Mariano Rivera’s Projections and the Value of the Closer

It’s not exactly surprising that of all the pitchers that the fans have projected so far, Mariano Rivera is the leader in saves, ERA, WHIP, and K/BB. After all, Mo has been hands down the greatest reliever of his era, if not all time, and has shown no indication of slowing down. Here’s what his fan projection line looks like (as of 2:40 AM ET on Friday)

77 IP, 4-4, 2.38 ERA, 41 SV, 9.23 K/9, 1.52 BB/9, 6.08 K/BB, 0.70 HR/9, 2.61 FIP

Pretty fantastic numbers. It’s hard to imagine a reliever with a better line, except for maybe Mariano himself. Of course, this brings up one of the more interesting questions regarding value in baseball, and that’s the value of the closer. Right now, this line by Rivera is projected to be worth 2.4 WAR. This is easily the best of the three relief pitchers so far projected, besting David Aardsma‘s 1.3 WAR and Takashi Saito‘s 0.5 WAR. If the crowd is truly wise, it will remain the highest. So then is 2.4 WAR truly as good as a reliever can be?

Probably not. That’s mostly because this measurement doesn’t account for leverage index. This certainly can be done, and it is done in some places. Sean Smith‘s WAR database at www.baseballprojection.com uses the mean of 1.0 and the pitcher’s actual leverage index. As Smith explains in his stat definitions, this is because of the effects of chaining. That is, the actual replacement for the closer’s innings isn’t a replacement level pitcher. The replacement is actually the setup man, who, on most teams, is nowhere near replacement level. As such, giving the closer full credit for the leverage of his innings isn’t a true representation of his value.

Even if we did, though, we still wouldn’t see any closer come near the value of the top-flight starters. Tim Lincecum right now is projected for an 8.0 win season by the fans (which is slightly ridiculous, but that’s another matter), and Mariano Rivera would be projected for a 4.8 win season with a LI of 2.0, slightly high for a closer. Similarly, Rivera comes nowhere near our top position players, Chase Utley (8.2 WAR) and Albert Pujols (8.0 WAR). The key here is that Rivera simply will not take part in enough events over the course of the season to impact a game nearly as much as the Utleys and the Greinkes, or the Pujolses and the Lincecums. His events may be better and they may be more important, but sometimes quantity does trump quality.

Mariano Rivera’s true value probably falls somewhere between the 2.4 WAR the base numbers give and the 4.8 WAR his fully leveraged numbers give. Remember that this isn’t only the best reliever out there, but a guy that is by a wide margin the best reliever. This is why many analysts, such as myself, are constantly wary of big money signings or marquee trades involving closers.



Print This Post



Jack Moore's work can be seen at VICE Sports and anywhere else you're willing to pay him to write. Buy his e-book.


Sort by:   newest | oldest | most voted
Bill
Guest
Bill
6 years 7 months ago

Tim Lincecum had an 8.2 WAR this season, so why is 8.0 WAR ridiculous?

Mike K
Member
Mike K
6 years 7 months ago

Because in the 8 years FanGraphs has computed WAR data for, there have only been 9 pitcher-seasons of at least 8 WAR. So you’re basically projecting him to have one of the 10 best seasons of the past decade. That’s not to say he *won’t* do it, but that seems like it should be more a +1 SD prediction than a baseline.

Bronnt
Member
Bronnt
6 years 7 months ago

WHAT!?

You’re saying the fans are projecting Lincecum’s line to fall somewhere in between his numbers from the last two years? What could they possibly be thinking?! This is completely outrageous. I demand an investigation.

Sky Kalkman
Member
6 years 7 months ago

1. There’s a better chance Lincecum falls short of 8 WAR than exceeds it.
2. There’s a lot more room to fall short of 8 WAR than there is to exceed it.

Another 8 WAR season is not ridiculous at all. *Expecting it* is.

Sky Kalkman
Member
6 years 7 months ago

Or try this — go find all the 7+ WAR seasons in the past few years. Then see what those players did the next year. Guessing they dropped off a win or two, on average. Sh*t happens.

Kevin S.
Member
Kevin S.
6 years 7 months ago

2.3 wins, actually.

I took all pitcher-seasons from ’02 to ’08 and compared them what that pitcher did the following season. There were 20 such seasons. 12 of those seasons were followed by a season with a decline of at least .5 WAR (that’s roughly-accepted error term, right?), and only Lincecum improved by more than .5 WAR after a seven-win season. Only seven pitchers had another seven-win season the following year, and only Johan had three straight 7+ WAR seasons (’04-’06).

Scottwood
Member
Scottwood
6 years 7 months ago

Interesting finding. Thanks for the research. So, in essence, based on regression, we could project for a WAR of 6 or so for Lincecum and Verlander and for Greinke to have a WAR around 7.

Obviously that is overly simplistic. But, it does give us a ballpark figure.

Ralph
Guest
Ralph
6 years 7 months ago

If the closer’s replacement is the setup man, do we also need to account for the fact that the setup man will also need to be replaced, and so will his replacement, etc?

John
Guest
John
6 years 7 months ago

There’s an argument Nathan is better

twinsfan
Member
twinsfan
6 years 7 months ago

There may be an argument, but it’s not a particularly good one.

Tom B
Guest
Tom B
6 years 7 months ago

obviously he doesn’t feel like sharing that argument with us. nathan can only hope to be as good as the GOAT each and every year.

Steve
Guest
Steve
6 years 7 months ago

no there isn’t.

Kevin S.
Member
Kevin S.
6 years 7 months ago

There’s an argument that Nathan’s approximately as good over the past three years. Even that’s somewhat remarkable.

adohaj
Member
adohaj
6 years 7 months ago

How isn’t there one!?

Since 2004 (when Nathan started closing)

Nathan
IP 418
era 1.87
whip .934
k/9 11.1
k/bb 4.32
SV% 91

Rivera
IP 440
era 1.90
whip .936
k/9 8.7
k/bb 5.37
SV% 93

Mariano lets up more hits but walks the same amount less note the almost equal whip. Nathan strikes out more batters. They both yield the same results since era and sv% are basically the same.

How then is there no argument? It seems to me that either pitcher could be said to be the “best” closer in the last 6 years.

Steve
Guest
Steve
6 years 7 months ago

and what in those numbers says Nathan is “better”?

the idential ERA in fewer innings? the lower save %?

the statement wasn’t “there is an argument that Nathan is almost as good as Rivera”. the statement was “There is an argument that Nathan is BETTER than Rivera”.

Explain where you’ve demonstrated that.

Sam
Guest
Sam
6 years 7 months ago

And that is even before accounting for the fact that Rivera pitches in AL East, and Nathan pitches in AL Central.

Nathan is an elite reliever. But he is not Mariano Rivera. No one is.

adohaj
Member
adohaj
6 years 7 months ago

By saying that there was no argument I guess I wrongly inferred you were saying Nathan was inferior. Nathan is not inferior to Rivera they are different pitchers but they produce the same results. As for the argument about Mariano in the post season. Nathan hasn’t had the same opportunity of 100+ innings to prove himself equal.

In my previous post I was saying that after looking at the numbers it is reasonable to say that Nathan is better than Mariano. There are also reasonable arguments saying that the opposite is true. I just think it is foolish to say the argument does not exist when the pitchers are so similar.

Dirty Water
Member
Dirty Water
6 years 7 months ago

Ha! love this one.

Yes, those numbers CLEARLY indicate that Nathan is the better reliever (as does ERA+), if only slightly. But he’s not, not even close. We all know that. So… Nathan is not a better reliever than Rivera for reasons not reflected in numbers.

And that’s suppose to be taboo here at Fangraphs. There is no such thing as anything that the numbers do not bear out, right? so it must be said that Nathan is a better reliever than Mo, so the author is wrong to state otherwise.

TCQ
Guest
TCQ
6 years 7 months ago

Water,

1) You the same dude that posts over on the CBS baseball ‘board? Used to read that semi-regularly.

2) The argument is whether or not Nathan has produced similar results(not better…ERA+ is flawed for reasons you really should know if you’ve been reading FanGraphs) to Rivera over the past few years. The thing that makes Rivera the better reliever is that he’s been doing it for a long, long time. Consistency and durability are virtues…

pm
Guest
pm
6 years 7 months ago

I think that a closer like Rivera is a lot more valuable than his numbers. There has to be some value in knowing that if you lead in the 7th inning that in 1 inning the game is basically over.

Tom B
Guest
Tom B
6 years 7 months ago

The only reason closers values are low that I can figure is that they are not always used at the optimal time of the game to achieve high value. If they were used when they should be, to stop the other team from winning at the highest-leverage point of the game, instead of just running them out there every time you have a 3 run lead, their values would likely be higher… right?

Jon
Member
Jon
6 years 7 months ago

In the 2010 Hardball Times Joe Nathan had the highest regular season championship expectency in 2009. A good bullpen is correlated with postseason success based on studies by Nate Silver.

I wonder if WAR doesn’t properly account for the value of a closer to a playoff caliber club. The financial impact of advancing in the playoffs is millions of dollars per round, so even a small change in win expectancy may be more significant.

Steve
Guest
Steve
6 years 7 months ago

what if the player in question consistently implodes in the playoffs?

adohaj
Member
adohaj
6 years 7 months ago

Nathan has a whole 8 innings of postseason data to judge from…

Tom B
Guest
Tom B
6 years 7 months ago

small sample size or not, nathan would have to pitch another 100 innings and only give up 3 runs to be even close.

Dirty Water
Member
Dirty Water
6 years 7 months ago

There is no difference between the regular season and playoffs except sample size. I know that cause I read it here. Ha!

TCQ
Guest
TCQ
6 years 7 months ago

You know, Water, that your viewpoint (judging from your posts on this article) isn’t really going to get any traction here, right? And it’s pretty unlikely that you’re going to get anybody riled up, so I’m not really getting your motivation.

Mike Green
Member
Mike Green
6 years 7 months ago

If you look at career raw WAR, Mariano Rivera’s is at about 40 (depending on which WAR you use). If you multiply that by his career LI, you still don’t get more than 75 and that is a significant under-estimation of his value because of his post-season performance.

In his particular case, because the Yankees use him during the regular season as though the post-season is most important, seasonal WAR really does not help you in valuing him. It sure would be nice if Mariano Rivera’s career post-season stat line- 133 innings, 0.74 ERA, 82 hits, 21 walks, 2 homers, 107 strikeouts- was better known than Derek Jeter’s clutch reputation.

don
Guest
don
6 years 7 months ago

I think his postseason exploits are pretty well known.

Either way it seems to me like projecting and valuing Rivera is a little different than most guys. He’s the best closer of all time and it’s really not close. He’s going to retire a Yankee, so his “value” on the market doesn’t matter much, and he’s a surefire first ballot Hall of Famer regardless of whether people consider postseason numbers (which they will).

Steve
Guest
Steve
6 years 7 months ago

next to Joe Nathan, of course.

by the way, I just saw that ball that A-Rod hit off Nathan fly by my window. It should land by Christmas though.

Kevin S.
Member
Kevin S.
6 years 7 months ago

I do have a question about the FIP calculations. It’s known that relievers typically out-perform their FIP, in terms of ERA. Why don’t we adjust our FIP calculations to recognize this? If Rivera’s 80% strand rate isn’t a career-long fluke, doesn’t that indicate that a FIP-based WAR calculation is somewhat undervaluing him?

Steve
Guest
Steve
6 years 7 months ago

Very good question.

Rivera almost always outperforms his FIP. That is, I would guess, mostly b/c his cutter induces a ton of weak contact. Broken bat grounders to 2B, etc.

timmy013
Member
timmy013
6 years 7 months ago

“I do have a question about the FIP calculations. It’s known that relievers typically out-perform their FIP, in terms of ERA.”

I could be mistaken, but it seems like part of this has to do with the fact that it is harder to give up an earned run when you don’t start an inning. If a reliever comes in with two outs and runners on in the 7th, his ERA is very unlikely to take a hit that inning.because he only needs to get one out before any of his hitters score. Even if the reliever starts the 8th, he will have an equal chance of giving up earned runs in the 8th as any starter of the same caliber starting any of his innings.

Therefore anytime a reliever enters a game with one or two outs his ERA will be helped out so we should use FIP or another similar stat to get the true value of relievers and how good they actually are.

Mike Green
Member
Mike Green
6 years 7 months ago

ERA is a poor measure for relievers for several reasons, not the least of which is the treatment of inherited runners. Rivera’s FIP also does not measure how good he has been because he has had a historically low career opposition BABIP due to the effectiveness of his cutter. WAR does not, I think, account, for this.

Kevin S.
Member
Kevin S.
6 years 7 months ago

Rivera’s hardly unique in being a reliever that strands extra runners or allows a lower BABiP over longer periods, though.

Rob in CT
Guest
Rob in CT
6 years 7 months ago

No, he’s just probably at or close to the extreme.

Kevin S.
Member
Kevin S.
6 years 7 months ago

I didn’t say they were as low as Rivera, just that they can tend to be lower.

Will
Guest
Will
6 years 7 months ago

Lots of good reasons for why Rivera’s true value is hard to calculate: psychological impact; proven control over balls hit in play; disproportionate contribution to close games and post season games; higher leverage; etc.

Sometimes, one doesn’t need to put a number on the player to determine is value. In fact, I don’t think it can be done with Rivera because the fact of the matter is he is priceless.

snapper
Member
snapper
6 years 7 months ago

Agree with all that, but you still woudln’t trade a good SP for him.

The Yankees would be MUCH,MUCH worse off if they lost CC Sabathia or AJ Burnett for the season than if they lost Rivera. The closer’s impact is limited b/c they only give you ~70 IP and there are only so many 1-run leads in the 9th.

A modern closer can’t approach the value of the 1970-80’s “stoppers” who pitched >100 IP often in multi-inning appearances.

Steve
Guest
Steve
6 years 7 months ago

Not arguing your larger point, but Mo WAS used in that “stopper” role all through the playoffs.

Will
Guest
Will
6 years 7 months ago

CC Sabathia is more than a good SP, so he is different. Also, that question would depend on the rest of the staff. As we saw in the playoffs, Mariano wound up pitching a higher percentage of innings than a normal closer does in the regular season. In spite of that, someone like Sabathia still played a bigger role. A pitcher like AJ, however, did not.

I’d break it down this way. In all cases in a post season, I would rather have a dominant closer (0 runs in 5 innings) than a solid starter (6 runs in 12 innings). If I had a strong rotation (like the 98 Yankees for example), I would also rather have the closer than a very good starter. But, if like the 2009 Yankees, the staff only had one true ace, I’d opt for the starter.

The way baseball is played now, you really need to have two rosters: one to win in the regular season and another to win in the post season. If you have more than enough to get you to the former, then the closer actually becomes alot more valuable.

mike
Guest
mike
6 years 7 months ago

I’ll end the conversation on Rivera vs. Nathan right now, eschewing all numbers: If you don’t live in New York, you have no idea how great Rivera is. I hate the Yankees and yet I have to admit how perfect the guy is. It’s laughable to think Nathan is in his class. Laughable. I don’t care what numbers you have backing your claim up. Again, live in New York for one baseball season and you will have no doubt.

twinsfan
Member
twinsfan
6 years 7 months ago

Are you sure you didn’t mean to post this on the ESPN forum?

Scottwood
Member
Scottwood
6 years 7 months ago

Just so I am clear, does the WAR for relievers account for their leverage index or do we have to manually do that?

NEPP
Guest
NEPP
6 years 7 months ago

Mo Rivera is and was a one-of-a-kind player. We will be lucky if we ever see another relief pitcher as good as him. Personally, I doubt it. Guys like Rivera, Pujols, Mays come around once a generation if at all.

Its amazing he’s still doing it at his age. I hate the Yankees but I have nothing but respect and admiration for Mo.

NBH
Guest
NBH
6 years 7 months ago

Poor Billy Wagner. He’s been pretty awesome as well:

Billy Wagner (career):
833 IP
11.79 K/9 (the best all time with a minimum of 800 IP)
2.39 ERA
1.01 WHIP
0.83 HR/9
385 saves vs. 33 blown saves (92% conversion rate)
2.79 FIP

Mariano Rivera (career):
1090 IP
8.31 K/9
2.25 ERA
1.01 WHIP
0.50 HR/9
526 saves against 28 blown saves (95% conversion rate)
2.78 FIP

Mariano Rivera = the greatest closer ever. But Wagner has been pretty fantastic and often gets overlooked.

Steve
Guest
Steve
6 years 7 months ago

Get back to me when Wagner does it in the majors instead of the NL.

/ducks

Steve
Guest
Steve
6 years 7 months ago

i’m kidding of course. Wagner has had a great career, and I wouldn’t be surprised if he sneaks into the HoF one day.

Will
Guest
Will
6 years 7 months ago

Mo’s ERA+ is 202, compared to 182 for Wagner. Also, Mo has 250 more innings. Wagner has had an excellent career, but Mariano has been much better. Also, Mariano has a silly 0.74 ERA in 133 post season innings, while Wagner has a 10.32 ERA in 11 post season innings. That speaks for itself.

NBH
Guest
NBH
6 years 7 months ago

Agreed, Mo is better and the AL/NL difference is big. But an 11.79 K/9 over 800+ IP is something special. It’s higher than Randy Johnson, Pedro and even Lincecum. Although it’s not as high as Lidge: 12.15 over 500+ IP. That’s sick.

Michael
Member
6 years 7 months ago

I though FG WAR took leverage chaining into account. Hmm, weird.

Mike D
Member
Member
Mike D
6 years 7 months ago

An argument can be made about Reliever A (Nathan, Wagner, Hoffman, etc.) vs. Mo that Mo has had so much more playoff experience than all of the other relievers, sample size, etc. but…

If any of those relievers were given that many playoff innings, does anyone really think that they would even come close to a 0.74 ERA, a number twice as good as an elite reliever’s ERA in the regular season.

Will
Guest
Will
6 years 7 months ago

Wagner not only has over 20% fewer innings, his high strikeout rate loses value since he gives up much more homers: those hits that do get through are far more impactive, which makes the fact that he has fewer saves despite far fewer saves pretty understandable.

Sky Kalkman
Member
6 years 7 months ago

“Deserved leverage” is something I’d love to see at Fangraphs. As in, if all relievers were properly distributed across all teams and all roles, what bullpen role would they have deserved and what’s the LI of that role? You can do some smoothing so there isn’t a sharp drop-off between tiers.

I’d also like to see a WPA based on PZR (to remove defense) for relievers (and starters and hitters) with the option of using WPA insetad of wRAA in WAR calculations. There are many different questions to answer with the available data, and WAR (as is) is only appropriate for some of them.

EK
Guest
EK
6 years 7 months ago

Can’t pitchers not control when they go in? So why should they get credit for the leverage index?

SharksRog
Guest
SharksRog
6 years 7 months ago

I agree that Tim LIncecum has much more room to regress than to progress. Nevertheless, I think there are reasons to believe that he will pitch significantly better next season.

While his K rate was flat last season, his hit rate, walk rate and OPS against each had similar drops as to 2008 — when his ERA plummeted from 4.00 to 2.62. I think Tim will continue to improve those rates next season, but even if he doesn’t, he could easily have a lower ERA than in 2009.

Either Tim’s ERA drop in 2008 was hugely more than should have been expected from his peripherals, or his drop in 2009 was far less than one could have anticipated from his peripheral drop.

Personally I think Tim’s ERA in 4.00 was abnormally high for his peripherals, so I think some of each is the case. I would be foolish to predict it, but I think there is a realistic chance that Tim’s ERA could drop below 2.00 next year.

In Sandy Koufax’s final season, he posted a career-best 1.73 ERA. His hit rate was the same as Tim’s in 2009, his walk rate was better, but his homer rate was a bit worse. Sandy’s OPS against that spectacular season was .545. Tim’s OPS against last year was .561 — a drop of 52 points from 2008 and of 111 points from 2007.

Believing Lincecum’s ERA might drop below two next season isn’t as unrealistic as it at first blush appears.

Kevin S.
Member
Kevin S.
6 years 7 months ago

There’s a difference between believing something might happen and expecting it to happen. I don’t think anybody’s said it’s impossible for Lincecum to improve – just unlikely that he would replicate this past season.

SharksRog
Guest
SharksRog
6 years 7 months ago

I understand your thinking here. My point is twofold:

First, Tim has steadily improved over his three seasons. Even though his ERA didn’t improve nearly as much last season as in the 2008 season, with the lone exception of strikeouts — which were already the best in the majors — his peripherals improved just about as much in 2009 as they had in 2008. What is to keep them from improving again?

Tim began last season somewhat weakened from two spring training injuries. The weakness showed, as he yielded 7 earned runs in just 8.1 innings in his first two starts — yet bounced back to win the Cy Young Award.

Second and perhaps more subtly and more importantly, Tim’s peripherals were pretty close to being as good as Sandy Koufax’s when Sandy had his career-best 1.73 ERA in 1966. It may be possible that Tim will be able to get significantly better results without even improving his peripherals, which have shown impressive improvement each of the past two seasons.

A decent chance to show a significant drop in ERA without peripheral improvement, and a histor of consistent and significant improvements from season to season — that’s a decent combination to provide an even lower ERA in 2010 than in either 2008 or 2009.

Can you point to any other pitcher who has dropped his OPS against by 111 points over the past two seasons? Can you point to any other pither who dropped his OPS by 52 points in 2009, yet saw his ERA decline by only a seventh of a run?

For that matter, can you point to any pitcher who has improved as much over the past two seasons as Tim has? There haven’t been many — and he started from a pretty high plane in his rookie season.

Julio
Guest
Julio
5 years 7 days ago

I predict that the Yankees Kei Igawa will have a WAR of 9.0 and Sergio Mitre will exceed a WAR of 10.

wpDiscuz