What’s the matter with relievers?

Take it away, CHONE:

PlayerInningsERARAR
Mike Gonzalez493.675
Brandon Lyon643.666
Rafael Soriano523.296
Billy Wagner373.415

So the Red Sox trade for Billy Wagner, with Wagner agreeing on the condition that they not pick up his option. However, he allows them to offer him arbitration, apparently not understanding how type-A status will put him in the Juan Cruz class. Well, the Braves had other plans. Before they even heard back from the two type-A relievers to whom they’d offered arbitration, Braves were so eager to give up their first round pick that they signed Wagner to a $7-million one-year deal, with a $6.25 vesting option. Rafael Soriano then decided to accept arbitration, and signed for $7.25 to pitch in 2010. So the Braves could have just held onto Soriano instead of signing Wagner. Wagner gives them similar production on a similar contract, but comes at the cost of draft picks. And then Brandon Lyon and Mike Gonzalez (the Braves other type A) sign multi-year contracts with uncompetitive teams that had just finally gotten out from under the costly contracts of Jose Valverde and Danys Baez. Teams are giving away draft picks and $5-7 million/year contracts for the sake of approximately five runs a year. I’m lost.

Can someone please tell me what’s going on?

Print Friendly
« Previous: Orioles get great value in Mike Gonzalez
Next: Christmas Holliday? »

Comments

  1. Steve Stein said...

    Well, the obvious answer is that the teams don’t know what they’re doing.

    Other, subtler answers might be:
    - Although Soriano and Wagner were 5-6 RAR in 2009, the teams expect better performance in 2010.  After all, you do want better than a mid-3 ERA out of your closer if you expect to be competitive.

    - Maybe the runs a closer saves or gives up are higher-leverage runs in terms of wins?  5 runs better than replacement value might be 5 wins.  Or something like that.

  2. ecp said...

    I find it ironic that this was posted an hour after another post on THT Live called “Orioles get great value in Mike Gonzalez” and praising the contract as a steal.

  3. Peter Jensen said...

    Jeremy – One of the reasons LI was invented was to provide a measure for the increased importance of relievers pitching time.  You definitely don’t want to use WPA/LI as a measure for relievers since dividing by LI completely negates that importance.

  4. Nick Steiner said...

    I do think that relievers are worth more than we give them credit for.  I find it hard to believe that all of these organizations, including the Sox, Rays, etc. are paying so much for an overrated position.

    An interesting thing to note is that a good reliever’s WPA will generally be a lot higher than his WAR.  In 2009, there were 23 relievers with better than a 2.0 WPA.

    http://www.fangraphs.com/leaders.aspx?pos=all&stats=rel&lg=all&qual=y&type=3&season=2009&month=0 

    However, there were only 10 relievers with at least a 2.0 WAR.

    http://www.fangraphs.com/leaders.aspx?pos=all&stats=rel&lg=all&qual=y&type=6&season=2009&month=0

  5. The Real Neal said...

    There’s a couple logical flaws in this argument.

    The first, as point out, is that high leverage reliever innings are of average importance.  They’re not.

    The second is the way that forecasting is done by individual teams then extrapolated to free agent contracts.  The team that signs a free agent will usually have met one of or both of the following conditions:

    1. They have the most optimistic projection for that free agent
    2. They value that free agent’s contributions more highly than other systems (i.e. teams)

    Which team spent the most on free agents last off-season again?

  6. Jeremy Greenhouse said...

    Steve, I posted projections. CHONE predicts 5-6 RAR next year for them. Of course, teams know a lot more than I do, and if the Braves like Wagner’s medicals, sure, that’s a pessimistic projection. But how about Lyon? The Astros know what they’re getting from him, and it’s not worth $15 million.

    As for your second point, what Dan said.

    Greg, fair point. I’m not alone in my thinking, though. If you have a case to make for reliever’s value, I’m all ears.

    Nick, please don’t use WPA to assess the value of relievers. And don’t compare a metric using league average as a baseline to one using replacement level. I’m a huge proponent of WPA/LI if you want to compare relievers to other types of players, but I doubt you’ll find anything there.

    Neal, the team that signs a free agent is also vulnerable to the winner’s curse. I’m sure free agents net positive revenue, but not relative to how else teams could be spending that money.

  7. Nick Steiner said...

    Jeremy – I’m not using WPA to value relievers, simply as a baseline for the expected range of values.  WPA is an approximation of the change in win probability, so it seems to be a pretty way to judge the impact a reliever can have. 

    Your point about average vs. replacement level is valid, but it goes against your case!  If you would compare those relievers projected ERA’s to average instead of replacement level, their RAA would be almost nothing.  So comparing WPA to replacement level would increase the expected value of relievers.

  8. Dave Studeman said...

    I think Jeremy is objecting to the use of leverage in a metric that is used to establish the contractual value of a reliever. Leverage is not something that is attributable to the reliever’s skill set—it is management’s decision when to use or not use a reliever.

    Having said that, I believe the sabermetric community often goes too far in neglecting leverage as a part of a pitcher’s skill. The proven ability to pitch well in high-leverage situations is something that the market values, as it should, IMO.  A fundamental question is how *much* the market should value it, and whether that skill is actually “proven” in any individual pitcher’s case.

  9. Greg said...

    WXRL seems to capture relievers value pretty well by taking leverage into account.  In 2009, Papelbon and Rivera both had over 6 wins above replacement according to WXRL.  22 relievers had over three wins.  This seems to justify the eight figure salaries relievers sometimes earn.

    P.S.: Let’s omit the universally panned Brandon Lyon signing from this discussion.

  10. Dave Studeman said...

    WXRL and WPA are pretty similar stats. I like the comparison to replacement in WXRL better, but that’s tricky to do. I wonder what the math is?  Was it published anywhere?

    Still, the underlying issue remains.  Pitchers shouldn’t get full credit for the leverage implicit in WPA or WXRL.  Leverage exposure is a management decision.  The key question is whether they should get any and, if so, how much?

    Personally, I’m not sure the Lyon deal was so bad, though I understand why some people feel differently about it.

  11. Greg said...

    Why not give them credit for all of the leverage?  The high leverage innings exist and if you let a replacement level pitcher pitch them you will lose a lot of games.  It seems reasonable to expect the Red Sox to have lost 6 more games had they non-tendered Papelbon and called up Fernando Cabrera from Pawtucket and let him try to close out their wins.  I’ll get back to you on the WXRL math.

  12. Jeremy Greenhouse said...

    ecp, yeah, that is ironic. I hadn’t seen that until after I wrote my post. Needless to say, I strongly disagree with that post.

    I’m going to start parading for a deserved leverage index for all pitchers. WPA and WXRL do not capture that deserved LI.

    I don’t really know if pitchers who have proven that they can Close should be paid more. And I don’t think that I’ve ever seen a reliever prove he can’t pitch as well in high leverage situations as low leverage situations. There is the Javy Vazquez effect of pitchers not doing so well with men on base, which is why we should use WPA/LI. But I think the Fausto Carmona of pitchers not being able to handle the ninth inning actually demonstrates why you should not pay too much for proven Closers, since they’re only relievers.

  13. Greg said...

    XXRL was first explained in the 2005 Prospectus Annual but here is how Keith Woolner explained WXRL back in 2005 at http://www.baseballprospectus.com/article.php?articleid=4378:

    “WXRL is the change in expected runs due to the reliever’s performance. As such, it matters very much the kinds of situations he comes to pitch in—and whether his good outings were in high or low leverage situations.

    A simple example. A pitcher appears in two games. He gets lit up for 3 runs in one inning of work in game A, strikes out the side in game B.

    Now consider two different game situations:

    Situation 1: 1 run lead, bottom of the ninth. Assume win probability
    upon entering is 80%.
    Situation 2: Down by 12 runs, bottom of the 7th. Assume win
    probability upon entering is 5%.

    Combining the games and situations, there are two possibilities: A1,B2 or A2,B1.

    A1: Blows a tight game, and his team loses. The reliever’s win expectation for this appearance is 0% – 80% = -0.800

    B2: Pitches a perfect inning, but the game is pretty much hopeless. Maybe the team has a 7% chance of coming back now. Win expectation for this game is 7% – 5% = +0.020

    The pitcher’s combined WX for these two games is -0.800 + 0.020 = -0.780. Let’s look at the other case:

    A2: Pitcher shuts down the other team to preserve a clutch win. WX = 100% – 80% = +0.200

    B1: Pitcher puts them further into an already deep hole. Suppose Win probability drops to 4%. WX = 4% – 5% = -0.010

    Pitcher’s combined WX for the second case is +0.200 – 0.010 = +0.190.”

  14. Dave Studeman said...

    Greg, thanks.  I understand WX—it’s WPA.  My question regarding the math is how he compares WX to replacement.  I’ve devised my own process for doing that, which you can see here:

    http://www.hardballtimes.com/main/statpages/wpab_form/

    I don’t know how Keith did it.

    You don’t want to give a reliever full credit for the “replacement” level of LI because of the chaining issue.  This is critical.  A closer isn’t replaced by a replacement-level pitcher.  He’s replaced by the setup man.

    This is the sort of thing that WAR includes.

  15. Dave Studeman said...

    But I think the Fausto Carmona of pitchers not being able to handle the ninth inning actually demonstrates why you should not pay too much for proven Closers, since they’re only relievers.

    This logic escapes me, Jeremy.  If you’re in the market for a closer, would you sign Carmona to be your closer?  And if not, wouldn’t that shrink the pool of available closers, which drives up price?

  16. Jeremy Greenhouse said...

    The Fausto Carmona example is hyperbole. He had one of the most disastrous runs as a Closer in the history of baseball, and then turned into a very successful starter. Though he can’t Close, he still has plenty of value.

    I’m not smart enough to argue with David Gassko. My point is that relievers and starters come from the same pool of players and should be evaluated the same way. Give every one of them a deserved leverage index. Anyway, it seems to me that all of these relievers deserve LIs between 1.1 and 1.3 or so.

  17. Dave Studeman said...

    I haven’t thought this all the way through, but my feeling is that starting and relief pitchers aren’t quite a fungible pool of talent.  First of all, there is preference; secondly, there is effectiveness.  Not every starting pitcher wants to be, or is equally effective as, a reliever, and vice versa.

    IOW, there are two semi-separate markets of pitchers.

    Plus, you can’t hide your bullpen, unlike other markets with a given number of players.  You can live with a weak-hitting shortstop, for instance, because you might have lots of other good hitters in your lineup.  You can’t do that with relievers. To be successful at all, you have to have good relievers in this day and age.

    You can use your bullpen less often, of course, but only if you have starters who can (literally) shoulder that burden.

    So, the pool of pitchers isn’t completely fungible, and you must have (at least) a couple of very good relievers to be successful.  Those factors drive up price compared to other positions.

  18. Greg said...

    Dave, Thanks for the link to Gassko’s article on chaining.  It’s a great explanation, but we still have the problem of relievers earning significantly more money than their WAR values would predict.  So either the teams are all overpaying for them or we’re undervaluing their performance.

  19. Dan Novick said...

    Steve,

    A high average leverage index for a closer will be closer to 2 (Mariano Rivera was at 1.94 and 1.84 the last two seasons). So if you want to multiply RAR by anything, don’t multiply it by more than 2. After accounting for chaining, it should be even less than that.

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 *