Statistics as an end to itself

Baseball Prospectus | Articles | Prospectus Today: Merit vs. Moments

Sheehan again. If you’re not a BP subscriber, I’ll summarize: he doesn’t think that Andruw should get any MVP consideration at all, apparently because defense in center field is only more important because of “highlight plays” and because of some ill-considered ranting about “merit versus moments”. Andruw hits a homer to break a tie in the midst of a pennant race? That’s a “moment” and should be thrown out, because winning games in a pennant race doesn’t matter. Andruw drives in 17 runs in 12 games as the Braves pull away from the pack? Same thing. This is why people dislike sabermetricists.

I’m on the stathead side most of the time, as my trolls will testify. Yet I’m a “soft” sabermetricist. I came to this from reading Bill James, and one thing I took away from James is that you consider all the evidence, which means while statistics are a tool you can’t ignore everything else and just give the award to the guy with the best statistics. We wouldn’t need an award then, because we all know who has the best statistics. The Braves are winning the division by 6 1/2 games, and Andruw’s won four or five games with walkoff hits. You can’t just ignore that in a pennant race because you’ve decided that the MVP is an award for the guy with the best statistics.

Baseball is not about compiling numbers. It is about winning ballgames, winning pennants, winning world championships. How can hitting a walkoff home run in extra innings against a team with whom you’re in direct competition (as Andruw did against the Natspos last week) be irrelevant to the discussion of the most valuable player? All evidence is relevant, not just what you can measure quantitatively.

29 thoughts on “Statistics as an end to itself”

  1. Hello… It’s the MVP. It doesn’t always go to the most deserving candidate.

    Sheehan can take a crash course in arts and crafts and make his own special award trophy for Albert Pujols and mail it to him. I’m sure that would mean more, anyways.

  2. Well said, especially the interpretation of Bill James.
    James approached baseball as a scientist, not as a statistician per se. He did an amazing job of posing the right questions, gathering the right data (and convincing others that the data needed gathering), and modifying his (and, eventually, others’) thinking accordingly. Statistics (as a science) is a way of thinking about information and, as you say, the statistics (in the sense of the data) rarely if ever speak for themselves.

  3. Puljos will win it. All the writers that vote on the MVP are big numbers guys that love Barry Bonds’s and Alber Puljos’s numbers. ANDRUW JONES IS WINNING TWO THIRDS OF THE TRIPLE FUCK CROWN! But it won’t matter. Let’s just win it all and stick it in guys like Peter Gammons’s face.

  4. Agreed completely, Mac. Of course, the statistic WPA measures an individual player’s contribution to winning, which, as you say, is what baseball is about. Does it not stand to reason, then, that the player who most contributes to winning should win?

    ‘Course, we don’t have full-season WPA numbers, so we might as well vote on things like “game-winning homers” because they are a rough proxy for WPA anyway (chance of Braves winning before Andruw homer: .45; chance after: 1.00; Andruw’s credit = .55 of a win, e.g.)

  5. I’m not too worried.
    If we say numbers are over rated then Andruw has come up with the big hit when we neede it.
    If we like numbers then Andruw is leading the league in hommers, RBI, and GWRBI.

  6. Bobby has heard everyone, the best lineup we have is playing tonight:


  7. Seriously, Johnny can’t hit righthanded, so why not leave McCann in there to get some at-bats against a soft-tossing lefty?

  8. Does this mean you think Andruw deserves the MVP Mac? I don’t, Derrek Lee is having a season of near Bondisan proportions. He is way, way ahead of anybody. He clearly is the best player in the league this season. If you don’t think he deserves the award because of who his SS is….well that argument isn’t really worth having, people either believe one or the other. My only problem with Andruw being an MVP canidate is that I don’t think he has been any more valuable to the Braves than Furcal or Giles. And Chipper would be well above all three if he hadn’t been hurt.

    BTW, it’s interesting the season Furcal and Andruw are having. At certain points most were just beggin to get rid of both of them(I agreed with Furcal, Andruw I wanted to keep), and they are both top 20 players in the NL.

  9. It’s just how you choose to define the word “valuable” in this context. Most of the definitions I hear (and, really, all of the ones I’ve read here) seem reasonable to me. I subscribe to an admittedly loose and flaky definition that, right now, allows me to give a slight edge to Pujols over Lee… even though I realize that the teams they play on is only a matter of circumstance. Even with the arguments that I do understand, actually adopting them probably wouldn’t allow me to comfortably give Andruw the edge. I agree with JPM and even (ugh) Sheehan that an equally compelling case could be made for Marcus Giles.

  10. Since it was already on my list of things to beat into the ground over the course of several days, I’ll say again that this would all be a lot easier if there was an award that actually meant something for the best hitter in the league. Why do we only have that for pitchers?

  11. In my definition, Lee isn’t even in the conversation.

    To me, having the best centerfielder in decades also lead the league in homeruns is such a rare commodity as to be more valuable than getting a base hit 6% more often.

  12. I posted about this in the previous thread that addressed the topic (let’s face it, if the Braves continue to pull away it will be a dominant topic for the next month), so I’ll just say that I appreciate your well-stated argument, but I disagree — to this point I have to give it to Lee, for reasons mentioned.

  13. That’s fine, though I can’t see Lee over Pujols. But my basic point remains: dismissing all other evidence in favor of choosing just by statistics doesn’t recommend itself any more than rejecting the statistics and choosing for whatever other reason you come up with.

  14. Albert Pujols is the best player in the league this year on the team that has won the most games. Usually that’s a pretty good indication that you’re the MVP. Andruw has played extremely well this year in a pennant race and that is much to his credit, but if you were a Cards fan you’d look at Andruw’s stats and notice that he still doesn’t know what a double is and lacks the high average/obp combo that both Lee and Pujols have and know who is better. I’ve enjoyed this as much as anyone else but I know that Pujols has been better this year and the past three, with only Bonds standing in his way.

  15. Look again, Lee’s year is hardly that awe inspiringly better than Andruws. Lee has a higher average, big deal. Andruw’s added value on defense more than makes up for it.

  16. Mac,
    I have copies of Bill James Baseball Abstract for 1986, 1987, and 1988. Say the word and they’re yours. It’d be my pleasure.
    All our hopes for an MVP for Andruw are worthwhile, but it’s an uphill battle. Pujols is beginning to acquire mythic status, what with that valentine of a book to LaRussa that writer put out this year.There’s nothing writer’s love more than a mythic hero with stats to back it!

  17. Mac, I seem to remember someone going through the rather arduous process of putting all of the Abstracts online. I’ll see if I can find them for you.

    As for the argument that Lee isn’t having a much better year at the plate than Andruw, are we kidding ourselves?

    He leads by .080 SLUG, his OBP is .075 higher. Lee ranks first in RC/G, Andruw ranks 30th(Marcus is 21). GPA, Andruw is 14th, Lee is surprise surprise, 1st. As for overall value, Lee has 33 Win Shares(ranking 2nd) while Andruw ranks 20th with 20. In fact, Lee has more WSAB than Andruw has WS.

    If you want to argue that Andruw is somehow more valuable than Lee, go ahead. But don’t even try and argue that Andruw is having nearly as good a season.

  18. Why did no one ever post an argument that Pujols and Lee are playing in a weaker Central Division and the two have better chance to pad up their stats facing the Pirates and Reds, while Andruw is playing the tougher competition in the Eastern Division? We can go on and discuss this MVP thing forever, but I agree with everything Mac wrote above.

    At the end, I still believe the award should be the competition between Pujols and Andruw. Of course, if one decides to take the pure stats approach, the award is either going to Lee or Pujols. It is really up to the voters themselves to decide what approach they are taking.

  19. Mac, although I fall solidly on the non-stat side of the analysis question, I have a lot of respect for what I would call the “level-headed” statistical community and, like you, was introduced to that through reading Bill James when he first started writing his tomes.

    I liked James a lot, but don’t spend an inordinate amount of time wading through what the statistical analysis community produces these days. Some of it is good (very good), but too much of it seems to be tinged with negativity (a player is not good because. . .). I know the traditionalists can be just as nasty and in my best Rodney King voice, I often plaintively wail, “Why can’t we just get along?”

    As for the MVP question, that’s been going on forever. I’m so old I remember when Nellie Fox won it the AL MVP in 1959 when Mantle, Colavito, Kuenn and others had stronger statistical seasons. But Nellie was the leader of a Chisox team that upset the Yanks (no small feat in those days, especially over a 154-game season) and that clearly counted for something in the voters mind. For years, voters seemed to opt for the “best (or most important) player on the best team” approach. But that changed somewhere along the line and I think the stat community did have a contribution to that departure.

    To me, I think individual player value has to be judged through the screen of team performance, at least to some extent. How the team ends up should matter. If it doesn’t, let’s just have a vote of all the Silver Slugger winners in each league and call the two league winners the MVPs. That doesn’t mean that A-Rod should have been disqualified when he played for those terrible Texas teams. But I think there is an additional burden for players whose teams did not perform well when the votes are cast and there probably should be.

    Tough question. I think Andruw deserves a lot of consideration and I do think Lee’s team’s performance does put an additional burden on him. If his stats were still far and away the best, he should be MVP, but I think the fact that Pujols and Andruw are playing on winning teams with stats that are comparable to Lee’s and that Lee’s advantage is not sufficient to overcome the importance of team performance.

  20. 50 pound, it is interesting that you brough up Mantle, because he is the root of this problem. The over-reliance on the value part of the mVp came about because of him, as people routinely searched for ways to NOT give it to him. He woul dhave won it 10 times if they hadn’t done that, so they did. The same thing happened to Mays and Bonds and others later, but he was the start of it.

    I just can’t justify taking points away from Lee over something he doesn’t control(the play of his teammates). If we traded Andruw straight up for Lee, would we be in a worse position? I think not, in fact I’d be willing to say that we would be in a much better position than we are now. It isn’t Lees fault that Neifi Perez is terrible, or that Dusty Baker is killing his pitching staff. And it isn’t Andruw’s fault that Marcus Giles and Rafael Furcal and John Smoltz are all excellent.

  21. Great discussion.

    I think it all boils down to how you define the MVP award. Is it the best hitter award (with maybe some consideration for defense) or is it more subjective with the emphasis on the V? If it is the statistical best hitter, then D. Lee is your man, otherwise I would argue Andruw only slightly ahead of Pujols.

    Personally, I think the advent of the Aaron award along with the Silver Slugger accolades mandated that the MVP focus more on the subjective “valuable.” Why have two awards that would always be awarded the same player, two best hitter awards?

    I really think Andruw is the MVP, based on how I define the award and how I believe it is meant to be defined currently. However, it wouldn’t take much to sway it Pujols.

  22. I’ve enjoyed this discussion more than most that have been posted here in recent months. Thanks for an articulate opener Mac and some wonderful comments from readers. My mythical vote is for Pujols (better hitter, better team, similarly injury riddled, with a bump for career merit), but Andruw shouldn’t be easily dismissed from consideration.

    Value is simply defined, but difficult to measure. It is simply the player who does the most to help his team win, including offense, defense, and leadership. Offense is pretty easily measured with some very precise tools. Defense is less easy, but zone based stats and some of the adjusted range factor stats (like Humphries DRA) can add a lot to the discussion. Leadership? Well, Terrence Moore aside, it is very difficult to get a grasp on, but it certainly exists and certainly contributes to “value.”

    I also think team performance can’t be ignored. Pendleton’s MVP in ’91 was deserved as was Kirk Gibson’s in ’88 even though others (Bonds & Strawberry respectively) had much better numbers. On the other hand, the award can’t simply be given to the guy with the most RBI or GWRBI as that is an overreliance on stats in a simplistic and ill-informed manner.

  23. … and a PS. If anyone can lay hands on the 1982 Bill James Abstract, it would complete my collection of the Ballentine published ones. And if anyone knows of copies of the self published ones available, I’d be thrilled.

Leave a Reply

Your email address will not be published.