cost of saves in arbitration is probably a function of both the perceived ‘closer role’ and the saves themselves. Statistically, ‘saves’ in arbitration years would have an exponential cost (not a linear one) attached to them. Reducing all of those saves per reliever to below… Read more »
Scott
Guest
Scott

In addition to the approx $5-7.5 million saved on Cook, could this also save the A’s $3 million on Doolittle over time by keeping Doolittle’s 8th innings and random saves down?

I.E. Doolittle first arb year he makes a little less, leading to a little less in arb 2 and 3, ultimately making Johnson Free?

Paul Swydan
Admin

Yes, it absolutely could/may. I think Matt just stuck with Cook only for simplicity sake.

Matt
Guest

It’s possible that they could save some money on Doolittle as well, especially if their closer suffered a significant injury and he would normally have to take over the 9th. However, it doesn’t appear that a handful of saves has a significant impact on relievers that aren’t full-time closers (see Parnell v. Clippard in the first table).

Scott
Guest
Scott

I saw that, just was wondering if a deeper look might reveal some smaller savings over a larger sample size of set-up men. . Is there any difference in arb pricing between a succefull 6-7th inning reliever ( what Doolittle would become pre-arb) vs a succefull 8th inning reliever ( what Doolittle was pre-arb). Is “holds” the only stat that matters or does “primary set-up man” have any value?

Steve
Guest
Steve
But to keep Cook out of the closer role, they don’t have to pay Johnson just this year, but they have to pay for someone like Johnson for three years. If the A’s had Johnson for just one year, at a number you consider a steal, $5M, but then had to pay Cook ARB2 and ARB3 money based on him becoming the closer, they probably don’t save a dime, and now they’ve used Johnson instead of Cook in important high leverage situations. The A’s aren’t shelling out $3-5M once (or in this case $10M), they’re going to have to do… Read more »
Matt
Guest

The final figures were calculated based on the assumption that Cook takes over the closer role in 2015, as clearly stated in the article. The $7M figure is based on the money they save simply by delaying Cook’s move to the 9th by one year.

Steve
Guest
Steve

Hmm, apparently I misread. But we’re still looking at the A’s spending $10M to gain $7M, and to be worse in the process. Call me when teams actually save money and get better.

Matt
Guest

The A’s bullpen must be really deep if adding a reliever who’s pitched 230 innings with a 2.70 ERA over the past three years is making them worse.

Steve
Guest
Steve

Seems a bit of a pissy response. I didnt say adding him made them worse, but their organization of the pen was sub-ideal. If thats not true, then theres no point to this article. If Johnson is actually their best reliever, then his addition is less about saving money (which, again, he isnt doing anyway).

Brian
Guest
Brian

To his point, the As will still need to replace Johnson’s roster spot in Year 2 and Year 3 when Webb takes over at closer. This very well could be a cheap pre-arb guy where the math works out that there are still net savings, but still.

Atreyu Jones
Guest
Atreyu Jones

Unless I missed it, this article doesn’t address the other available option: sign a FA set-up man instead of a proven closer and use him as closer. A Uehara or a Benoit is not always available, but if they are teams could save at both ends – by not paying for magic closer pixie dust, and buy keeping their young players’ arb prices down.

Matt
Guest
This is an excellent point. The important idea is simply signing a free agent to handle the ninth inning, and it doesn’t matter whether they’re a closer. However, pretty much all of the teams employing this strategy are looking to be competitive in 2014, and if you’re going to be leaving an elite reliever in the 8th, you’d better be pretty sure that the guy you’re signing can handle the 9th, or the fans and media are going to be up in arms if he falters. Also, I think the convergence of contracts for set-up men and closers also plays… Read more »
Atreyu Jones
Guest
Atreyu Jones

I actually meant Benoit’s last contract with Detroit. It was still big compared to other set-up men, but much smaller than good closers’.

This year, I’m not sure if their actually was a FA who was a good candidate to be signed and then promoted to closer. Joe Smith maybe? J.P. Howell? Scott Downs? Ed Mujica? Jesse Crain? These guys project to be worse than Jim Johnson, but not all that much.

Matt
Guest
Those would be the guys. However, none of them have the track record of performance that Johnson has. Smith might be the best of the bunch, but even he signed for $15.75M over 3 years. Downs was cheap and has been solid for a few years, but has some platoon issues. Crain would have been a perfect candidate were it not for his health concerns. I think the key is that these teams are hoping to make the playoffs and contend for a world series. If you try to save money by going with Downs and he blows a few… Read more »
Atreyu Jones
Guest
Atreyu Jones

Yes, this year might not have been the off-season to elevate a FA set-up man into a closer. But the risk that a Joe Smith type might fail and cause you to promote your pre-arb set-up man to closer is also there if you sign Jim Johnson (although slightly lower).

Matt
Guest

There’s definitely still some risk, but it’s going to be a heck of a lot easier to explain to the average fan if Johnson hits a rough patch and they stick with him in the 9th.

Atreyu Jones
Guest
Atreyu Jones

Well fans turn on any closers who blow saves in bunches whether or not they have closing history (with the exceptions of Rivera or Hoffman types).

Sanjay
Guest
Sanjay

This is fascinating analysis, and it illustrates a new market efficiency. I don’t think, however, that we can really conclude that Johnson was ‘free’ if the expected savings on arb > $10M. While the Johnson acquisition looks defensible in isolation, it looks a lot worse when you consider that the Rays acquired Grant Balfour for 2/$14M (or was it 12?–can’t remember).

Which I guess would have made Balfour’s effective cost way below zero?

Matt
Guest
This is a good point. My main argument is simply that these teams can save a large sum of money by preventing their pre-arb relivers from getting saves. The Johnson/Balfour argument is another story entirely. When the A’s traded for Johnson, it looked like Balfour might get 2/$16 or more, and I understand not wanting to guarantee a second year to a reliever who just turned 36. Also keep in mind that the “savings” will be different for every team. In the case of the A’s it’s easiest to project, because their set-up guys are one year away from arbitration… Read more »
Alex McIntyre
Guest
Alex McIntyre
This is absolutley fantastic analysis. But I think another point needing mention, besides the financial benefits to restricting controllable relievers from the closing role, is the non cost-related benefits to having a possibly superior reliever closing in, say, the eighth or seventh inning, where they may actually encounter more valuable situations. Often, though many fans overglorify the ninth inning as the most pressure-filled, situations that may appear in the innings prior may require a more skilled or apt reliever. Paying Jim Johnson the $3 million also frees the team to use Cook in a situation where he may be needed… Read more »
Luis
Guest

This is outstanding work Matt. People were wondering if there was some market inefficiency teams were exploiting with all these strange free agent closer signings, and you’ve proven what it is. Very smart analysis on your part.

Luis
Guest

I also wonder if the teams realize this or if they’re just making the same old “pay for the saves” mistake and this is all just a coincidence. Doesn’t take away from your analysis, you’re clearly right, but there are some teams that I hesitate giving too much credit to.

Matt
Guest
Thanks for the feedback. This is something I’ve thought about as well, and it’s certainly possible that some of teams doing this are just kind of getting lucky. I think it’s most likely that the smaller budget teams know what they’re doing, because committing big money to a free agent closer is a bit of an unusual move for them. For a team like the Mariners who had money to burn and want to compete this season, it may have been more of an accident. I do, however, think that it’s a little silly for us to assume that any… Read more »
rubesandbabes
Guest
rubesandbabes

This is an outstanding article even before it trots out the analysis/proofs.

Jim Pagels
Guest

Would it not make sense to simply sign an elite reliever rather than an “established closer” to be the new closer, though?

Then you save money in preventing your elite relievers from racking up saves and you save money in terms of the new free agent you signed.

What am I missing?

Matt
Guest
This has been brought up a couple times in comments or via twitter, but the gap between the free agent prices of “proven” closer and set-up guys is shrinking, and it’s not easy (especially this offseason) to find a good reliever that’s cheap just because he doesn’t have saves. A team looking to make the postseason will want to make sure that their free-agent closer is at least close to being as good as the guy they’re blocking. For teams that aren’t competitive, this strategy makes a lot of sense. Although we need to remember that losing lots of games… Read more »
Matt
Guest
Just to follow up on this, there are only 56 that have accumulated at least 2 fWAR over the past three seasons. Only ten of them became free agents this offseason (doesn’t include Darren Oliver, who retired): Joe Nathan (4.3 WAR) – Closer, paid $20M over 2 years to close for Tigers Jesse Crain (3.7 WAR) – Set-up guy, but coming off a shoulder injury ($3.2M/1yr) Fernando Rodney (3.2 WAR) – Closer, paid $14M over 2 years to close for Mariners Joaquin Benoit (3.0 WAR) – Closed for most of 2013, paid $15.5M/2yrs by Padres, probably as set-up man Matt… Read more »
Jim Pagels
Guest

Thanks for the thorough response!

I suppose my question, then, is… If it seems that any good free agent set-up guys or closers are going to be expensive, then your post should simply be titled, “Paying established relievers (rather than “closers”) saves teams money,” right?

Matt
Guest

That could certainly work, but in the examples we’ve seen this offseason it’s been the “closers” who are being paid to close, the set-up men are being paid more but still usually to set up.

B Mac
Guest
B Mac

Jason Motte; last year of contract, coming back from TJ, proven closer now throwing 96 mph again. What are his chances of taking the job back from Rosenthal? I have scoured the news, and nobody thinks this is going to happen, but based on your article, it looks like a guarantee.

And it nicely illustrates the utility of this theory to fantasy baseball players, too!

wpDiscuz