FanGraphs Baseball


RSS feed for comments on this post.

  1. Can past RE24 be predictive of future RE24?

    Comment by tinmanryan — April 26, 2012 @ 1:26 pm

  2. That is a good question. An even more appropriate question is: what correlates with future RE24 more: (a) past RE24, or (b) past wRAA?

    wRAA is Linear Weights, and takes “standard” weights, while RE24 gives each event a specific weight by the 24-base-out states.

    If there is such a thing as “situational hitting” as RE24 is designed to address, then we SHOULD see a better relationship of past RE24 than with past wRAA.

    Anyway, good question.

    Comment by Tangotiger — April 26, 2012 @ 1:41 pm

  3. Here’s a link to a RE24 leaderboard I created:,65,6,8,9,10,11,16,14,15,21,22,37,38,40,61,70,58&season=2012&month=0&season1=1871&ind=0&team=0&rost=0&age=0&players=0&sort=2,d

    It appears that the stat only goes back to the 1980s. The top 5 are Bonds, Thomas, Chipper Jones, Bagwell, and Manny Ramirez.

    Brian Giles makes a surprise appearance at #29.

    Comment by swieker — April 26, 2012 @ 2:10 pm

  4. The pitchers who threw to Bonds and Sosa with runners on second and third had a special mix of stupidity and courage.

    Comment by monkey business — April 26, 2012 @ 2:13 pm

  5. I’m having a little trouble wrapping my head around some of the HR values. Wouldn’t, for example, a home run with a runner on 2nd with two outs be worth 2 runs (the home run) + the average run expectancy of 2 outs with no runners on? I’m sure it’s a simple thing I’m overlooking, or I’m confused on a concept, I’d just like to know why that is.

    Comment by Stevesaxon — April 26, 2012 @ 2:23 pm

  6. You want the CHANGE in the 24 base-out states. You are forgetting that the ending state has a positive run value on its own. You have to take that out.

    Comment by Tangotiger — April 26, 2012 @ 2:26 pm

  7. Is this in The Book?

    Comment by Tyler — April 26, 2012 @ 2:31 pm

  8. A fairly detailed discussion of the run expectancy by the 24 base-out state IS in The Book.

    However, let me recommend that you read it for free from Amazon’s Look Inside feature. Just go to my site, click on the Amazon link, and do a search for


    That’ll bring you to page 24. Page back to page 17, and start reading from there until page 29.

    Comment by Tangotiger — April 26, 2012 @ 2:36 pm

  9. The Book opens with RE24 as the basic building block of sabermetrics. Awesome stuff. The greatest joy for me, however, was explaining base-out-states to my Dad and watching the light go on in his eyes. Thanks, Tango!

    Comment by Choo — April 26, 2012 @ 3:00 pm

  10. Good Guy Tango Tiger. Writes a book, shows you how to read the discussion-relevant section for free.

    If you needed another reason to have The Book on your shelf, here it is folks.

    Comment by Tim_the_Beaver — April 26, 2012 @ 3:02 pm

  11. Thanks Tango but I will definitely be buying your work. When you say “large enough career” what would you equate that too? And are you saying that RE24 is the better method of comparing players who have had large enough careers? My main go-to has been wOBA and wRC+.

    Comment by Tyler — April 26, 2012 @ 3:07 pm

  12. If you take an exaggerated example, say all of someone’s hits and walks were with bases empty (say he was .500/.700/.800), and with runners on base, he had a .000/.000/.000 line. Overall, he’s .275/.385/.440, but he ended up with a negative RE24.

    Is he actually “pretty good”, or was he actually “below average”? It depends what interests you.

    If you think of baseball as being based on the context, and the outcomes matter with respect to that context, then RE24 is your focus.

    Now, in reality, we’re not going to have anything close to that example. But, when it comes to “close cases”, then RE24 is going to act as your tie-breaker. It adds another dimension to the discussion.

    Comment by Tangotiger — April 26, 2012 @ 3:16 pm

  13. Thanks for sharing your story!

    Comment by Tangotiger — April 26, 2012 @ 3:17 pm

  14. Perfect I was thinking along those lines as well. Thanks so much for the insight.

    Comment by Tyler — April 26, 2012 @ 3:23 pm

  15. A good example is Tony Gwynn, who, we won’t be surprised, was a good situational hitter. His wRAA is +392 runs, which is great. But his RE24 is +551, which is tremendous.

    So, RE24 gives him credit for all the adjustments he made, and it also includes by the way any “moving runners over by out” he may have had.

    Comment by Tangotiger — April 26, 2012 @ 3:31 pm

  16. Curious how Gwynn compares to Raines and, more broadly, if RE24 measures tend to rate high BA hitters higher than batters who walk more (as compared to, say, wOBA).

    Also – I should know this — do RE24 values reflect steals/caught steals too?

    Comment by craigtyle — April 26, 2012 @ 3:48 pm

  17. Feel free to search for any player you want. Raines is +423 in wRAA and +503 in RE24. So, a situational hitter for sure.

    And since Vince Coleman is around average in all these stats, I can only infer that SB/CS is part of all the stats (wOBA, wRC+, wRAA, RE24).

    Comment by Tangotiger — April 26, 2012 @ 4:08 pm

  18. How does A-Rod’s RE24 compare to his WRAA+?

    Comment by Chadam189 — April 26, 2012 @ 4:41 pm

  19. AROD

    +691 wRAA
    +655 RE24

    And I know SOMEONE is going to ask for Jeter:

    +382 wRAA
    +394 RE24

    Anyway, you can look up anyone yourself. Just do a search like I did…

    Comment by Tangotiger — April 26, 2012 @ 4:54 pm

  20. Take the case of a runner on first and 2 outs. Since it says the “value” of an out is -0.26, the generic run expectancy in this situation is 0.26 runs. By hitting a home run, you score 2 runs and enter a new state with 2 outs and nobody on. This new state is worth 0.12 runs. So you scored +2 runs, but then moved from a generic +0.26 state to a generic +0.12 state. So the result was +2 – 0.26 + 0.12 = 1.87 net runs better than the average outcome.

    Comment by Jon — April 26, 2012 @ 5:22 pm

  21. Shorter version: you have to subtract the generic run expectancy of the state before the home run to find the net effect.

    Comment by Jon — April 26, 2012 @ 5:23 pm

  22. Jim Rice:

    +392 wRAA
    +247 RE24

    Comment by DavidJ — April 26, 2012 @ 5:44 pm

  23. is the RE24 stat listed on a player page or the leaderboard links equivalent to “runs created” or “runs above average” or “runs above replacement”?

    e.g. in 2007, A-Rod had 76.95 RE24. Does that mean, according to his complete performance in all the base-out states he had in 2007, he generated 76.95 runs? Or did he generated 76.95 runs above AVERAGE? or what?

    just want to make sure I understand the context, as this seems like a great stat for judging results with context included.

    Comment by batpig — April 26, 2012 @ 6:06 pm

  24. !!!

    Those GIDP really killed him there, didn’t they!

    Great find!

    Comment by tangotiger — April 26, 2012 @ 6:31 pm

  25. Runs above average, thanks for asking.

    Comment by tangotiger — April 26, 2012 @ 6:32 pm

  26. thanks for the reply. Is “average” defined as the average expectation for the base-out state based on historical data? e.g. runner on 1st, 1 out would historically be expected to produce 0.34 runs on average (just making up a number) but if the batter hits a homerun (worth 1.73 runs according to the link you posted) then he would get credit for 1.39 runs above average?

    Comment by batpig — April 26, 2012 @ 6:55 pm

  27. Dave Parker, interestingly, is almost the exact reverse of Rice:

    +260 wRAA
    +384 RE24

    Comment by DavidJ — April 26, 2012 @ 7:17 pm

  28. Great find!

    Comment by Tangotiger — April 26, 2012 @ 7:29 pm

  29. Right, it’s based on historical data then “tuned” for a particular run environment for that year-park.

    You can see a standard set here. Someone at my site suggested taking a screen capture with your phone (of the 1993-2010 chart), and reference it at the park!

    Comment by Tangotiger — April 26, 2012 @ 7:31 pm

  30. How is it that one of the linear weights defensive indifference values is negative?

    Comment by monkey business — April 26, 2012 @ 9:24 pm

  31. If you are referring to this chart, then I explained the sampling issues:

    Comment by Tangotiger — April 26, 2012 @ 9:49 pm

  32. Ryan Howard’s career RE24 is 50 points higher than his wRAA, and was 20 points higher this year.

    Comment by YanksFanInBeantown — April 26, 2012 @ 10:59 pm

  33. It’s runs not “points”. Great find!

    Comment by Tangotiger — April 26, 2012 @ 11:10 pm

  34. B-ref’s goes further back: RE24 is computed from play-by-play data which is only complete from 1974 to the present. From 1948-1973, the data is incomplete, though for most seasons only less than 20 games per season total are missing.

    Comment by Jonathan C. Mitchell — April 27, 2012 @ 9:03 am

  35. According to B-Ref’s Barry Bonds is tops with 1356.277 (fangraphs is 1348.26) and a distant second is Hank Aaron at 976.728. Yeah, Bonds was the good.

    Comment by Jonathan C. Mitchell — April 27, 2012 @ 9:04 am

  36. Right, sorry.

    I was trying to figure out of the value of defensive indifference really should be positive after there was such a SB in a game I was listening to that increased expected output for the offensive team. I was hoping that table could tell me, but without error bars, I’m not sure if the non-zero value is significant.

    Comment by monkey business — April 27, 2012 @ 9:33 am

  37. Ah, okay, I was adding the RE of the new base-out state instead of subtracting it. Thank you!

    Comment by Stevesaxon — April 27, 2012 @ 10:55 am

  38. Just check Fangraphs’ play-by-play log for the game/play in question, and you’ll see how much the win expectancy improved because of the defensive indifference.

    Comment by Tangotiger — April 27, 2012 @ 11:09 am

  39. I grabbed the 778 batters with 5000+ PA from 1974 to 2012; here are the leaders and trailers in RE24/600PA – wRAA/600PA:

    Jose Cruz +12.4
    Terry Pendleton +11.2
    Ryan Klesko +11.0
    Edgardo Alfonzo +10.4
    Steve Garvey +10.1
    Carl Everett +9.7
    Darryl Strawberry +9.6
    Mike Piazza +9.4
    Tony Gwynn +9.3
    Derek Bell +9.2

    Brook Jacoby -7.0
    Chet Lemon -7.1
    Tom Brunansky -7.3
    Ivan Rodriguez -7.4
    Lance Parrish -7.4
    Rick Dempsey -7.6
    Neifi Perez -7.6
    Jason Varitek -7.9
    Jim Rice -9.6
    Rick Burleson -9.7

    Comment by Eric R — April 27, 2012 @ 11:13 am

  40. Wonderful work!

    Comment by Tangotiger — April 27, 2012 @ 11:16 am

  41. Whoops– not sure where I got 778 batters… it was actually 380 :)

    Comment by Eric R — April 27, 2012 @ 11:45 am

  42. Awesome stuff. If nothing else, this is a great stat to show old-school fans who believe in RBI and whatnot. It’s a nice compromise between advanced stats and traditional wisdom about situational hitting.

    Comment by Matt — April 27, 2012 @ 1:42 pm

  43. Excellent characterization. I agree it’s got that nice blend that it bridges the old school and new school.

    What prompted this thread was that I asked readers on my blog with metric they had little to no use for (for whatever reason, either because they didn’t like it, or they didn’t know what it was), and RE24 was #1 for least useful.

    Basically, RE24 got bad or no P.R., and hopefully it’ll gain at least enough traction that it becomes another tool in the saber-utility-belt. Productive outs, situational hitting, GIDP, etc. It captures all of it pretty well, and in terms of “real” runs.

    Comment by Tangotiger — April 27, 2012 @ 1:48 pm

  44. Am I correct in understanding that RE24 doesn’t consider fielding at all? For example, if the bases are empty there is no difference in RE24 between reaching first by a single or an error, right?

    Comment by Jim Lin — April 27, 2012 @ 5:55 pm

  45. Correct.

    Comment by Tangotiger — April 27, 2012 @ 9:51 pm

Leave a comment

Line and paragraph breaks automatic, e-mail address never displayed, HTML allowed: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Close this window.

0.250 Powered by WordPress