You Are Here > Baseball-Reference.com > Blog >

SITE NEWS: We are moving all of our site and company news into a single blog for Sports-Reference.com. We'll tag all B-R content, so you can quickly and easily find the content you want.

Also, our existing B-R blog rss feed will be redirected to the new site's feed.

Baseball-Reference.com » Sports Reference

For more from Andy and the gang, check out their new site High Heat Stats.

Does the W-L record matter for closers?

Posted by Andy on January 13, 2011

On our Trevor Hoffman Hall of Fame post, a couple of readers remarked on his career W-L record of 61-75 as being a detriment to his HOF credentials. This doesn't make a lot of sense to me, but let's explore the issue a little bit.

I've already written about how overrated W-L records are. Read that here first.

Closers have a hard time earning wins. Given that they usually enter a game when their team already has the lead, the typical way they can earn a win is:

1. Blow the save but remain in the game

2. Have their team take the lead while they still remain the pitcher of record

3. While they (or someone else) is pitching, have that lead stand up

Sure, this happens. It's called vulturing a win. But it's not all that likely. The other way that closers get wins is when they come in to a tie game, hold the score, and have their own team score to win the game.

Let's take a look at Hoffman's splits by situation for example.

I Split W L W-L% ERA G SV
in Wins 61 0 1.000 2.50 61 0
in Losses 0 75 .000 19.62 75 0
in No Dec. 0 0 3.37 298 0
in Saves 0 0 0.84 601 601
in Sv Situ 12 31 .279 2.71 696 601
in non-Sv 49 44 .527 3.15 339 0
Provided by Baseball-Reference.com: View Original Table
Generated 1/13/2011.

The two key splits to look at the Save Situations and non-Save Situations (the last two lines of the table). Let's talk about non-save situations first. These are games that were either tied, Hoffman's team was behind, or they were so far ahead that it didn't qualify as a save situation. I'd bet my bottom dollar that most of these appearances (which comprise about a third of his total appearances) came in tie games. As one might expect, Hoffman's decisions in these games are about .500 W-L% at 49-44. This is basically the same as a starting pitcher at the beginning of a game (when the score is tied at 0). Most starting pitchers finish with records fairly close to .500.

Now look at Save Situations. These are the cases where it's tough to get a win unless it's vultured. Unsurprisingly, Hoffman has a lot more losses than wins, because in the games where he pitches well, he gets a Save and not a Win.

Add these two situations together, and that's why closers usually have losing records.]

Mariano Rivera is a fairly big exception. His splits, in terms of how he was used, is fairly similar to Hoffman, but he's won a lot more than he's lost when pitching in non-save situations. In a similar number of non-save situational games as Hoffman, Rivera has 4 more wins and 14 fewer losses (and a similar fraction of no-decisions.) The biggest reason for this is probably the Yankees' strong offense, enabling Rivera to do better (in W-L terms) in tie games.

In save situations, Rivera has 4 more losses than wins, as compared to 19 more losses for Hoffman. The biggest difference here is probably that Rivera simply blows fewer saves.

Here's a handy list of closers....guys with at least 300 career saves:

Rk Player W-L% SV G GS CG SHO GF W L
1 Mariano Rivera .574 559 978 10 0 0 829 74 55
2 Billy Wagner .540 422 853 0 0 0 703 47 40
3 Rich Gossage .537 310 1002 37 16 0 681 124 107
4 Dennis Eckersley .535 390 1071 361 100 20 577 197 171
5 Robb Nen .517 314 643 4 0 0 549 45 42
6 John Wetteland .516 330 618 17 0 0 523 48 45
7 Rick Aguilera .515 318 732 89 10 0 557 86 81
8 John Franco .508 424 1119 0 0 0 774 90 87
9 Tom Henke .494 311 642 0 0 0 548 41 42
10 Rollie Fingers .491 341 944 37 4 2 709 114 118
11 Bruce Sutter .489 300 661 0 0 0 512 68 71
12 Jeff Reardon .487 367 880 0 0 0 695 73 77
13 Roberto Hernandez .486 326 1010 3 0 0 667 67 71
14 Todd Jones .479 319 982 1 0 0 619 58 63
15 Jeff Montgomery .469 304 700 1 0 0 549 46 52
16 Doug Jones .466 303 846 4 0 0 640 69 79
17 Trevor Hoffman .449 601 1035 0 0 0 856 61 75
18 Troy Percival .449 358 703 1 0 0 546 35 43
19 Lee Smith .436 478 1022 6 0 0 802 71 92
20 Jose Mesa .423 321 1022 95 6 2 633 80 109
21 Randy Myers .411 347 728 12 1 0 548 44 63
Provided by Baseball-Reference.com: View Play Index Tool Used
Generated 1/13/2011.

Rivera is, as usual, an outlier. A lot of Eckersley's wins came as a starter (although even as a reliever he had a .539 W-L%).

Overall, though, you can see that most of these guys tend to be a little under .500. Yeah, Hoffman has a worse W-L% than most, but it's not wildly atypical. The sum of all W and L in the table above is 1538-1583 (.493), and that's before considering what these guys did as starters. Aguilera, for example, was 40-29 as a starter.

I just don't feel like wins and losses for closers are all that meaningful...what do you think?

This entry was posted on Thursday, January 13th, 2011 at 10:22 am and is filed under Splits. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.

86 Responses to “Does the W-L record matter for closers?”

  1. I don't think win-loss records matter for anybody, but especially not a relief pitcher.

  2. I think it goes without saying that generally W/L records don't matter and they matter even less for relief pitchers and even less than that for closers, but I do think on a season-by-season basis (not necessarily career), "losses" can say something about a reliever generally and a closer specifically. Not wins, only losses. A reliever that has a lot of losses generally is blowing leads or tie games that were passed on to him. Very few relievers regularly pitch more than two innings at a time in the modern game, nor have they for years, so losses do a decent job of isolating a massive failure on their behalf. It still isn't the tool that I'd look to first to evaluate a relievers performance, but it matters a little I'd say. Wins, not so much.

  3. I agree with Larry Smith completely. Losses by a reliever is underlooked. In general W-L is overrated, but I don't think losses by relievers (or at least closers) is overrated.

  4. Closers are often used in tie games at home once the possibility of a save is gone (innings 9+ where a home win requires a walk-off). I don't know how often that happens.

  5. Wins are certainly not the definitive stat for a pitcher. However, it seems to be that a starter has to win at least 250 games to get in the HOF. With that said, Tommy John is one of the few pitchers in top 50 WAR who isn't or isn't slated to be voted in the HOF. I think its time the HOF should be looking at TJ through the same lens as Blyleven.

  6. To the extent that "losses" matter in the way Larry suggests, "wins" are almost exactly as bad. At least for a closer. They'll usually mean you came into the game with the lead, gave it up, and then your offense came back and bailed you out.

    I'd much rather just ignore them both completely (and saves and blown saves) and look at how well the guy actually pitched in the big picture, but if losses are a negative for a closer, so are wins.

  7. People just so cavalierly say "W-L is irrelevant". It cracks me up.

    But, where relievers are concerned I couldn't agree more. They get thrown into these sudden-death showdowns, and so many things get weird in the ninth inning.

    The offense wins it in the bottom of the ninth after you pitched a single inning, and you get the win? Crazy. And there's a million crazier scenarios.

  8. Frank Clingenpeel Says:

    Even before the advent of the modern closer, won-loss records for relievers were at best a secondary stat. Remember Darold Knowles, a pitcher for the second Washington Senators who posted a sub-2.oo ERA, yet had a 2-14 won-loss record.

  9. Wait..what?

    So, we're talking on another thread about the qualifications for closer's as HOFers, and the ONE thing they're responsible for each game....doesn't matter?

    (Slaps self in forehead in disbelief).

    W/L may not matter for a starter, but it ABSOLUTELY matters for a closer.

    If you take away Hoffman's non save situation games, he has the same AMOUNT of saves, but his record is much worse.

    IMO, Andy, you just provided inarguable proof Hoffman is NOT a HOFer.

    These numbers are from before the 2008 season, and I don't have the time now to update them, although I doubt there would be any changes.

    Dennis Eckersley, Hoffman and Rivera have a combined total of eight saves of seven inings or more, with Eck having five of them.

    Gossage had 52..out of 310 career saves.

    While I'm not a big fan of Gossage in the Hall, it's clear he is more deserving than either Rivera or Hoffman because his saves/opportunities were more meaningful.

    Ask yourselves something;

    If Mariano Rivera had pitched the last sixteen seasons with the Cubs, and had posted EXACTLY the same numbers, but had no rings or Yankee mystique, would he be a HOFer?

    No chance.

  10. There's also the matter of closers pitching on the road. Unlike any other pitcher in any other situation, a blown lead on the road in the bottom of the ninth equates to a loss.

    It wouldn't surprise me to find that Hoffman pitched more on the road than other closers.

  11. #10, let's check:

    Appearances:

    Rivera: 505 G at home, 473 away
    Hoffman: 573 home, 462 away
    Wagner: 460 home, 393 home
    Percival: 378 home, 325 home
    Mesa: 531 home, 491 away

    So home percentage:

    Hoffman 55.4%
    Wagner 53.9%
    Percival 53.8%
    Mesa 52.0%
    Rivera 51.6%

  12. so, Hoffman actually pitched more at home than other top closers. My list of "top" closers is totally arbitrary--I just tried to pick the most recent guys from my list in the original post.

  13. "If Mariano Rivera had pitched the last sixteen seasons with the Cubs, and had posted EXACTLY the same numbers, but had no rings or Yankee mystique, would he be a HOFer?"

    If Trevor Hoffman makes the hall of fame, then yes.

  14. Johnny Twisto Says:

    W/L records *matter* for all pitchers. They're just not a great indicator of pitcher quality for starters, and almost useless for relievers.

  15. It doesn't surprise me that closers pitch more at home than away. For the most part in tie games in late innings, a home team's closer comes in to keep the game tied. And for the most part, an away team closer doesn't often come in the game unless his team is in the lead.

    There are some games of course where an away closer will come in late to keep the game tied as well. But more often than not, away closers usually only come in when their team is ahead.

  16. Johnny Twisto, I disagree. Barely anybody pays attention to a reliever (well msot notably a closers) W-L record. I assure you half the wins, especially for closers come at the expense of them blowing a lead only to see the home team win the game for them int he bottom half of the inning. That's another stat I'd be interested in trying to find out.

    Since Rivera and Hoffman don't have a ton of wins, perhaps it's not difficult to look at their Win game logs and see if they blew a save in those games.

  17. Johnny Twisto Says:

    We can see on the pitcher pages how often they came into games when ahead, tied, behind. During his closing years of 1994-2009, Hoffman entered when ahead 76% of the time, tied 16%, behind 8%.

    Rivera, from 1997-2010, has been ahead 83%, tied 12%, behind 5%.

    The next question would be whether their respective managers are actually choosing to use Hoffman more often in those tie games. Since the Yankees have been good to great ever since Rivera arrived, and the Padres were up and down, I'd assume there were just more tie games where the option existed to put in Hoffman than there were for Rivera.

  18. Johnny Twisto Says:

    Mike, I understand that. I just said that they are practically meaningless for evaluating relievers. I just disagree that they don't matter. They are, after all, wins and losses, the whole point of the game. Regardless whether they've been credited to the appropriate person.

  19. There are too many pitfalls/variables with W/L record as far as closers are concerned to give it much credence.

    Say a closer comes in to a game with a 2-0 lead in the top of the 9th and gives up 2 runs. In the bottom of the 9th his team scores 1 run and the closer gets the Win?

    Or say an even more extreme scenario where the closer comes into a 3-0 game in the top of the 9th and gives up 4 runs. In the bottom of the ninth the Home team scores a run and the closer is removed in extra innings and gets a no decision.

    A closer comes into a game 2-2 in the 13th and records the final out of the inning and then the home team scores a run in the bottom of the 10th and he gets a win for recorded 1 out in a 13 inning game?

    A closer comes into a game in the 9th with the score 5-5, man on third no outs. He strikes out the first two batters and then the third baseman makes an error on a ball hit by the third batter and a run scores. The closer strikes out the last batter of the inning and then the home team doesn't score in the bottom of the 9th and the closer gets the loss.

  20. Damn, JT, I spent all morning trying to figure out where that data was....I became convinced it wasn't available, especially after I asked Neil and he didn't know either.

    Here's Rivera's for example:
    http://www.baseball-reference.com/players/r/riverma01-pitch.shtml#pitching_reliever::none

    Columns on the right show # of games entered when behind, tied, or ahead.

  21. From what I have seen on other threads on this site, it seems that many of us are not in love with W-L record for starters even...So for closers I think it is even less meaningful (unless you look at splits as Andy does above), but even within each split, the team offense has a lot to do with the the W-L records (even moreso than a starter because we are only talking about 1 or 2 innings - i.e. better offenses score in more innings making the probablility higher that the offense will score at the correct time to aid the closer in getting a win)

    In other words, W-L really doesn't mean that much even if you compare only closers.

  22. Johnny Twisto Says:

    Per Mike's suggestion, I checked the game logs. Looks like 12 of Hoffman's wins followed blown saves, and 17 of Rivera's.

  23. One note about Rivera and how Torre/Girardi have tended to use him...

    He always pitches the 9th inning of a tie game at home (because there cannot be a save situation).

    Typically he would not pitch until it was a save situation (not come into a tie game) on the road (because of the automatic 2 innings required).

  24. Thanks Johnny.

    I think W-L has to be adjusted, at least for the loss side. They should do it like how Game Winning Goals are done in the NHL. Let's say the final score of a hockey game is 5-3 Penguins over Flyers. Whoever got the 4th goal for the Penugins gets the game winning goal in the book. Makes sense right?

    Can't they do the same for losses for a pitcher and maybe adjust it and the loss to a pitcher who gave up the game winning run? Ex. Dodgers beat Giants 4-3 (Dodgers are the road team). But let's say the score is 1-0 going into the top of the 9th. The starting pitcher went 8 innings, gives up the the 1 run, the Dodgers tack on 3 in the 9th. Then the Giants get 3 of their own in the 9th but fall short 4-3. Do you think a starting pitcher is more deserving of a loss than the relief pitcher(s) who gave up 3 runs in the 9th?

    Now of course it's hard to re-write history, but I don't see how that wouldn't work. Now in terms of who gets the W in the game (let's say the Giants won 5-4 instead of losing 4-3), that might be more difficult of a task. Because it's just as bad to give the relief pitcher a W when he gives up multiple runnings in only 1 inning of work.

  25. John Q, just for the record, you wrote:

    :A closer comes into a game in the 9th with the score 5-5, man on third no outs. He strikes out the first two batters and then the third baseman makes an error on a ball hit by the third batter and a run scores. The closer strikes out the last batter of the inning and then the home team doesn't score in the bottom of the 9th and the closer gets the loss."

    Unless I'm missing something, you are incorrect. If the closer entered the game with a runner on third, when that run crosses the plate the run will be charged to the previous pitcher who allowed said runner to reach base. The previous pitcher will get the Loss. If that runner reached base AFTER the closer entered the game, then regardless of whether or not he ultimately scores on an error, yes the loss would then be charged to the closer.

  26. LudaCris, to add on to what you said about John Q.

    "A closer comes into a game 2-2 in the 13th and records the final out of the inning and then the home team scores a run in the bottom of the 10th and he gets a win for recorded 1 out in a 13 inning game?"

    This is not necessarily true. An official scorer is allowed to determine who is credited with a win. If he feels there is/was a pitcher who pitched more effectively, they can be credited with a win upon his decision.

    So let's say the closer replaced a pitcher who pitched 2.1 shutout innings. An official score, if he chooses (in this case I would), is allowed to actualy give the W to that pitcher even though he didn't get the final out.

  27. Mike,

    While I do not have the MLB rulebook in front of me to confirm/deny that the official scorer makes a determination as to the winning pitcher, in practice today this is rarely used (if true).

    I will say that playing years and years of fantasy baseball and looking closely at box scores, the pitcher who recorded the final out is the pitcher of record and the pitcher of record always (if not, almost always) receives the Win. So for all intents and purposes, he is correct on that regard.

    I am of the firm belief that W-L record is one of the least important categories for any pitcher. For the team it is the only important category. For individuals, it doesn't come close to telling the full story. It should be factored into Hall of Fame consideration for starting pitchers, but for Trevor Hoffman should not even be a question.

  28. LudaChris, yet it is indeed rarely used, but the official scorer can indeed make a decision on it if he feels another pitcher pitched better than one who would qualify as getting the W.

    I know Wikipedia isn't the most trustworthy source, but this is what it shows on it. I'm only using it here because I'm at work and don't want to download the officiail rule book in the PDF format here lol

    "In Major League Baseball, the winning pitcher is defined as the pitcher who last pitched prior to the half-inning when the winning team took the lead for the last time.

    There are two exceptions to this rule. The more common exception is that a starting pitcher must complete five innings to earn a win (four innings for a game that lasted five innings on defense). If the starting pitcher fails to meet the innings requirement, the official scorer awards the win to the relief pitcher who, in the official scorer’s judgment, was the most effective.

    The second exception applies only to a relief pitcher who makes a "brief appearance" and is himself later relieved. If, in the official scorer's judgment, the relief pitcher was "ineffective," the win is awarded to the succeeding relief pitcher who was most effective, in the official scorer's judgment"

  29. I do agree it is more of a team stat than a individual pitcher stat and is overrated for the most part. However like the article mentions, I do not think that Losses for a closer are overrated

  30. Johnny Twisto Says:

    I think that second exception was only added to the rules a couple years ago.

  31. Andy, the link that you put in the article does not work. The one where you put to read Here about why W-L is overrated.

  32. Interesting stuff... It's good to know that sound judgment can be used in determining the winning pitcher, but like we both acknowledged, it is rarely (virtually never) used. There are times when the official scorer's judgment seems unreasonable as well, so maybe it's for the best that for stat-keeping purposes we just award it to the pitcher of record.

    It is also fair to say that losses, for a relief pitcher, should be considered. Given Hoffman's 1,035 game appearances, and the fact that you are only appearing in games where the decision is in reach (+/- 3 runs), 75 career losses should not even come close to keeping him from the Hall.

  33. @26,

    I can't think of a time that the official scorer ever chose the winner of a game for the scenario (or a sililar one) that you outlined.

    Typically the O.S. would only choose the winner in a situation where the Starting pitcher would have gotten the win if not for the 5 inning requirement.

    An example of this is where the Yanks jump out to a 7-0 lead going into the bottom of the 5th inning when AJ Burnett gives up 5 quick runs and gets pulled for a reliever before the 5th inning ends. But then the game ends 7-5. AJ Burnett is not eligible to get the win because he did not finish 5 innings. Therefore the official scorer would choose the most worthy (for lack of a better term) relief pitcher (which also would probably not be the guy that finished the game if he gets a save).

    A don't think the O.S. ever steps in and changes the eligible winner (by rule) just because someone else might have been more worthy.

  34. One related item, IMO, DOES matter.

    Save percentage. I think save percentage matters more for a closer than W-L percentage for a starter.

    A closer usually comes in with the game ahead, and faces nobody in the opposing lineup more than once. If you blow that, it is serous.

  35. To 34:

    I agree with you on save percentage. To add to this, I am always curious to see splits on how good a closer's save percentage is with a 1-run, 2-run, and 3-run lead. When we're talking about a potential Hall-of-Famer like Hoffman, the law of averages would keep this constant with other Hall-of-Hamer relief pitchers. But in evaluating one performance over a couple of seasons (like if considering to sign him as a free agent), I would like to see some splits to dig deeper.

  36. #33, such occurences has happened with the Yankees this year. Twice as a matter of fact. (If not twice this year then twice the past 2 years) It was a game where the Yankees gave up the lead in the 7th, then re-took the lead in the bottom of the 7th. Kerry Wood and Mariano finished if off. I believe Joba gave up the lead. Joba would be credited with a W but the official scorer gave it to Kerry Wood.

    I could have that box score wrong, but I know there have been instances when such has happened. I will try to find it.

  37. B-Ref nows has a stat that is remarkably useful for relief pitchers called Win Probability Added (WPA). Traditional "Wins" and "Losses" take the entire stream of the game and rather arbitarily assign responsibility for the entire team's win or loss onto the back of one single pitcher. That can make some sense most of the time for starters, who on average pitch enough innings during a game that, at least more often than not, if you are going to assign the "Win" or "Loss" to a pitcher it will make some sense to assign it to the starter. That's why Win-Loss records for starters often (not always, but often) accurately tell us a lot (not everything, but a lot) about the quality of the pitcher.

    But for relievers, who often face just a handful or fewer of batters, to assign the entire team win or loss the one pitcher, based on the strict, narrow rules about who gets "assigned" the win or loss, turns out to make very little sense and tells us almost nothing about the quality of the pitcher.

    WPA does something similar to what traditional "Wins" or "Losses" are intended to do, but much more carefully reflects each pitcher's actual contribution to improving or reducing his team's chances of winning the game. WPA looks at the probability the pitcher's team had of winning the game (based on the score, inning, outs and men on base) before each plate appearance against the pitcher and then, after each such plate appearance, adds or deducts the increase or decrease in that probability resulting from that PA to the pitcher's total "Win Probability Added" .

    So, for example, when a closer enters the game at the begining of the 9th inning, if his team is a head by one run, history suggests that his team will win the game about 85% of the time. If the closer then successfully closes out that one-run win for his team, he has increased his team's win probability from 85% when he entered the game to 100% when he has finished (an increase of 15%, or 0.15). So he gets 0.15 of Win Probability Added for that game. If he had come in with his team ahead by three runs going into the bottom of the ninth, he would only get 0.03 of Win Probability Added because when he entered the game his team already had a 97% chance of winning.

    On the other hand if the closer comes in with his team ahead by one run at the start of the bottom of the ninth and he gives up a single and a homer to lose the game, he will have 0.85 subtracted from his Win Probability Added, because his team has gone from an 85% chance of winning when he entered to 0% when he finished. This is a much, much, much more accurate system for allocating wins and losses to relief pitchers than the traditional system of "Wins" and "Losses".

    According to the Play Index, the all-time top five career Win Probability Added totals for relief pitchers (at least 80% of their games pitched being relief appearances) are:
    1. Mariano Rivera +50.8
    2. Trevor Hoffman +34.4
    3. Goose Gossage +32.6
    4. Hoyt Wilhelm +30.1
    5. Billy Wagner +29.3

    The gap between this group and the sixth place guy (Joe Nathan) is pretty large, and I think a very good argument can be made that these five guys represent the five relief pitchers most deserving of a Hall of Fame spot.

    By the way, the career WPA numbers above cover only regular season games, so don't even take in to account Mariano's extraordinary post-season performance.

  38. Found one right here. I don't know if I can post the link on here but I can tell you the date and game. September 3, 2010 Blue Jays at Yankees. Kerry Wood gets the W

    http://www.baseball-reference.com/boxes/NYA/NYA201009030.shtml

    Ivan Nova the starter pitched 4.2 innings, (Yankees ahead 5-3 at the time). Ivan Nova game in and pitched 2/3 an inning. Technically he would be credited with the W.

    David Robertson pitched 1 inning
    Kerry Wood pitched 1.2 innings
    Mariano Rivera pitched 1 inning.

    The win this game went to Kerry Wood.

  39. Correction on my stat in that game, Boone Logan pitched the 2/3 inning after Ivan Nova. I put Nova's name twice, sorry about that.

  40. LudaChris @35 ... some stuff related to that is available under the "more stats." Not the number of runs in the lead, per se, but number of game where the closer came in with either high, medium or low leverage, and a couple of other things.

  41. So as mentioned, the "eligible" winner would have been Logan, but the scorer gave it to Kerry Wood.

  42. Oh, my thoughts on how to judge if a closer is HOF worthy or not: http://socraticgadfly.blogspot.com/2011/01/how-do-you-judge-if-closer-is-mlb-hof.html

  43. Good find Mike, but 33 was already agreeing with you than the official scorer might jump in and select the "most-deserving" relief pitcher when the starter fails to go 5.0 IP although that team already has the lead and never loses it. This would be Exception #1 that you posted earlier.

    He is looking for a scenario that goes along with Exception #2, which I think we are all in agreement that it is generally not used. It is possible, but I cannot recall ever seeing it.

  44. Johnny Twisto Says:

    Rivera has added another +11.6 WPA in the postseason.

    Gossage was +0.8 Hoffman was -0.8. Wagner -0.9. Armando Benitez -1.1.

    Rollie Fingers +2.5. Dennis Eckersley +0.7 (as a reliever).

  45. @44 Johnny Twisto - where do you find postseason WPA out of curiosity? I'm wondering since Rivera is #6 all-time amongst pitchers in regular season WPA, where he would rank adding postseason WPA to the mix.

  46. Johnny Twisto Says:

    I think the only way you can get it is to sum up the gamelogs. Unfortunately I don't believe there is a leaderboard or a way to search it on PI.

    Rivera moves to second (since 1950) when including postseason. The guys ahead of him don't add enough.

  47. #31 Mike, thanks. Fixed. I can't believe we got to 46 comments before someone pointed that out.

  48. Mike @ #38

    The rule you're referring to, which your link confirms, ONLY applies when the starting pitcher does not pitch the necessary five innings to become the pitcher of record.

  49. WHAT?! Are you serious?! The role of the closer is to save the game and not lose it, and if they're in the position to win it then that's what they should be doing. Yes I agree the W/L record is overrated but that doesn't take the closer off the hook either. A lose is a lose no matter who's pitching. Espacially for a closer who gets paid millions to maintain the win in the first place!

  50. @46 Johnny Twisto - thanks for the check on that. That's pretty interesting.

    #1 in ERA+
    #1 in RA (Runs Allowed)
    #2 (since 1950) in overall WPA

    That's HOF-worthy to me!

  51. Another stat that I find useful in evaluating relief pitchers is the OPS+ of the batters that the pitcher faced. You can find this stat on PI. Lowest all-time career OPS+ of batters faced for pitchers with at least 500 career relief appearances:
    1. Mariano Rivera 45
    2. Billy Wagner 48
    3. Francisco Rodriguez 57
    4. Troy Percival 58
    5. Joe Nathan 61
    6. Tom Henke 63
    7. Hoyt Wilhelm 66
    8. Trevor Hoffman 67

    This is a rate stat, so active players may tend to drop some on this list as declining performance level in their later years drags them down a bit.

    Note that the OPS+ for hitters vs. Mariano means that the average hitter who has faced him across his career has performed against him at a level no better than what is the normal performance of a good-hitting pitcher (Don Drysdale and Ted Lyons each had a career OPS+ of 45 as hitters).

  52. A game where the official scorer decided who would get the win. June 23,1934; Dodgers in St. Louis. The Cards starter gives up 2 runs and is lifted after 3 innings. Their next pitcher allows 1 run in 2 innings. The 3rd hurler lasts 1 inning and gives up 1 more run. Middle of the 6th, down 4-0. The Cards offense explodes in the bottom of the inning for 5 runs. Jay Hanna Dean comes in to start the 7th and finishes the game. No more runs are scored. The official scorer gave the win to the Dizzy one even though the pitcher of record when the team got the lead was was a reliever, not an ineligible starter. This was important historically because Dazzy's brother ended the season with 30 wins, and may not have made the HOF without that numerical milestone.

    Say a team wins 5-3. Who gets the win depends upon the timing of the scoring. Each pitcher is responsible for not allowing runs, but scoring them is up to the offense, not the pitcher. A reliever comes in 6th inning or later and finishes the game. He cannot get a win in any circumstances unless the offense scores while he is in the game. If they are tied or behind, of course they need runs to take the lead. If they are ahead, it will be a save unless the pitcher blows it and then is bailed out by the offense scoring. So essentially all reliever wins are in a large part luck, did the offense score at the right time to give you, not one of your teammates the win.

  53. Kds @ 52,

    Nice job. You are quick at scouring those game logs! :)

    I am not sure what the rules were in the 30's but I really don't think an official scorer would do that today. Even if it were an HOF-type pitcher getting the win in place of a middle reliever. But I guess if as JT says above, they added the rule recently, then maybe the plan is to make similar decisions in the future.

  54. I am trying to figure out if WPA favors closers too much. I was looking at the overall list of WPA leaders (pitchers).

    Rivera slides in OK between Johnson and Palmer, but I have a problem with Hoffman slotting between Marichal and Koufax, but I guess Hoffman is bound to compile given his long career as a quality pitcher. Then the next several relievers on the list (Gossage, Wilhelm, Wagner, Eck) are OK in their positioning compared to equivalent WPA starters.

    But then once you reach about 50th on the list, there is a flood of closers (not all of them seemingly that great). From #50 through #70 there are way more than half closers.

    So just using my quick assessment, it seems like for the most part WPA is pretty accurate for the real stars (top 50), but after that it seems to benefit closers more than starters...maybe it is just me. But I guess that is good for people wo claim that the closer is truly an important position.

  55. Johnny Twisto Says:

    Well, it does favor closers in the sense that they pitch at times of the game when the chances of winning can swing most dramatically. That's exactly what the stat is measuring. If you want to remove that leverage aspect, look at WPA/LI. There, Rivera ranks 25, Hoffman and Wagner are at 50.

  56. @54

    Of course WPA favors closers. The end of a close game is a higher-leverage situation than any other, so the Leverage Index in those situations is huge. Likewise, you'll find that a career pinch-hitter likely has a much higher Leverage Index for his plays than a regular position player. So, if a closer pitched a similar number of innings to a starter with similar results, his WPA will be MUCH higher than the starter's. The thing is, though, WPA is intentionally measuring with that bias in mind.

  57. #43 and #48, thanks guy.

    Andy, no problem.

  58. Guys, not guy.

    #52, excellent find.

    And I just realized I'm posting on my other name now haha.

  59. @54: I tend not to use WPA for starters, or to compare relievers and starters. I prefer to use it to compare relievers to one another. One thing about WPA as currently calculated is that it, if I understand it correctly, is calcluated based on a comparison to "average" rather than to "replacement level". So a starter who pitches 8 innings and just keeps his team at the same 50% chance of winning that it had when he entered the game will get zero added to and zero subtracted from his WPA. But that doesn't really reflect his contribution as compared to a "replacement level pitcher" who would have left his team over the same eight innings worse off in terms of win probability. In other words, I don't think starters get full credit for the large number of merely above-replacement level innings they throw. That I think is one reason you see a bunch of relief pitchers with WPAs higher compared to starters who intuitively seem like they contributed more over thier careers. to solve this problem, I would supplement any use of WPA in comparing relievers and starters with Wins Above Replacement (WAR), under which starters will (appropriately, I think) rank higher.

  60. @ 9

    Chuck... It's hard enough to judge players based on what they do. Let's not get into "'what if Rivera played for the Cubs." HOF votes can't be based on revisionist history like that. And for you to say "no chance" Rivera is in if he had the same numbers playing for the Cubs is ridiculous. His stats are staggering.

    Also, Rivera's playoff and WS numbers can't be disqualified as part of your HOF/ what if question. This isn't an MVP/ CY Young vote where only regular season counts. Anything players do in October goes on their resumes and helps HOF eligibility.

  61. I think the answer to Andy's original question is, "No."

    Under the modern closer usage pattern, the closer can only get a decision when he fails to hold the lead. And if they do blow the lead, their only chance to get a win is if they do that in the top of 9th (i.e., at home), or if on the road, they only allow the tying run and then pitch another inning.

    Throw in the inherent randomness of W-L records, and I can't see any value to them for modern closers.

    P.S. I posted without having had time read the prior posts, so I'm sorry if this is redundant.

  62. P.S. to my #61: I don't think even losses are meaningful for a closer, unless it's a really big number. Ordinarily, closers get just 6-12 decisions per year. That's such a small sampling that any closer's season and even career record is bound to contain more randomness than meaning.

  63. Is there a way to quickly find: Out of a closers total save opp. appearances, how many or of what percentage of 'losses' came when blowing a 2 or 3 run lead?

    Would be interesting to see who blew the freebies more often.

  64. #59: "One thing about WPA as currently calculated is that it, if I understand it correctly, is calcluated based on a comparison to "average" rather than to "replacement level". So a starter who pitches 8 innings and just keeps his team at the same 50% chance of winning that it had when he entered the game will get zero added to and zero subtracted from his WPA."

    This actually is not true; a starting pitcher can accumulate a lot of WPA in such games. For a great example, check out the Travis Wood-Roy Halladay pitching duel from last season, in which each pitcher threw nine shutout innings and neither factored in the decision:

    http://www.baseball-reference.com/boxes/PHI/PHI201007100.shtml

    Note that both Wood and Halladay earned 0.692 WPA in the game--extremely impressive for a single game, and reflective of how well each guy pitched. You can follow the changed in WPA in the play-by-play section to see how it adds up.

    Also note that Felix Hernandez led all MLB pitchers in WPA last season (and by a half-win margin, too) despite just a 13-12 record.

  65. Just to follow up my last post and explain how it works: at the start of the game, it's 50-50. But if the home pitcher throws a scoreless top-of-the-first, he has slightly increased his team's chance of winning because, even though the game is still tied, the other team now only has eight more chances to score, whereas his team still has nine. So, the visiting pitcher now inherits a situation in which his team has slightly less than a 50-50 chance of winning. But if he pitches a scoreless bottom-of-the-first, he brings it back to 50-50 because now both teams have the same number of chances left to score. So, in a game that keeps going back and forth like this, the WPA for the pitchers starts to add up, even as the game remains tied.

  66. Johnny Twisto Says:

    Anon #63, I can't see any easy way to do that. In searching around, I did look this up. Interesting, but I'm not sure if it's meaningful. Percentage of successful saves in which runs were allowed:

    Rivera 7.9
    Hoffman 8.8
    Wagner 7.6

    That only includes runs charged to the closer, not inherited runners.

  67. DavidJ #65, good explanation. Let me add this; a low scoring close game will give most of the WPA to the pitchers. A high scoring close game will give most of the WPA to the hitters.

  68. DavidJ @64-65: I did not mean to suggest that the two pitchers who pitch matching shutouts do not generate lots of positive WPA. You are absolutely right that such pitchers will indeed get lots of plus-WPA credit even though they leave them game at the same 50-50 win probability that they started. Halladay will got lots of WPA credit when he pitches in a 0-0 game because the Phillies hitters by failing to score ininning after inning are offsetting Halladay's positive WPA with negative WPA of their own (leaving the team overall at the same 50-50 they began with).

    When I referred to an "average" starting performance generating zero WPA, what I mean was (and I was not as clear as I should have been) a pitcher who gives up an average number of runs in a relatively average circumstances (say a pitcher who gives up four runs in eight innings in a 4-2 or 6-4 game). That pitcher has performed better than a replacement pitcher would likely have performed but will get zero WPA credit (his team's probability of winning will not have changed from the original 50-50 start due to his efforts -- any change will be the result of the his team's hitters positive or negative WPA). It's because players are not given WPA credit for that portion of their performance that is above-replacement but below or at average that players with a lot of playing time are not valued by WPA as highly as they should be if you are comparing them to players with less playing time. So starting pitchers suffer by comparison to relief pitchers in WPA. In your example, Halladay received 0.692 in WPA for his fine pitching perfomance, but that compares him to what an average pitcher would have done that day. If you compared him to what a replacement level pitcher would have done, he would have received more than 0.692, and its that missing increment that affects starters more than relievers.

  69. @66 - cheers. There's always meaning in something to someone ;)

  70. Dropping back to #30 and others:

    The "brief" and "ineffective" exception dates back to 1950, also the year when the 5-IP rule for starting pitchers went into effect. "A win will not be credited to any relief pitcher who pitches briefly and ineffectively. In such a case the official scorer shall credit the victory to another relief pitcher."

    In 1978 it was clarified to "A victory will not be credited to a relief pitcher who is ineffective in a brief appearance when a succeeding relief pitcher pitches effectively in helping his team maintain the lead. In such cases, the succeeding relief pitcher will be credited with the victory."

    That's very close to the current language of the rule.

    Source for the above 2 quotes is the 1993 MacMillan Encyclopedia.

    So it has been around for quite a while but doesn't happen that often. There probably aren't a lot of situations where it would be applicable.

  71. Johnny Twisto Says:

    Huh, I could have sworn that was a change made very recently. I wonder if it was just slightly revised again recently. Will have to look into this.

  72. Since I was curious I checked the pitching lines in Rivera's 8 postseason wins. It's some late-night note-taking so there might be some minor errors.

    8 Wins. 16.1 IP with only two one-inning games. 10 hits. 14 Ks. 0 walks. Oh, and a 0.00 ERA for zero vultured wins.

    I have to think that counts for something.

  73. Oh, and I'm not sure that if Mariano had been on the Cubs his whole career that would necessarily mean he'd have no rings.

  74. @10, Jon: "a blown lead on the road in the bottom of the ninth equates to a loss."

    Except when the blown lead leave the game tied.

  75. Johnny Twisto @14: "W/L records *matter* for all pitchers. They're just not a great indicator of pitcher quality for starters, and almost useless for relievers."

    @18: "They are, after all, wins and losses, the whole point of the game. Regardless whether they've been credited to the appropriate person."

    Johnny, I think this is the first time I've been unable to follow your reasoning. I've read these over several times, but it still seems like you're directly contradicting yourself in both instances. So I think I must not be getting your meaning.

    In particular, the second line I quoted seems to be mixing apples and oranges. Sure, wins and losses are the whole point of the game -- as team events. But those wins and losses are very different thing from "wins" and "losses" as decisions on the records of individual pitchers.

  76. Followup to my #75: Perhaps what's confusing me is the meaning we're attaching to "matter." I'm using it in the sense of, "are meaningful," which is how I took Andy to mean it, based on his closing line:

    "I just don't feel like wins and losses for closers are all that meaningful...what do you think?"

    So, while W/L records for closers might "matter" in the sense of salary or award votes or something, I think that, with very few exceptions, they aren't meaningful for closers.

  77. Johnny Twisto Says:

    Sorry for being unclear. I posted 14 after seeing Bark's 7: People just so cavalierly say "W-L is irrelevant". It cracks me up. I guess a lot of my posts here are attempts to bridge the so-called divide between saber-inclined fans and traditionalists, either by explaining what the new stats mean and measure, or showing how the sides are not necessarily so far apart. I thought it was great on that other thread how Bark seemed to have his mind blown by the concept of BABIP and wanted to read more about it. But then he sees everyone writing here that "wins and losses don't matter" and it seems like the kind of thing which will drive him back away. So I was crudely trying to say that of course they matter, on the team level. And since team wins and losses are credited to individual pitchers, the pitcher's record matters, in the sense that it reflects how well the team has played. Is it an important component in evaluating how well the pitcher himself has performed? No, especially for relievers.

  78. #68 Birtelcom:

    Ah, ok, I see what you're saying. Sorry I misread your post. And I agree: WPA is a nice stat for telling the "story" of a game (or even a season)--really, it's a great measure of how well a pitcher "pitched to the score"--but you're right that it's not a great tool for comparing starters with relievers, in that it doesn't account for the extra value that starters provide in their ability to pitch a lot of innings.

  79. DavidJ: I'd even go further than arguing that WPA is good for telling the story of the game or season and that it is a good, objective measure (not the only one, but an important one) of the performance by relievers of their tasks. Such a large percentage of the role of a relief pitcher (in contrast to starting pitchers and everyday players) is caught up in the game context of his performance that WPA seems like one of the best ways to truly measure their value. Starting pitchers and everyday players engage in large numbers of plate appearances a season in a wide variety of game contexts, so WPA is less important in measuring their performance. but relief pitchers take on a relatively small number of PAs a season and a lot of their value depends on their level of success in high leverage situations.

  80. If all managers used their closers the same way and if closers spent their whole relieving career in that role, then save percentage would be a very important indicator, as some folks above are saying it is. But in the world as it exists, I doubt it's has an acceptable signal/noise ratio.

  81. wins do not have solid meaning for relievers. a reliever could lose a lead and then win the game. but just about every loss a reliever earns is his, and his alone.

  82. I normally wouldn't cite a single game to make the point that not all "blown saves" are equally bad, but this one has always stuck in my mind so here it is -- one of the four blown saves charged to Mariano Rivera's postseason account (out of 46 tries):

    -- 2004 ALCS, game 5: Yankees open the 8th with a 4-2 lead, but Ortiz cuts it to 1 run with a leadoff HR against Gordon; a walk and a single put runners on the corners with nobody out. And here comes Mariano! All he has to do to earn a save is (a) retire the side in the 8th without letting that runner score from 3rd, and (b) pitch a 2nd full inning ... after going 2 IP the night before. He couldn't do it; Varitek hit a sac fly to plate the tying run.

    In fact, as many of you know, all four of Mariano's postseason blown saves featured both a 1-run lead and Rivera entering the game in the 8th inning with 0 or 1 out.

  83. @82 John Autin - that Red Sox-Yanks game has always stuck out in my mind as well. It really irritates me when some people say Rivera choked by blowing two games in a row there. Anyone who knows anything knows that Tom Gordon blew that game.

  84. P.S. to #82 -- In that "blown save" charged to Mariano Rivera in game 5 of the 2004 ALCS, his WPA for the game was positive 0.307 -- higher than 12 of the 13 other pitchers who appeared in that game. (Tim Wakefield, who earned the extra-inning win with 3 scoreless innings, had a 0.468 WPA.)

  85. [...] Autin recently pointed out that Mariano Rivera once earned 0.307 Win Probability Added for his blown save appearance in game 5 [...]

  86. Adjusted ERA and save percentage, taken into consideration the skill level of the fielders behind the pitcher should only matter. W-L record is not accurate because that is a team accomplishment.