Stat of the Day: RE24

Of course, it needs a better name than RE24. That won’t work at all. We’ll work on the name later.

I have in my mind a friend who is a big baseball fan but does not like advanced statistics at all. I think he would love RE24 (with a different name) and a few other advanced stats if given a chance. So I’m going to explain RE24 with him in mind — I suspect it will be too slow for the rest of you. Apologies in advance.

The statistic is called RE24 because, as you might know, there are 24 different possibilities when a batter comes up in any given inning. There are eight “states.” They are:

Bases empty.

Man on first.

Man on second.

Man on third.

Man on first and second.

Man on first and third.

Man on second and third.

Bases loaded.

So that’s eight states. You get 24 total possibilities because of the outs — all of these states are possible with zero outs, with one out or with two outs. That makes 24 different possibilities in any inning.

Now, each of these possibilities offers a run expectation — that is to say how many runs a team might be expected to score in the inning. You might not be able to do the math, but you intuitively understand this. If your team has the bases loaded and there’s nobody out, you have a run expectation in your mind. If your team scores just one run, you undoubtedly will be disappointed and feel like they let some get away. If your team has nobody on and two outs, any run they score feels like a bonus.

Well, smart people have calculated the run expectation for each situation. So far this year, for instance, when a team gets bases loaded with nobody out, they are expected to score 2.2 runs. When a team has nobody on and two outs, the run expectation is less than one-tenth of a run (0.0937 if you want to extend it out). So, while there’s some fairly sophisticated math happening here, it’s pretty intuitive.

All RE24 does is add up the run value a player adds (or takes away) from any given situation. Let’s say you lead off the first inning. The run expectation for every inning is about half a run (.47). That too makes sense — teams average about 4.5 runs per game (4.3 this year in the AL, 4.08 in the NL) — so that’s roughly half a run per inning.

Let’s say our batter singles. Well, the run expectation obviously goes up — man on first with nobody out obviously leads to more runs than nobody on, nobody out. The run expectation with a man on first and nobody out is about .83. That’s adding .36 runs of value. And what RE24 does is put that that .36 runs into the players bank account. Every hit, walk, HBP, sacrifice is thrown into the RE24 account.

OK, let’s try another situation. A runner comes up with men on first and second and nobody out. The run expectation is 1.4 runs. He hits into a double play. So, now, there is a runner on third with two outs. That run expectation is about .36 — which means the hitter just cost his team 1.04 expected runs. RE24 counts that too and puts it into the players bank account.

So, that’s really all you need to know. There are numerous calculations and variables, but don’t worry about those for now (unless you want to know more). The point is that RE24 — which is Tom Tango’s preferred metric, by the way — adds up a players value over the season. You might have heard of WPA — Win Probability Added — which works in a similar way. But there’s a difference: WPA adds up the WIN expectation rather than RUN expectation. That means that a leadoff double in the ninth inning of a tie game is worth A LOT more than a leadoff double in the third inning when the team is down by five runs.

Maybe that kind of measurement speaks more to you — I like RE24 better because it doesn’t have the wild swings that WPA has and isn’t as context driven (if you play on a lousy team that is often down five runs, it really doesn’t matter what you do).

Here’s the main reason why I think you will like RE24 better than other statistics.

RE24 AL Leaders:

  1. Miguel Cabrera, 75.37
  2. Chris Davis, 66.32
  3. Mike Trout, 65.97
  4. Edwin Encarnacion, 41.80
  5. Robinson Cano, 38.57

RE24 NL Leaders

  1. Paul Goldschmidt, 55.44
  2. Allen Craig, 48.16
  3. Freddie Freeman, 47.30
  4. Joey Votto, 42.97
  5. Shin-Soo Choo, 42.87

I think that will come closer to matching people’s MVP votes than just about any other stat, including WAR (especially in the American League). Yes, it does miss Pittsburgh’s Andrew McCutchen — my personal MVP because of his brilliant all-around game (he ranks 10th with a 31.88 RE24) — but I think in general RE24 is the advanced stat that comes closest to matching what many so-called “anti-stat” people really think of as “value.”

What seems to make many Cabrera fans so angry about the Trout lead in WAR (and, as expected, the lead is widening — Baseball Reference now has Trout with 8.2 WAR, Cabrera with 6.9) is that it just does not seem to give him enough credit for his offensive awesomeness. The guy’s a hitting Terminator. He’s hitting .358. He’s got 130 RBIs in 127 games. He’s a demigod. And WAR just doesn’t speak to these fans. The argument goes that WAR, with all its contextual adjustments and its various attempts to give value to things that statistics have generally not valued in the past, does not give Cabrera enough credit for simply being the Incredible Hulk of hitting.

Well RE24 does give him credit. As you can see, Cabrera has a MASSIVE lead in RE24 over everyone else in baseball. Not only that, it is the highest RE24 in baseball since Albert Pujols in 2009, and the season ain’t over yet — he could finish with the highest RE24 in a decade.*

*Just to give you one more idea of how ridiculous the big-headed Barry Bonds was, his RE24 in 2004 was 128.8 which was almost DOUBLE anybody else in baseball, and is almost 30 runs better than Mickey Mantle’s Triple Crown season in 1956.

So, maybe RE24 is a good stat for you when you think about MVP. I mean, I wouldn’t put TOO much stock in it. Last year, while the Trout-Cabrera MVP debate raged, it was actually Edwin Encarnacion who led the American League in RE24. But, anyway, RE24 is fun to talk about, though it definitely needs a better name.*

*How about Mantle Points? The Mick led the league in RE24 eight times — maybe you name it for him. Barry Bonds led the league in RE24 10 times, but as cool a name as “Baseball Bonds” might be, I doubt anyone would go for it.

64 thoughts on “Stat of the Day: RE24

  1. Tracey

    A) Miggy fans will like it because it doesn’t incorporate defense at all and incorporates speed only tangentially

    B) I like Mantle Points, but how about this: Mantle Factor? “Cabrera’s Mantle Factor is over 75 this year, tops in the majors” has a good ring.

    Reply
    1. Rob Smith

      Given Joe’s explanation, which may have left some things out, it doesn’t factor in speed at all…. except if a fast runner gets more doubles or triples, which they may, or may not. All it does is see whether the player’s hitting outcome (single, double, triple, HR, out, DP, etc) increases or decreases the expected amountof runs scored (old situation, outs/runners on, compared to situation after the at bat). It doesn’t seem, and maybe Joe left this out, to include what happens once a player is on base.

      Reply
    2. bluwood

      Sorry for interjecting here with something completely different, but I wanted to run something past the intelligent posters here:

      It’s about making saves a relevant stat again.

      When a starting pitcher is removed from a game prior to completing five innings (either for ineffectiveness or injury or whatever), and he is removed with a lead that is never relinquished, then it is the responsibility of the official scorer to determine which reliever “earned” the win.

      Perfect example (but there are tons of them): April 22, Red Sox vs. A’s. Junichi Tazawa comes into the game in the 8th with three runs in (making the score 9-6 Sox) and two runners on with only one out. He gets the Sox out of the jam with no more damage. Bailey comes in in the 9th with no one on and the the lead at three, and gets the save.

      If MLB allowed the official scorer to choose the reliever that “deserves” the save, it could go to Tazawa, who obviously was more integral in “saving” the game for the Red Sox.

      And Joe would like it because while we add that responsibility to the Official Scorers duties, we could remove errors! Win-win!

      Thoughts?

      Reply
    3. bluwood

      To add: By having the current save rule changed to the one I proposed, managers would no longer feel constrained to use closers as they currently do. You could bring in your closer when he’s needed, not when he’s going to accumulate a semi-worthless statistic.

      Thanks!

      Reply
    4. invitro

      I might say that managers who let the save rule control how they use relievers are being stupid, and their teams will pay by losing a few more games, and rewarding stupidity with failure is a good thing.

      My favorite stat for relievers might be WPA (and WAR for predictions). Fun fact: the AL pitching WPA leader is Joe Nathan, and #2 in the NL (after Kershaw) is Rex Brothers.

      Reply
    5. DJM

      @bluwood

      I tend to go for a simpler way to change the save rule.

      Any time a pitcher enters the game with the tying run on second, third, or at the plate, and does not allow the tying run to score before the end of the inning, they receive a save for that appearance.

      And by any time, I mean ANY time. If a reliever enters in the fourth inning in that situation and successfully completes the inning, they receive a save. If another reliever comes in during the same situation in the eighth inning of the same game and does the same thing, they get a save, too. If a pitcher enters the game at the start of the ninth inning with more than a one-run lead, they cannot receive a save.

      This does two things. First, it returns the rule to the original spirit of the save: giving relief pitchers who “save” leads credit. Second, it increases the number of saves, possibly many times (I haven’t done any research, so I cannot say for sure). Basic supply-and-demand says that by increasing the supply you lower the cost, which may both get fans and managers and front offices deeper into alternative forms of analysis as well as prevent teams from overpaying for saves on the market.

      So this change would both make the save a more-effective stat for analysis and a less-effective stat for increasing a player’s perceived value–though it could also give some middle relievers more leverage with their teams.

      Reply
    6. Patrick

      I think your re-imagining of the save is better than the current use. However, I don’t think it goes far enough as there is still only one save. What about a situation where 2 or 3 pitchers to amazing work to preserve a game. Which one gets credit?

      Personally, I would simply dump the save and use the Fangraphs Shutdown vs Meltdown paradigm. The 9th inning isn’t special and more than one pitcher is usually need to win the game. Lets reward all of the pitchers that contributed.

      Here is the link: http://www.fangraphs.com/blogs/shutdowns-meltdowns/

      Reply
    1. Ross Holden

      Exactly. Isn’t that why they created the Hank Aaron award? To ensure Cabrera gets an award when he’s a hitting demigod even if he’s not the all-around most valuable?

      Reply
    2. John Gale

      And what’s your point? I don’t think Joe or anyone else is saying this is the *only* stat that should be used (though that basically is what the statheads think about WAR). It’s just another one to consider. And I hate to get into the umpteenth semantic debate over this, but WAR doesn’t really measure all aspects of “value” either. It might tell us who is a better player (you know, assuming everything in WAR is accurate, which is a dubious claim at best) on paper. But if he’s on a last-place team, who cares? Last year, Trout got crushed in the voting, and that’s when he was on a good team that barely missed the playoffs. He’s now on a bad team, and Cabrera is having an even better year at the plate on a team that is cruising to a 90-win season. Personally, I don’t think a guy on a bad team has any place in an MVP debate unless he’s otherworldly *and* no one from a contender is having a great year. And I’m hardly alone. Last year’s debate showed that a strong majority of sports fans favored Cabrera over Trout, and Cabrera won the MVP in a landslide. I’d be shocked if Trout gets more than a couple of MVP votes this time. He’s been outstanding, but he frankly doesn’t deserve consideration. I’m not saying it’s his fault. Take it up with the rest of the Angels.

      Reply
  2. Brendan Kennealy

    “A runner comes up with men on first and second and nobody out. The run expectation is 1.4 runs. He hits into a double play. Now, there is a runner on third with two outs. That run expectation is about .36, which means the hitter just cost his team 1.04 expected runs.”

    How does this stat account for outcomes that are out of the batter’s control? Say, what if the 2nd baseman bobbles the relay and makes an error, or what if the baserunner who was on first breaks up the double play? Then we have Runners at 1st and 3rd with one out, but the batter still hit the same ground ball that should have resulted in a double play. Does he get penalized the 1.04 expected runs outlines above, even though he isn’t responsible for 2 outs (just 1) and also advanced a runner, or is he credited for advancing the runner and increasing his team’s chance of scoring, even though all he did was get lucky and benefit from botched defense?

    I supports better stats and believe most conversational stats are woefully unreliable indicators of offensive value, but I don’t see how this one controls for variables outside of the batter’s control.

    Reply
    1. Todd Bernstein

      I do believe this is very similar to the Batting Runs portion of a WAR calculation, which is separate from the Baserunning Runs portion. The latter would give credit to the baserunner for running 1st to 3rd on a single instead of to the batter.

      Does a stat exist that controls for a runner breaking up the double play?

      Reply
    2. BobDD

      If you discount it because it does not account for all bobbles, then you’d have to throw out all statistics – that just is not a good reason to throw it out.

      Reply
  3. Bonzi

    I would call it something like SitScore (For “Situational Scoring”) or something like that.

    It’s definitely an interesting concept, but its biggest difference from WAR is also it’s biggest drawback: It only factors in hitting, to the exclusion of defense and baserunning.

    There are problems with WAR as well: That it couples a mathematically sound context-independent hitting evaluation that most can agree on with a far less well-established defensive evaluation is among them. But I’m wary of an approach to stat-building that justifies itself by “this confirms my eye-ball test”, which is essentially what you are saying by touting its list of leaders.

    Reply
  4. RedsManRick

    So Cabrera is 10 runs ahead of Trout in terms of run value as a batter. The funny thing is that WAR has pretty much that exact same difference.

    Of course, Trout is a decent CF while Cabrera is a poor 3B. Trout is one of the best baserunners in baseball and Cabrera is poor there as well. And Cabrera has hit in to a lot more double plays.

    Could it be that those things add up to more than the 10 run lead Cabrera has as a hitter. It seems reasonable.

    If RE24 resonates with MVP voters, it’s only because it does the same thing as the other stats MVP voters like: it ignores everything else but hitting, those things at which Mike Trout is much, much better at than Miguel Cabrera, Chris Davis et. al.

    Ironically, that’s why McCutchen isn’t showing up on the NL list.

    I’m glad you’re drawing attention to RE24, but I’m not sure the MVP vote is the best context for it.

    Reply
    1. Fabian Stolzenburg

      I agree with you for the most part, but this stat clearly already accounts for Miggy’s double plays, so we don’t want to double count them.
      However, what I am really interested in is if this stat is park-adjusted? If not someone should get on that because that would make it even more impressive (although maybe the error bars get too large when trying to study run-expectation values for specific parks).

      Reply
    2. Ian R.

      To be fair, RE24 takes Cabrera’s double plays into account, and I think (THINK) it also accounts for baserunning. Caberera is hardly a poor baserunner, also – he’s stolen three bases without being caught and he generally makes the most of what speed he has.

      Of course, RE24 doesn’t account for defense and it doesn’t make a positional adjustment. Those things should matter in an MVP debate. But if you want a stat that accounts for Cabrera’s performance in high-leverage situation, RE24 will do that.

      Reply
    3. kehnn13

      Regarding defense making up for it- in Joe’s story last week, he specifically mentioned that it doesn’t. It is not until park factor was brought in (which has its own flaws) that Trout passed Cabrera.

      Reply
    4. Dodger300

      Of course, when Bourjos is healthy Trout’s value is greatly diminished.

      Because when he starts the game standing on top of a different piece of real estate, WAR changes its mind about him.

      Suddenly, Trout isn’t nearly as good a ball player as you thought he was. Silly you!

      And seeing how WAR is both modern AND definitive, if anyone attempts to disagree you are simply outing yourself as a baseball Neanderthal.

      So it is far better for you to keep quiet and be thought a fool than to ever question WAR and prove it.

      Reply
  5. MtheL

    Joe -

    I like the idea, but how does RE24 take into account runs a player actually drives in? In your situation above, what if the player hits a double scoring both runs instead of into a double play? Would he get the value of a man on second with two runs scored minus 1.4 (the runs expected prior to his double)?

    I also see two problems with RE24 (both of which are similar to counting RBI’s or many other stats for that matter): (1) This seems to favor good hitters in the middle of the lineup more than it does good hitters at the top of the lineup. A good hitter at the top of the lineup who rarely comes up with RISP would seem to less often have the chance to significantly boost his RE24 score, while good hitters in the heart of the lineup can frequently get a bigger boost. (I realize the opposite is true – bigger drops for middle lineup guys than lead off hitters, but good hitters (as your list shows) will generally stay in the black and not the red on the RE24 score, thus additions to the scores will happen more often than subtractions.) (2) Won’t a player’s RE24 score be hurt by having poor runners on base? If I have a speedy guy at second who can score on line drive, my RE24 score will be higher than having a slower guy who will only end up at 3rd. And unlike my problem (1) above, the reverse is not true. The RE24 score will go down the same if the hitter does poorly in either situation, but will not go up the same if he does the same in both situations.

    Reply
    1. kkurt23

      To your first point I think RE24 (mantle factor) would take that into account. If he hit a bases clearing double he would get (2 for the runs scored) + (.8 for having a guy of second with the correct number of outs)=2.8 – (1.4 what was expected before the at bat) = 1.4 RE24 into his account.

      Reply
    2. Todd Bernstein

      With runners on first and second (120), run expectancy is 1.4 runs. The batter hits a double and the base state is now (020), run expectancy 1.1 runs. The batter gets credit for (2 – 1.4) + 1.1 = 1.7 runs above average. The runners are splitting credit for the 0.3 additional runs actually scored.

      It doesn’t matter where you are in the lineup. If you increase the expectancy of your team to score runs, you earn RE24 runs. Heart of the lineup guys, who you would expect to come to bat with runners in scoring position, have high expectations to meet (1.4 runs in the above example). Leadoff hitters do not have as high expectations (0.47 runs above). What’s being measured is a player’s ability to exceed the average player’s production by advancing runners (including themselves) and not making outs.

      Reply
  6. W. Blake Gray

    Mike Trout appears to once again have a lock on being Grand Marshal at the Angels’ third-place parade.

    Shouldn’t that be enough of an honor? Isn’t he just happy to be here?

    Reply
    1. Rob Smith

      I think it’s amazing that Trout is hitting so well at Angels Stadium…. and not just as a slap hitter either. As Hamilton and Pujols are finding out, they can crush the ball and the thick night air just kills so many balls at the warning track. If I was Hamilton or Pujols, I wouldn’t go to the Angels in a million years!! If I was a pitcher, I’d sign up in a heartbeat!

      I think that’s something a lot of people have a hard time getting their heads around…. the park effect at Angels Stadium. The dimensions are pretty average…. maybe a couple of feet longer than average, but not obviously large dimensions. The foul territory is definitely more than average. But the night air is the thing people don’t see.

      Reply
    2. W. Blake Gray

      People often said last year like Mike Trout was entitled to the MVP and he was being penalized for the Angels finishing third.

      In fact, the MVP is an award, a reward. Only one can win it. Cabrera was rewarded with it.

      Why is this so hard for people to accept? I could care less about Cabrera. I’m not a Tigers fan. What is it about Trout that makes a certain percentage of baseball fans swoon like schoolgirls?

      Reply
    3. BobDD

      But of course this isn’t about Cabrera and Trout; they are just stand-ins for the particular belief system one has for how to measure individual baseball activity.

      There are many fans why believe Tony Gwynn a better hitter than Wade Boggs because Gwynn had a .338 – .328 batting average advantage. While others look at Boggs lead in OB% of .415 – .388 and proclaim Boggs the much better hitter. The pendulum is swinging away from batting average (Gwynn/old style stats) and strongly towards OB% (Boggs/new Sabermetrics). I’m just surprised to find out from the comments here that it isn’t unanimous yet.

      Reply
  7. Tangotiger

    A couple of notes:

    1. RE24 as shown on Fangraphs DOES include SB and CS (and BK and PB and moving runners over on outs and everything else). I think Baseball-Reference does the same.

    Basically ALL changes in base-out states are accounted for in someone’s piggy bank.

    2. Yes, they are park adjusted, as best I know.

    Reply
  8. Stephanie Rotz

    How does this work in reverse… I mean, for pitchers?
    I imagine clayton kershaws run expectancy is always lower than the stock average numbers for a given situation. Do pitchers get credit for runs not given up with this stat?

    Reply
    1. Robert

      This is a hitting stat only. You could theoretically calculate it for pitchers, except it wouldn’t be nearly as valuable. Pitchers start every new inning with nobody on and nobody out, so you’d have many situations where you’d be “rewarding” a player for getting out of a jam that he put himself in.

      You’d really be calculating a pretty convoluted stat that isn’t meaningfully different than ones that are more straight forward.

      Reply
  9. Stephanie Rotz

    How does this work in reverse… I mean, for pitchers?
    I imagine clayton kershaws run expectancy is always lower than the stock average numbers for a given situation. Do pitchers get credit for runs not given up with this stat?

    Reply
  10. invitro

    I dunno… if you’re gonna get context-sensitive and give a batter more credit for a grand slam than an empty-bases HR (which I think is a good thing when appropriate), why not go all the way and give a batter more credit for a grand slam when the team is down three runs than a grand slam with the team up by five? I think I like WPA for certain things, including MVP and HoF and other awards, and don’t care about the wild swings. I know clutch hitting is not a skill, but I think it should be rewarded when awards are given out… but I’m not sure if that’s the right mentality. Should good fortune be rewarded? What is the purpose of awards? I think it’s a hard question.

    Also, ignoring the fact that not all runs are equal is a very bad thing to do, whether trying to predict (with WAR) or to reward (with RE24). I assume there’s a RE24+ somewhere that adjusts for run context, although I don’t see it on b-r. I also assume that WPA includes run context naturally, if the win probability tables are calculated well.

    In 2013, CDavis and Goldschmidt lead Cabrera in WPA, because (I suppose) CDavis and Goldschmidt have been clutch, while Migs has been a tad anti-clutch. Trout has been -very- anti-clutch – the second-most anti-clutch hitter in baseball after Weeks – and so is only 7th in WPA. I think I would vote for CDavis for MVP right now because of the clutch factor. Seriously! I hope Joe writes a post on CDavis for MVP.

    I’d like to hear from people who are very against using clutch stats when voting for MVP (or HoF).

    (McCutchen has now passed Cabrera in WAR for 2nd… the plot thickens.)

    Reply
  11. nscadu 9

    So all this to justify or not justify Cabrera or Trout for MVP. I back Trout as the best player in the league the past 2 years. He had a double digit WAR last year and an historic season that no player had ever accomplished. Cabrera won the Triple Crown, but this year he is even better at the plate, but may not win the crown again. Triple Crown is arbitrary and has more to do with luck or slightly poorer luck by other hitters. Last year we heard all this talk about playoff Tigers and non-playoff Angels even though Trout’s team actually had more wins. This year if playoff context is considered, then there is no argument between Trout and Cabrera. Cabrera is on a team with a better record and playoff bound, but the Tigers also have some a strong lineup and solid pitching. Most valuable to a playoff team, my MVP is McCutchen. Again, a player is not measured by batting only. Without McCutchen the Bucs are definitely out of contention, he has neither the batting nor pitching supporting cast that Cabrera has and at least according to the eyeball test looks more valuable to his team than Cabrera is to his. Also, as pointed out above, McCutchen just passed Cabrera in WAR and he is playing for a team that has just about the same record as the Tigers, is potentially playoff bound in a much tighter race.

    Reply
    1. John Gale

      It’s “a historic,” not “an historic.” The “h” in “historic” is not silent. This is my all-time grammar pet peeve, and it’s much more infuriating than anything in the debate over Cabrera and Trout. Please use it correctly (I’m not actually blaming you–just all the people who keep perpetuating). Also, I’m not sure what your point about McCutchen and Cabrera is supposed to be. They play in different leagues, so it’s not relevant.

      Reply
    2. Stephanie Rotz

      “an historic” was very common up until somewhat recently and is still common in britain.

      In some regions of the US it is pronounced with a silent h. Both my parents use “an.” It is still fairly common in some areas. Ive never thought of it as correct or incorrect.

      I dont know what the elephant thing is about though.

      Reply
    3. Stephanie Rotz

      When speaking it is possible to create confusion when saying “a historic.” Listeners may hear “Ahistoric” which means “lacking historic documentation.”

      Obviously this problem is solved in writing because of the added space between the words.

      Reply
    4. Dodger300

      John Gale, the great thing Bout language is that there is NO rule book!

      If enough people choose to say something a certain way, then it automatically becomes part of the language The very fact that you hear the phrase “an historic” often enough for it to become your pet peeve is prima facie evidence that it is a perfectly fine thing to say!

      So even though my first grade teacher told me there is no such word as ain’t, she was wrong. And so are you.

      Ain’t it grand!

      Reply
  12. BobDD

    I think that by the end of the season it is quite likely that one player will stand out relatively more than anyone does now. Who stands out at this exact point makes no difference. So we are talking in theory for now about such things as MVP.

    Reply
  13. James Whitney

    While I think it’s fun to play around with numbers, RE24 strikes me as just that: playing around with numbers for the sake of doing so. As a stat it is rather opaque: you have to trust not only that the states have been recorded properly but also that the run expectation matrix has been calculated accurately. (Do you adjust for team and park effects here? I guarantee you that for every base/out state the Tigers have a higher run expectancy than the Royals).

    In the case of WAR I can appreciate it because the result is a generalized value that can’t be gotten in any other way. But RE24? Well, consider this stat for the American League:

    1. Miguel Cabrera 329
    2. Chris Davis 325
    3. Mike Trout 293
    4. Adrian Beltre 285
    5. Adam Jones 282

    That stat? The ludicrously simple Total Bases, something that could be explained in a sentence rather than the multiple paragraphs RE24 requires.

    The question I ask is: what does RE24 offer above and beyond Total Bases as a stat? Is that value worth having to wade through the explanation of the stat? I think it’s a fun number to play with but if you’re looking to define value for a casual fan I don’t think it adds all that much.

    Reply
  14. Rick Johnson

    This approach may be the oldest form of statistically determining offensive value. George Lindsay wrote an article in “Operations Research” in 1963 that assigned run values to the various offensive events: Runs=(.41)1B+(.82)2B+(1.06)3B+(1.42)HR. Steve Mann created “Run Productivity Average” around the same time, and derived his values from analyzing 12,000 plate appearances: RPA=((.51(1B)+.82(2B)+1.38(3B)+2.63(HR)+.25(BB)+.15(SB)-.25(CS). In 1978, Pete Palmer determined the values for the Total Baseball linear weights formula by running a computer simulation of all MLB games from 1901. His formula: Runs=.47(1B)+.78(2B)+1.09(3B)+1.40(HR)+.33(BB+HBP)+.30(SB)-.60(CS)-.25(AB-H)-.50(OOB), where OOB is “out on base.” These are actually 50-year old concepts upon which Bill James and others built their work, and they are the foundations for WAR, Runs Created, Win Shares, and virtually every other advanced statistic that attempts to values and individual player’s productivity.

    Reply
  15. Derek Legler

    Can you extend this to running and fielding to get a complete player value? Then compare it to WAR.
    For example, a man is on first base with 1 out. The batter hits a double. The average runner scores 70% of the time. You get credit for +.3 or -.7 depending on weather you score or not. Or maybe the math would be much more complicated, but i think it could be done.

    Reply
    1. invitro

      If you change “this” to WPA from RE24, I think the answer is yes, you just add the “Clutch” column on b-r to WAR. I think that gives you the context-sensitive stuff along with the context-neutral stuff.

      But I have no idea if that’s right. WPA – Clutch is WPA in context-neutral form, which should approximate batting wins, or whatever that number is called in the WAR formula.

      Here are number of some players this season:

      AL WAR Clutch WAR+Clutch
      CDavis 6.1 1.9 8.0
      Donaldson 5.8 1.3 7.1
      Cabrera 6.9 -0.3 6.6
      Cano 6.4 0.1 6.5
      Machado 6.2 0.3 6.5
      Pedroia 5.6 0.6 6.2
      Ellsbury 5.5 0.7 6.2
      Kipnis 5.4 0.7 6.1
      Trout 8.2 -2.2 6.0
      Victorino 5.7 -0.1 5.6
      AJones 4.0 1.3 5.3
      CSantana 3.2 2.0 5.2
      Beltre 5.6 -1.7 3.9

      NL WAR Clutch WAR+Clutch
      Goldschmidt 5.9 1.8 7.7
      MCarpenter 5.2 0.9 6.1
      McCutchen 7.0 -1.0 6.0
      YMolina 4.9 1.1 6.0
      CGomez 6.6 -0.9 5.7
      AGonzalez 3.2 2.3 5.5
      Freeman 4.1 1.4 5.5
      Simmons 5.5 -0.1 5.4
      Wright 5.6 -0.6 5.0
      Votto 5.5 -0.7 4.8
      Tulowitzki 5.0 -0.4 4.6
      Posey 4.9 -0.3 4.6
      ACraig 2.1 1.0 3.1

      Reply
  16. Flip Kromer

    Proposed name: “Blackjack Score”. The explanation you gave has a good analogy to Blackjack. You know that each combination of cards in front of you and visible dealer card gives a certain chance of winning the pot, and each additional card can make large or small changes in that chance

    Reply
  17. Ian Armstrong

    If you run things through a bit of math, you can get an re24 average (R2A) that compares well to ERA. It’s the only sane metric for evaluating a relief pitcher, whose job it is to stop a rally dead.

    Looking at RE24 every 9 (like runs per 9 but situationally agnostic), then calculating that the way R9 is used to get ERA, you get a very interesting metric.

    Reply
  18. Eric Walker

    Whether RE24 has much value depends on the extent to which one thinks “clutch hitting” is a real skill. If it is, RE24 reports accurately; if it is not, RE24 rewards blind luck a lot. If a player doubles, RE24 assigns wildly differing “bank credit” values depending on which of the 24 states prevailed at the time; but if “clutch hitting” is not a skill, there is no reason to give a credit for the event having occurred in a fortunate situation (or a debit for occurring in a different situation).

    To my understanding, “clutch hitting” is not widely regarded as a repeatable process, and is thus not a skill.

    Reply
    1. Dodger300

      Does it reward “blind luck,” or “reality?”

      There are already plenty of statistics which reflect what one THINKS should have happened, according to some formula, which Fangraphs and Baseball-Reference.com and Baseball Prospectus all agree upon (or maybe not).

      So why do you have a need to pooh-pooh a statistic that describes actual facts which we all witnessed occur? Is it somehow more accurate to ignore them or pretend they never happened?

      Reply

Comment: