The Best Reliever of All-Time, Mariano Rivera

Detractors of advanced statistics love to point out specific instances where the advanced and traditional stats don’t align, as evidence that the saber-stats are far from perfect. You mean to tell me that Derek Jeter isn’t good at defense? That Ben Zobrist is as valuable as Robinson Cano? Or that Tim Raines deserves a Hall of Fame vote? Many of these sort of statements go against people’s first reactions and assumptions, making the stats backing them up an easy target for ridicule.

But these sort of debates miss the larger point: quite often, advanced stats agree with traditional assessments.  Torii Hunter was dang good at defense there for a while, and players like Cal Ripken Jr. and Ricky Henderson easily deserve to be in the Hall of Fame. Sports fans like to argue, though, and what fun is it debating topics with only one side? It’s much more entertaining to talk about things that are contentious, and as such, the distances between mainstream and saber stats can look like a gaping chasm when in actuality, it’s more like a meandering brook with handy stepping stones.

Enter Mariano Rivera. As you have surely heard, Mariano Rivera got his 602nd save yesterday, passing Trevor Hoffman on the all-time list and entering into first place all by his lonesome. As much as us saberists love to disparage the save, this record highlights an all too evident truth: Mo is the best relief pitcher of all time. And it’s not even close.

If you prefer old-school statistics, Rivera has the lowest career ERA (2.06) of any relief pitcher to throw more than 200 innings.* That’s considerably lower than Trevor Hoffman (2.87 ERA), and better than any of the current relievers in the Hall of Fame: Hoyt Wilhelm (2.52), Goose Gossage (2.63), Bruce Sutter (2.83), and Rollie Fingers (2.83), and Dennis Eckersley (2.85**). And not only that, but in an age where relievers typically don’t shoulder large workloads, Mo has thrown 1,159 innings — the 14th most in major league history.

Mo has the lowest walk rate of the Hall of Fame relievers (1.9 walks per nine), and his strikeout rate is by far the highest (8.3 per nine). He’s only blown 65 saves, while Goose Gossage blew 97 and Bruce Sutter blew 90…yet Mo has pitched in over 1,000 games while Gossage appeared in 900 and Sutter in 650. You can go on and on with the comparisons, but they’re still just as unremarkable; in short, the traditional stats make Mariano out to be a soul-destroying, bat breaking machine.

And when you turn to the new-school statistics, you find that it’s not just saves that really likes Mariano; it’s everything. He has the most career Wins Above Replacement (WAR) of any relief pitcher and by a longshot too. Heck, only nine relievers have ever pitched good and long enough to accrue over 20 WAR, and of those nine, only one has ever cracked 30. With 38.3 WAR, the next closest player (Goose Gossage) is a full 9 wins behind.

But it only gets better. Mo has by far the most Win Probability Added of any relief pitcher (55 WPA) over the course of his career, a full 23 ahead of the next closest guy (Trevor Hoffman). And he may have recently set the all-time saves record, but he’s already the all-time leader in Shutdowns (539).

This one save doesn’t change his place in history, but it does serves as yet another of how the saber and mainstream aren’t as far apart as you’d think. It doesn’t matter what statistics you like to look at, the final conclusion is the same: Mariano Rivera is not only the best old reliever of all-time; he’s easily the best reliever the game has ever seen. The role of the closer may be a young one and not steeped in legend yet, but Mo is crafting his own over in New York.

*You guys are so smart. One reader caught that Mo’s career ERA is actually 2.22, while I listed above that it’s 2.06. The numbers I referenced above are Mo’s stats as a relief pitcher; he had 10 starts early in his career that inflate his career ERA. And since Dennis Eckersley also started some games, I excluded all games started when comparing the relievers against each other.

**Again, this is Eckersley’s ERA as a relief pitcher.




Print This Post



Steve is the editor-in-chief of DRaysBay and the keeper of the FanGraphs Library. You can follow him on Twitter at @steveslow.

74 Responses to “The Best Reliever of All-Time, Mariano Rivera”

You can follow any responses to this entry through the RSS 2.0 feed.
  1. Tim says:

    You just gotta love Mo. Classy guy and one of the most dominant pitchers of this era. It’ll be a shame when the day comes that he isn’t shutting the door in the Bronx

    Vote -1 Vote +1

  2. Ryan says:

    FanGraphs taking the ol’ contrarian stance again?

    Vote -1 Vote +1

  3. Nathan says:

    I think WPA is a good stat to consider when talking about a pitcher’s (particularly a relief pitcher’s) effectiveness. And Rivera’s WPA (54) is second only to Clemen’s (76) for pitchers since 1974.

    I’d say that’s a strong argument for keeping effective relief pitchers in the relief role. One wonders how much higher his WPA might be if Rivera was used more in high leverage situations, rather that save situations.

    Vote -1 Vote +1

    • CircleChange11 says:

      I was watching a DVD the other day about my Cardinals’ 3 WS trips in the 80s.

      There’s lots of interesting things to notice, from the poder blue away jerseys, to gold necklaces, to high socks and on and on … but one thing that really stuck out to me the usage of the term “relief ace”.

      Relief ace as oppossed to “the closer”.

      I wondered if labelling the pitcher as “relief ace” rather than ‘closer” would change how managers used the pitcher. Might they bring him in earlier in the game in a higher leverage situation rtaher than saving him to “close the game”?

      I like the idea of a “relief ace”. Bases loaded 1-out, 8th inning with a 3-2 lead … here comes the relief ace.

      Actually, doing away with “saves” would probably accomplish the same thing. When we measure relievers by their “saves”, it’s no wonder that managers and players want the “relief ace” to pitch in the 9th.

      +8 Vote -1 Vote +1

      • Person says:

        There is something that turns top closers into relief aces: the postseason. In the playoffs (or a pennant race), Rivera is certainly coming in for that 1-run, 5-out save.

        There are also some teams where, whether by accident or design, the closer is the second best reliever. Managers will in that case use their best reliever for the highest leverage situations. The first team that comes to mind here is the Nationals, where Storen closes but Clippard often works the tougher innings, comes in with runners on, etc.

        Vote -1 Vote +1

      • Notrotographs says:

        “There is something that turns top closers into relief aces: the postseason.”

        Ron Washington is confused.

        +6 Vote -1 Vote +1

      • Black_Rose says:

        “I like the idea of a “relief ace”. Bases loaded 1-out, 8th inning with a 3-2 lead … here comes the relief ace.

        Actually, doing away with “saves”

        Mariano would have more blown saves, and possibly more meltdowns, if he was brought in really high leverage situations such as the one mentioned above.

        The article forgot to mention his low BABIP, which is probably skill-related.

        Vote -1 Vote +1

      • JimNYC says:

        There’s a real problem with Fangraphs’ WAR as it relates to Rivera — it massively discounts his obvious ability to suppress BABIP. 1100 innings isn’t a fluke. Watching him pitch, it’s obvious why his BABIP is lower than the average pitcher — the ball breaks so much that it’s impossible to get good wood on it, and you get tons of broken bat popups or lazy grounders. Fangraphs WAR just assumes that’s all random, which is… just wrong.

        Vote -1 Vote +1

  4. Jason says:

    His career ERA is 2.22, not 2.06.

    Vote -1 Vote +1

  5. CircleChange11 says:

    When it comes to Mo Rivera, I don;t think most people even look at his regular season accomplishments.

    What he has done in the post-season (and so much of it) is far above what everyone else has done, that it makes his regular season accomplishments seem boring.

    Vote -1 Vote +1

  6. tomhaywood says:

    Hi there,

    Just wondering where the discrepancy in blown save stats between yourselves and baseball refernce comes from. You quote 65 blown saves, they quote 72 blown saves.

    Cheers in advance.

    Vote -1 Vote +1

  7. Spunky says:

    Great article! Congrats Mo!

    Vote -1 Vote +1

  8. MikeS says:

    About the only argument you could make against this highly controversial stance (sarcasm) is that guys like Gossage and Sutter routinely threw more than 1 inning at a time and although Rivera may do it more often than many, he still doesn’t do it all that often.

    Gossage averaged 1.81 IP/appearance, Sutter 1.58 and Rivera 1.16. I didn’t take out there starts and Gossage made the most (37, none for Sutter) so it might pull up his average some. I bet Rivera would have had more longer outings 30 years ago but Tony LaRussa changed the game so he never really had a chance.

    Vote -1 Vote +1

  9. test says:

    Blown Saves should not be compared to pitchers who routinely had to pitch more than one inning and often came in with runners on (the exact opposite of Rivera’s regular season usage). It’s disingenuous to compared the Blown saves of anyone but Hoffman (I think he has 5-10 more, a negligible difference), the already HoFers, aside from Eckersley, just weren’t used the same way. I wouldn’t give Wilhelm exact credit for all those wins that modern closers just don’t have a chance at either…

    There is also a real concern about Rivera’s awesomeness being overkill for his role. Holding a two run lead the same or giving up one run, for a closer, results in exactly the same outcome for the team. Not his fault, but it does mean his greatness isn’t as useful as it should be. Yes, it’s another “stop managing the bullpen to maximize saves for one guy” comment.

    Vote -1 Vote +1

  10. Hurtlocker says:

    Any stats freak worth his salt knows how important it is to have a great bullpen. Mo made the Yankees better late in the game, period. A great reliever and sure HOF in my opinion.

    Vote -1 Vote +1

  11. Norm says:

    off topic-
    “But these sort of debates miss the larger point: quite often, advanced stats agree with traditional assessments. Torii Hunter was dang good at defense there for a while, ”
    -
    Interesting that UZR shows Hunter *really* good in only two seasons….

    Vote -1 Vote +1

  12. Tom says:

    Begs the question… would he have been as effective on a lesser team?

    Vote -1 Vote +1

    • Brian Cashman is Watching says:

      Funny you mention that. There has been this fight from Trevor Hoffman fans arguing Trevor Hoffman was better. The argument goes: Mariano Rivera pitched for the Yankees, a team that was always winning, so he got many more opportunities to save games, while the Padres generally had bad teams while Trevor Hoffman was pitching. However, that simply is not true.

      Using Baseball Reference for comparisons (Fangraphs does not seem to have early data on blown saves), Trevor Hoffman has 601 Saves out of 677 Save Opportunities, and 76 blown saves. Mariano Rivera had 602 Saves out of 674 Save Opportunities, and 72 blown saves. Someone else noted that “lesser” teams generally have more save opportunities, but I do not have that statistic to back it up.

      Vote -1 Vote +1

      • Mike says:

        I calculated this online

        Trevor Hoffman was a Padre from 1994 – 2008. He closed every year except 2003 when he pitched only a few games (also had 0 saves).

        From 1994 – 2008, Hoffman had 554 saves. The Padres in that time span (again excluding 2003) had won 1086 games. So almost exactly 50% of the Padres wins consisted of Hoffman getting a save.

        From 1997 – 2011, the Yankees have won 1460 games. Rivera has 597 saves in that time span. That would equal 40.9%

        So 41% of Yankee wins came with a Rivera save, 50% of Padres wins came with a Hoffman save.

        Wouldn’t that tell you that a vast majority of Yankee wins in that span were for the most part, 4 runs or more?

        Anyone saying he played for the Yankees got him more saves is wrong. Also as I mentioned, people also forget Hoffman missed almost a whole year, so he would probably have 640 career saves if he was healthy that year.

        Both pitchers averaged 39 saves per 162 games.

        I still think the save is a very overrated stat, but then again, so are W-L and we still cherish 300 Wins. Why not cherish 600 saves?

        Vote -1 Vote +1

      • Mike says:

        Actually I’m mistaken, Hoffman was on the Padres in 93 as well, but he wasn’t really their full time closer, so I’ll just leave it at 94 – 08

        Vote -1 Vote +1

      • Mike says:

        Actually let me put it this way since the 09 Brewers weren’t a good team either

        1994 was a strike season (not full time obviously so I won’t include that)

        From 1995 – 2009 (excluding 2003), Hoffman had 561 saves in 14 full seasons
        Rivera’s season isn’t over yet, but from 1997 – 2010, aslo 14 full seasons, Rivera had 559 saves. Only a 2 save difference, but Hoffman still had more

        Vote -1 Vote +1

      • Mike says:

        Ughh I wish there was an edit button and now I gotta post again lol, Rivera had 5 saves in 1996, therefore he had 554 saves from 97 – 10

        Vote -1 Vote +1

      • Jonathan says:

        “The argument goes: Mariano Rivera pitched for the Yankees, a team that was always winning, so he got many more opportunities to save games, while the Padres generally had bad teams while Trevor Hoffman was pitching.”

        The funny thing about that argument is the fallacy you didn’t point out:

        One could rationalize that the Padres would statistically offer more Save Opportunities for Hoffman because they weren’t always winning. The Yankees have had an elite offense for pretty much Rivera’s entire career, meaning that more often than not, they won by a crooked score that would leave Rivera without a save opportunity.

        People know full well you can’t get a save when you’re losing, but so few people seem to account for the fact you can’t when you’re steamrolling most of your opponents either.

        Vote -1 Vote +1

      • AMusingFool says:

        People know full well you can’t get a save when you’re losing, but so few people seem to account for the fact you can’t when you’re steamrolling most of your opponents either.

        Aw, c’mon… You haven’t seen the save in the 30-3 game before? (The case I’m thinking of was an O’s-Rangers game, but I forget which year, let alone which game.)

        Vote -1 Vote +1

  13. designated quitter says:

    I agree with the article, but on the other hand I was at Yankee Stadium this Labor Day and saw Mo come in with a two run lead, and give up one run and leave runners on 2nd and 3rd. He got a save. A Rivera save and a Gossage or Lyle save are not the same animal. The most valuable Yankee releiever this year is David Robertson, whom I suspect will inherit the closer role when Rivera is elected el Presidente of Panama.

    Vote -1 Vote +1

    • Dan says:

      Does this actually mean that one specific Rivera save is a different animal than another specific Gossage or Lyle save? Because that’s what it seems like.

      Vote -1 Vote +1

    • Jacob says:

      I believe this is what you call a logical fallacy

      Vote -1 Vote +1

    • Jonathan says:

      “David Robertson, whom I suspect will inherit the closer role ”

      So marks contestent number 453 in the last four seasons in the “Who Will Replace Mo” sweepstakes. Robertson needs to get his walk rate down a whole lot before he can consider stepping into Rivera’s role.

      Vote -1 Vote +1

  14. Barkey Walker says:

    Joe Nathan, 543.1 IP in relief, 2.23 ERA. His ERA+ as a closer is higher than Mo’s too.

    Vote -1 Vote +1

    • Mike says:

      Nathan’s ERA is 2.36 as a reliever. He’s already 36 years old so I don’t see him obviously matching Rivera’s inning total

      Vote -1 Vote +1

      • Barkey Walker says:

        I didn’t say that Nathan is a better player, only that his ERA+ as a closer is higher and that his ERA should be on the above table. To get relief, I subtracted the years where he started, what did you do?

        Vote -1 Vote +1

      • Mike says:

        Misunderstanding on my part, considering Rivera’s 96 season he wasn’t a closer so I’m sure that accounted for a lower relief ERA. Sorry about that.

        Vote -1 Vote +1

    • Mike says:

      And as pointed out, Rivera’s ERA in relief is 2.06. 2.22 is his career ERA

      Vote -1 Vote +1

  15. Devon says:

    I think the most telling & impressive stat about Mariano Rivera is one that is rarely ever talked about yet. His ERA in save situation is 1.91. And he did that through the steroids offense era!

    The only other relievers close to that ERA, are Joe Nathan and Hoyt Wilhelm, who are both sitting at 2.16 right now. Joe Nathan looks like he won’t last up there though. The next closest closer, is Sparky Lyle with a 2.39 ERA in save sit’s.

    If your job is to save the game, then what you do in save situations says everything. Mariano is way beyond them all.

    Vote -1 Vote +1

    • Jonathan says:

      “The next closest closer, is Sparky Lyle with a 2.39 ERA in save sit’s. ”

      Papelbon’s sitting at 2.27 in save situations right now, and that’s accounting for an awful 4.37 in save situations last year. And he’s only 30.

      Vote -1 Vote +1

    • OCP says:

      Hoyt Wilhelm’s ERA in save situations was 2.35 according to baseball-reference.

      Vote -1 Vote +1

  16. The Nicker says:

    I wouldn’t exactly say the SABR community is unanimously agreeing that Rivera is awesome (he is) when fWAR states that he’s been less valuable throughout his whole career than Kevin Tapani.

    Then again, this just reinforces the idea that fWAR should probably be recalibrated to use ERA for career WAR once a pitcher is over 1,000 innings. I think it’s safe to say that Rivera, with a .262 career BABIP and a partial career HR/FB of 6.2%, bucks a lot of trends and should be rewarded for that. Change his 2.76 to a 2.22 and he’s probably closer to 50 WAR than 40 WAR.

    Vote -1 Vote +1

    • JimNYC says:

      Thank you. I’ve been saying for years that, once you get the kind of IP Rivera has, it’s no fluke that his BABIP is so low and his HR/FB ratio is so low. He really can control that stuff — and it’s easy to see how just by watching him pitch.

      Vote -1 Vote +1

  17. Danny says:

    So, if a reliever comes in a tie game, it doesn’t matter what he does?

    Can you explain that to me?

    Vote -1 Vote +1

  18. Mike B. says:

    What jumps out at me is Mo’s 1996 season, his first full one as a pure relief pitcher: 61 G, 107 2/3 IP (!), 10.87 K/9, 0.08 HR/9 (!!), 1.88 FIP, 4.4 WAR.

    Allowed 1 HR in 107 2/3 innings pitched + 4.4 WAR for a reliever? Yikes.

    Vote -1 Vote +1

  19. Barkey Walker says:

    I always wondered, isn’t your ERA pretty seriously biased if you do a 4 out close? That first out, it is pretty difficult to get an earned run on you.

    Vote -1 Vote +1

  20. joecool says:

    You didn’t say if you thought he deserved in the hall or not. I guess I would put him there, but no other relivers.

    Mo surely “isn’t good enough to start” but he’s also better in a relief role than most starters would be. But better than HOF starters would be?

    I would put him in but only because his post season numbers are the extra bit he needs. In season numbers weren’t quite enough for me, and I would not put Hoffman in the hall.

    I am very down on the contributions of relievers though. Not that it’s “easy” to find good relievers from anywhere, but it’s much easier than finding starters.

    Vote -1 Vote +1

    • JimNYC says:

      There’s a logical fallacy in saying that just because a closer failed as a starter, a non-failed starter would do better as a closer. Mo Rivera wouldn’t have had the length to be able to play full seasons into his 40′s if he’d been a starter; that doesn’t mean somebody with a more durable arm who can start would be better able to limit run production in a more innings-reduced role.

      Vote -1 Vote +1

  21. Rusty says:

    All these stats are only counting the regular season. Don’t forget, he is also the best postseason pitcher ever:

    94 games, 139.2 innings, 2 homers, 0.71 ERA, 0.77 WHIP, 8-1 record, 42 saves.

    Add that to his regular season stats and he’s 83-58 with 644 saves and a career 1.91 ERA as a reliever.

    Vote -1 Vote +1

    • Black_Rose says:

      What’s is the second home run? I know that Sandy Alomar hit one of them in 1997.

      Vote -1 Vote +1

      • Mike says:

        Someone on the Mets did in the 2000 WS in Game 2, I think it was Jay Payton Yankees were up 6-0 in the 9th, hung on and won 6-5. Rivera was charged with giving up 2 ER. The only postseason game ever he allowed 2 ER.

        Vote -1 Vote +1

  22. psiogen says:

    The elephant in this article’s proverbial room is the fact that Rivera is a case study in the weaknesses of Fangraphs’ pitcher WAR. It’s plenty obvious at this point that FIP does not fully account for Rivera’s run-prevention skills. fWAR probably underrates him by as much as a full win per season. I love Fangraphs but y’all need to step it up on the pitcher value science.

    Vote -1 Vote +1

    • Kevin S. says:

      Weakness, or outlier? Mo has a fairly unique ability to skew FIP. Show me any pitcher with a comparable skill set and we’ll talk.

      Vote -1 Vote +1

      • JimNYC says:

        Either way, I think it’s pretty hard to say that — after a thousand games and nearly twelve hundred innings pitched — Rivera’s run suppression abilities not encompassed in FIP is just a sample size or luck issue.

        Vote -1 Vote +1

      • Kevin S. says:

        I completely agree, and I think most everybody in the analytic community would as well. The point is that given the singularity of Mo’s career, it’s hard to say he invalidates the system. By the way, rWAR, which is based off of RA with an adjustment for quality of defense, has him 53rd among all pitchers, ahead of HOFers like Whitey Ford, Sandy Koufax, Red Ruffing and Early Wynn.

        Vote -1 Vote +1

      • psiogen says:

        Many pitchers have over- or under-performed their FIPs year after year. Off the top of my head:

        OVER:
        Tom Glavine
        Billy Wagner
        Johan Santana

        UNDER:
        Javier Vazquez
        Glendon Rusch (who has the distinction of racking up 19.6 fWAR but only 3.2rWAR over a 12-year career)

        I’m sure those names are just the tip of the iceberg.

        Vote -1 Vote +1

      • The Nicker says:

        psiogen: Great point, I’m going to write a Community post about this. Glavine isthe best example of an ERA overperformer. Some other notables: Charlie Hough, Steve Trachsel, Tim Hudson, Jim Palmer were overperformers; Vazquez, John Burkett, Bobby Witt underperform their fWAR.

        Vote -1 Vote +1

  23. Jordan says:

    This tribute actually sells Mo short. He’s 5th all-time in WPA, including SPs, behind only Clemens, Maddux, Seaver, and Pedro. And – even adjusting for the fact that he pitched almost exclusively in high leverage situations – he’s 20th all-time in WPA/LI. Oh, and everyone ahead of him has thrown at least 2000 more innings, and WPA/LI is a counting stat.

    Perhaps most impressive of all is Mo’s career ERA+: 206. Next best is 154. And I haven’t even touched his post-season performance – he’s quite probably the best post-season pitcher ever: .71 ERA (best ever) in 139 IP (7th most ever).

    Vote -1 Vote +1

  24. Dan says:

    ERA with NO mention of FIP?

    Come on Slow, your better than that!

    Vote -1 Vote +1

  25. James says:

    I’m pretty sure that BABIP isn’t part of the fWAR equation. Right?

    Vote -1 Vote +1

  26. Jim schlichting says:

    All BB players know he is and was the best ever! That’s when you know his true value when selected by his
    Professional playing peers!

    Vote -1 Vote +1

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Current day month ye@r *