In Retrospect…

Some things work a lot better in your mind than they do in the real world, and after a couple of days, I think we’ve learned that the idea of taking “official positions” was one of these things. Our intention for the series was to try and use the awards voting to help clarify some misconceptions about our metrics and how they should be used to evaluate player seasons, but in practice, the titles of the post – and the fact that we’re really a big conglomerate of individuals with their own take on things – probably created more confusion than anything else.

Life is full of lessons, and hopefully we’ve learned from this one. We’ll be publishing some thoughts on how our metrics pertain to awards voting, but you won’t be seeing any more “official positions”, and we’ll try to do a better job focusing on the validity of multiple viewpoints. In almost every case, there are legitimate disagreements that can be had, and we value the different voices that make valid points from different angles. We’ll try to make that more clear going forward.

Just like our metrics, we’re not perfect. When we do something that doesn’t work, we’ll try to correct it. The “official position” series didn’t work as intended, and so we’ll make some steps to correct that.




Print This Post



Dave is a co-founder of USSMariner.com and contributes to the Wall Street Journal.


121 Responses to “In Retrospect…”

You can follow any responses to this entry through the RSS 2.0 feed.
  1. Ryan says:

    Oh man, I totally called that.

    Vote -1 Vote +1

  2. teenarcher says:

    So is this your official position on official positions?

    +76 Vote -1 Vote +1

  3. Telo says:

    Still baffling to me that a site with so many eyes on it doesn’t have an editor.

    Vote -1 Vote +1

    • Dave Cameron says:

      The site has an editor. That guy happens to be fighting cancer, and is doing the best he can while juggling about 47 things in his life.

      +36 Vote -1 Vote +1

      • Telo says:

        Hope you’re doing better, Dave.

        But if you are actually the “editor”, well, you may want to crack a grammar book, get a new prescription on your glasses, and take a college course on style, because some of the writing that gets published on this site is just plain poor. Which, to be honest, is fine most of the time. It doesn’t obscure the analysis (usually). But it could be so much better.

        I always wondered why Cistulli was relegated to NotGraphs obscurity when the one thing he is actually good for, grammar/style/general writing ability (you’ll notice comedy is not listed), is exactly what FanGraphs proper lacks.

        -43 Vote -1 Vote +1

      • Brad Johnson says:

        I’m surprised by that. I wrote for a five writer blog that had “an editor.” From experience elsewhere, I would expect Fangraphs to have 3-5 editors.

        Vote -1 Vote +1

      • batpig says:

        In the most respectful way possible, I agree with Telo completely.

        This site needs a dedicated editor who can quickly proof and edit grammar/spelling/etc when articles get posted. The volume of posts with glaring errors is alarming and unprofessional for an otherwise classy organization.

        This is not at all to bash Dave, but even before his illness the above was all true. I think Dave has far too much on his plate (even in healthy times) to engage in the daily tasks that a basic copy editor performs.

        Vote -1 Vote +1

      • Jason461 says:

        I have to jump on the bandwagon here. I think the analysis at Fangraphs is great. The writing… eh, it’s okay. A copy editor would help a lot.

        Vote -1 Vote +1

      • Swingdoc says:

        IMO more editing will add absolutely NOTHING to the quality of the work here or my enjoyment of the site. Maybe there are a few “in the box types” who it really bothers but my guess is they are in the minority. I see the comments often and it annoys me just as those errors bother them.

        No Dave, please spend your time on those other 46 more important things and don’t sweat the small stuff. I say we take a vote and be done with the issue for good!

        Vote -1 Vote +1

      • Bill says:

        Dave, I love your work and I really like this site, but for once, I agree with Telo. Delegate some of your copy editor responsibilities. This site is too good to be held back by the lack of a dedicated copy editor.

        Vote -1 Vote +1

      • CircleChange11 says:

        For me, the grammer and spelling don’t matter much. I read right past them. I recall a demonstration at a conference where a paragraph was constructed lacking any punctuation and full of spelling mistakes. everyone could read and comprehend the passage.

        The most important thing to me is that the stats and analysis is accurate.

        In that regard I think the most important thing an editor could do is just to repeat the question “and why is that?” over and over until the author has dug as deeply as he can.

        We ask this site to provide numerous articles, on a daily basis, that are interesting, revealing, analytical … and for free. IMHO, our expectations are far greater than what the site should reasonably be expected to deliver.

        The topic has been brought up in other discussions, but peer review might be the way to go, where each article is sent to multiple peers to proof, question, challenge, support, etc. I don’t know how this stuff works logistically, but I can imagine it’s not easy.

        I guess I don’t view this site as a professional journal, but more along the lines of an online magazine/blog.

        +7 Vote -1 Vote +1

      • Tangotiger says:

        Circle: well-said.

        I was trying to find a polite way to tell people to be realistic with a section of the site that is the cherry on top of the Fangraph Stats 7-layer cake, all of which you get free at the company picnic.

        I couldn’t do it, and you did it fantastically well. So, ditto.

        Vote -1 Vote +1

      • Jason461 says:

        Circle and Tango – I get that. I do. That said, I’ve noticed (as I’m sure you have) that there are occasionally serious comprehension/misinterpretation problems in the comments. Some of that is user error (readers not reading closely), but some of it is stylistic. I teach writing for a living, so this is kind of my wheelhouse, but I don’t think it’s unfair to say that the smoother the writing is, the fewer issues you’re going to have with misunderstanding. Also, as a side note, of course you can understand a horribly constructed paragraph, but read thirty of them in a row and tell me how you feel.

        There’s also something to be said for enjoying the writing. I can nearly always understand the articles here (and when I can’t, it’s more my failing than the writer’s), but I often cringe at the style. It’s a distraction that, frankly, occasionally keeps me from reading what’s posted here.

        Now, to use an extreme example, set that against something like Joe Posnanski’s blog, which I’m sure we all read, he fudges the analysis from time to time when he writes about stats (something he admits to freely), but virtually everyone gets what he’s trying to say. I realize this is something of an apples and oranges comparison as he is more of a storyteller and this is an almost purely analytical site, but good writing is good writing.

        A site like Fangraphs, which presents advanced concepts to a wide audience needs to pay special attention to clarity. Some of this stuff is hard for the general population to understand and muddled writing doesn’t make it any easier. Clarity is clarity and the writers at Fangraphs aren’t always the clearest presenters of their ideas.

        Last, the way the site occasionally chafes at criticism is off-putting. Fangraphs is huge now, whether it wants to be or not. That means people are going to criticize its weak spots. When your front page content is written and you then try to claim that the writing isn’t important, well, you’re kind of shooting yourself in the foot.

        Vote -1 Vote +1

      • Tangotiger says:

        Then the best option is for the commenters to submit their issues directly to Dave and David, rather than polluting the thread with their personal concerns of style.

        I was at a parent-teacher get-together, and I was flabbergasted that a parent criticized a teacher’s method of teaching in view of everyone.

        Concerns, valid or not, need to be addressed in the proper venue. While we’re in the middle of a discussion of WAR, talking about the poor grammar is itself more annoying than the grammar being questioned.

        Vote -1 Vote +1

      • Jason461 says:

        But isn’t this thread exactly the place for that then? Dave’s post is entirely about how they screwed up and did a poor job conveying their intentions. Wouldn’t comments about how this is a reasonably common occurrence be exactly to the point?

        For the record, I’m talking very little about style here. I’m talking about writing that is clear and easily understood.

        Vote -1 Vote +1

  4. NF says:

    You’re sorry you posted it becuase everyone laughed at you, not that there was “confusion.”

    -40 Vote -1 Vote +1

    • Telo says:

      Well, let’s not go for the throat when they are trying to do the right thing. People weren’t laughing as much as pointing out that Bradley was making some strong claims without enough logical support. UZR is just fundamentally flaky. It’s a black box based on inaccurate data. There are so many fundamental issues with it that Bradley ignored. He talked well about the concepts of the data, but applied them in too ambitious a way.

      Vote -1 Vote +1

      • mcbrown says:

        Agree on not going for the throat, disagree that Bradley made any mistakes. I think the “anti-” argument boils down to “UZR is noisy” without any quantification (and there are lots of measurements in life that are noisy but still “good enough”).

        Vote -1 Vote +1

      • Anon21 says:

        He absolutely did make a significant mistake, as was pointed out in comments. People generally weren’t taking issue with the “single season UZR is enough” point, but they were pointing out that Bradley’s comparison between UZR and BABIP misses the point that UZR has significant measurement error and BABIP has virtually none.

        Vote -1 Vote +1

      • Telo says:

        The thing is, nothing Bradley said was really wrong (as I recall). So you are right, in a way. He just started the article from a really awful place: assuming UZR did a great job of measuring what actually happened on the field. It just doesn’t. It has several fundamental flaws, namely the actual source data itself. You just can’t write an article like that without addressing the elephant in the room. Pretending that UZR does an awesome job of telling us what happened doesn’t help anyone.

        Vote -1 Vote +1

      • mcbrown says:

        Anon21, see my comment re: measurement error below. What is “significant”? BABIP certainly does have measurement error – hits and errors are subjectively judged by the scorer. We all assume (correctly) that the error is low, so we ignore it.

        Vote -1 Vote +1

      • mcbrown says:

        Telo, I disagree. I think sophisticated readers (the ones objecting to Bradley’s article) already know the issues with UZR. There is no more point in bringing them up than there is in bringing up scorer subjectivity as a component of ERA in arguing whether ERA or FIP is a better measure of “what happened” over a season – we all know that many people would prefer RA over ERA for that reason.

        Vote -1 Vote +1

      • Anon21 says:

        So yes, the readers who objected to Bradley’s article clearly understood the measurement error point. I think most of the people who read Bradley’s article did not understand that point, as Bradley’s analysis would have seemed familiar from the context of offensive statistics. Trouble is, it just doesn’t work that way in the UZR context–again, the type of error at issue is just different.

        Vote -1 Vote +1

      • Telo says:

        I think you overestimate the number of readers who truly understand UZR. Plus, in an article titled “Official Stance on UZR” or whatever, to lead with “1 season of UZR is enough” and NOT to touch on ANY of the inherent measurement and data challenges with UZR is extremely misleading and a very poor decision.

        However, I understand what you’re saying – you understand UZR and the point Bradley was trying to make by likening UZR to BABIP – why we shouldn’t normalize it. But just because we understood his intention (I happen to disagree with him) doesn’t mean it was a well executed article. Cameron did the right thing.

        Vote -1 Vote +1

      • Barkey Walker says:

        UZR also fails at the situation where multiple player could field a ball. Is Longoria’s UZR inflated by him getting more pop flys in four territory than other third basemen?

        The denominator needs to be fraction of balls that are played in that location. What is the value of a shortstop that also covers third? Not much since there is already someone there, but UZR would say it was a good fraction of a run per stop in front of the flummoxed third baseman.

        Sometimes it is pointed out that it doesn’t matter much… okay then just make it right.

        Vote -1 Vote +1

    • JD says:

      “Hey what are you guys doing today?”

      “Oh, probably hangin around Fangraphs and trolling all day.”

      “Ok, cool, see you there.”

      +66 Vote -1 Vote +1

    • CircleChange11 says:

      Does it matter why it was posted?

      Don’t be that guy that doesn’t accept an admittance of a mistake because it’s not self-humiliating enough.

      Just like our metrics, we’re not perfect. When we do something that doesn’t work, we’ll try to correct it. The “official position” series didn’t work as intended, and so we’ll make some steps to correct that.

      Seems reasonable to me.

      +20 Vote -1 Vote +1

    • Hurtlocker says:

      Assholes do vex me

      Vote -1 Vote +1

  5. Mark Geoffriau says:

    Is there any reason the UZR article can’t be reposted with the above qualifier appended to it? There was some interesting discussion going on.

    +10 Vote -1 Vote +1

    • Telo says:

      I agree. I had just posted a comment that I believe would’ve been a great discussion topic.

      Vote -1 Vote +1

    • Anon21 says:

      I think the unfortunate thing is that people will continue to make the exact same mistaken point Woodrum made. It pops up constantly in discussions about UZR; mgl has punctured it at The Book, but Fangraphs obviously has a much wider readership. An article here simply explaining how UZR has measurement error, and how measurement error differs from the more familiar results versus true talent level point, would go a long way towards educating people about what UZR can and cannot tell us.

      And if Fangraphs still thinks that single-season UZR is worth using (which again, I think is a new position or at least a big change in emphasis over the past few months), then that’s fine. Just be clear about the facts of the situation, and why small samples of UZR have certain problems that small samples of hitting results do not.

      Vote -1 Vote +1

      • mcbrown says:

        Argh, the measurement error again… will someone complaining about the measurement error please make an attempt to quantify it? Lots of measurements in life have error; the presence of error doesn’t render them useless unless the error can be quantified as large.

        MGL’s own argument against UZR as a good measure of “what happened” isn’t very strong, in my opinion.

        Vote -1 Vote +1

      • JDanger says:

        The Margin of Error in UZR needs to be hammered home a lot more, absolutely. This issue keeps popping up, especially as we near awards time. Well said.

        Vote -1 Vote +1

      • Anon21 says:

        mcbrown: That’s a ridiculous request. UZR is a proprietary formula; no reader has access to its guts, or can make any attempt whatsoever at quantifying the measurement error. Even mgl, who designed the tool, says that he cannot quantify the measurement error.

        And no, the presence of error does not render a metric useless, but understanding the difference between measurement error on the one hand and the difference between true talent and observed results on the other (which I don’t think of as “error,” per se) is key to making accurate arguments about a metric’s value. A lot of people, I think Bradley included, do not seem to understand or be aware of the measurement error point at all with regard to UZR.

        Vote -1 Vote +1

      • mcbrown says:

        Anon21, I don’t think it’s ridiculous. Many people are raving at Bradley over an error they believe is huge but can’t back up – that seems ridiculous to me.

        I can’t say whether Bradley understands measurement error. I for one do understand measurement error, and I understand how sample size diminishes error. And that leads me to conclude that if 3 seasons’ worth of data is “enough” then 1 season can’t be “meaningless”. If one season is so worthless a record of what happened as to be “meaningless”, then it is highly unlikely that 3 seasons is “enough” to measure true talent. Assuming the errors are symmetric and uncorrelated then 3 seasons of data is likely only about 1.7x as good as 1 season of data (though of course it is tough to say without lifting the kilt of UZR). So if UZR has say a 10% margin of error at 3 seasons, the error should still be <20% at 1 season – i.e. not the greatest thing in the world, but not useless.

        I will continue to view it this way until someone explains differently.

        Vote -1 Vote +1

      • Jason says:

        McBrown,

        They absolutely could calculate the error in UZR, they just don’t. We can’t do it because we don’t have the data. You choose to just trust them that the sample size is large enough? Why? This is not how science is done. I will continue to ignore UZR and WAR until they publish the actual data and give confidence numbers on the estimates so I can assess how meaningful they are.

        Vote -1 Vote +1

      • Sandy Kazmir says:

        @mcbrown

        The point is that if the error term for one season’s worth of data (1500 innings) is around 10 runs (hypothetical) then you’re essentially saying that all but the most extreme defenders are somewhere between + and – 5 runs of defense. Any player could fall anywhere in that range despite what that one year is actually saying. That essentially tells us nothing and is essentially the same thing as regressing a player to the mean for any number based on 10 plate appearances. A nice, neat number will be spit out, but it’s so heavily regressed that it’s virtually meaningless.

        Vote -1 Vote +1

      • mcbrown says:

        Anon21, that is a valid point – probably I should just be ignoring UZR until they publish more on it. That would certainly be more scientific.

        Sandy, I understand what you are saying, and I agree if the error band is that large (actually I think it would need to be larger for your hypothetical, since UZR seems to range from around -20 to +20). However, I would point out that if the 1 year error term is +/- 5 runs, then I estimate that the 3 year error term is +/- 8.7, or +/- 2.9 runs per season, which means that with 3 years of data only a few less players will be bunched within the error band of 0 than they are with 1 year of data. So by some variant of proof by contradiction (again, assuming we accept that 3 years of UZR data is “significant”) the 1 year error band must be lower.

        Vote -1 Vote +1

      • Cliff says:

        The problem is that the measurement error is systemic, which means it never goes away no matter how large the sample size. In fact, the larger the sample size, the worse it is, because the random error “drops away” and so the measurement error is assumed to not be there. The error is in judging the path of the ball- when the fielder has difficulty getting to the ball, it is systematically judged to be further away from the starting point of the defender. When the fielder reaches it easily, it is systematically judged to have been closer to the defender. That it correlates reversely with the true talent level.

        Vote -1 Vote +1

      • joe says:

        Cliff… great point. I think most people assume it’s random variation and just a matter of getting enough of a sample size to address it.

        But there are also a lot of sources of systemic error and it’s hard to understand how significant that might be:
        – input data bias (defining zone, speed of ball, runner speed)… I think mgl(?) has done a bit on this over at Tango’s site
        – scoring bias (what’s an error that feeds the errR component)
        – the model itself may have some systemic error
        – the park adjustment factors
        – not assessing the starting position of a player
        – dpR (which would seem to not fully isolate a player’s skill at a DP and mix in teammate ability and contextual issues)
        – armR is also contextual (score, # of outs will, pitcher on mound, hitter coming up will impact whether a runner tries to take a base on an outfielder)

        Vote -1 Vote +1

      • Tangotiger says:

        “- not assessing the starting position of a player”

        UZR takes the position that the starting spot of a player is a skill. And it assigns it to the player.

        You can agree or disagree. But, it’s not a bias. It’s a feature, insofar as UZR sees it.

        And yes, there are random biases (which are solved by larger sample size) and systematic biases (which are exacerbated by larger sample size) to recording data.

        There’s also random and systematic biases to the “eye test”.

        Pick your poison.

        Vote -1 Vote +1

    • Yirmiyahu says:

      +1.

      Seems to go against the spirit of the site to remove an article and kill what I assume was a lively debate.

      +9 Vote -1 Vote +1

      • mb21 says:

        I don’t think completely deleting the article does anybody any good. The same mistakes will again be made by I assume even writers here at Fangraphs and certainly other places. The comments pointed out the errors and all the article needed was an admission that UZR is not necessarily representative of what the player accomplished and all would have been good. Part of writing an article on a blog is accepting that you’re going to say things that are wrong. You can’t hide the mistakes that people have already seen. Leave them in plain view and correct them in an additional section of the article.

        +8 Vote -1 Vote +1

      • Sandy Kazmir says:

        Well said, mb. Some folks just can’t handle great dialogue which can be perceived as criticism. I’m sure if you give Bradley more time, he’ll shed that thin skin.

        Vote -1 Vote +1

      • Sandy Kazmir says:

        Well said, mb. Some folks just can’t handle great dialogue which can be perceived as criticism. I’m sure if you give Bradley more time, he’ll shed that thin skin.

        Vote -1 Vote +1

      • The taking down of the article nothing to do with its criticism. I didn’t even read the comments before requesting it be taken down. The article was simply not what either Dave or I would have intended for an “official” stance on UZR.

        It slipped by both of us and is not a reflection on how thick Bradley’s skin is. Bradley does great work for FanGraphs.

        Many people do not read the comments in an article and instead of leaving a piece up for everyone to read that had inaccuracies, it made more sense to take it down in this case. I’d rather we have time to get it right than put an addendum on the existing piece.

        Vote -1 Vote +1

      • Peerless Priceless says:

        But there is value in making mistakes and having a discussion about it. We can benefit from reading the post and the comments, and those less familiar with UZR or who have mistaken ideas about it have the most to gain from doing so.

        Plus: big fights online are fun.

        Vote -1 Vote +1

      • Mark Geoffriau says:

        I can understand not wanting to propagate confusion by leaving an errant article up (and relying on an addendum or the user comments to provide the correction), but simply yanking the whole page (and useful discussion) is a terribly uncreative solution.

        If Fangraphs is really that worried about misleading readers by leaving the article up, why not simply replace the text of the article with a brief explanation of why the original article was posted and then removed, and allow the discussion to continue? As it is, it’s hard not to assume that there’s some embarrassment over the original article and an unwillingness to let the discussion continue.

        +6 Vote -1 Vote +1

      • mb21 says:

        David, thanks for the reply. i appreciate the honesty even if I don’t agree with it. I hope others do too.

        Vote -1 Vote +1

      • Dave says:

        David, if the “official response” was what bothered fangraphs the most, why not just remove that part of the title and (like others have noted) add a disclaimer at the beginning of the article. You can’t learn from your mistakes if you cover them up.

        You really underestimate how many people read comments and follow the debates. I know I do it most of the time, and only recently have i started chiming in on blogs and its still rare that I do. But I always read the comments. Some of the best discussion happens after the jump.

        Vote -1 Vote +1

      • Dave says:

        my comment should be “official position” not “official response”…

        Vote -1 Vote +1

  6. LorenzoStDuBios says:

    THIS STINKS! THIS IS TOTAL BS! STICK TO YOUR GUNS RIGHT OR WRONG!

    -8 Vote -1 Vote +1

  7. AndyS says:

    I totally missed this series, when did it run?

    Vote -1 Vote +1

  8. Anon21 says:

    I don’t think there was anything wrong with the concept of doing “official positions” on various controversial topics that Fangraphs, rightly or wrongly, is perceived to be associated with. You just want to be sure that the people who formulate your official positions are careful to understand the stat or concept they’re explaining or making an argument about.

    Vote -1 Vote +1

  9. Eminor3rd says:

    What? The UZR article was great.

    Vote -1 Vote +1

    • Telo says:

      It failed to address major facets of UZR. If UZR was built on OBP quality data, perfectly known, binary data, the article would’ve been OK. But it didn’t address in a single breath the biggest shortcomings of UZR: that stringer data is inherently inaccurate and biased, and that the methodology is a black box and imperfect as well.

      He started from the assumption that UZR = what happened in the field. And that’s just so far from the truth.

      Vote -1 Vote +1

  10. Patricio says:

    So…can we take FIP out of WAR? No? Okay…I’ll just keep looking at the chat….

    Vote -1 Vote +1

    • Santos says:

      Why would you want to take FIP out of WAR? It only measures what a pitcher does. It doesn’t include runs allowed, which are influenced by defense. It measures strikeouts, walks, homeruns, and innings pitched. A pitcher doesn’t have control of much else (batted ball typed to a degree) so when evaluating a pitcher based on WAR you should only be measuring those things that the pitcher is responsible for. Don’t mix other players’ contributions into it.

      Vote -1 Vote +1

      • jimiu says:

        But don’t we take fielding into account when looking at WAR for hitters? Maybe I don’t understand hitters WAR clearly, but to my knowledge we don’t calculate a hitter’s WAR based on what he should have done based on walks, k, hr and/or hit trajectory. We base it on what actually happened and if a ground ball snuck under a fielders glove for a double that is as good as a double in the gap off the wall. Personally I think this is a fundamental issue with comparing WAR of a pitcher vs WAR of a hitter. One is based on what happened, the other is based on what theoretically may have happened if everything went according to a formula. I think FIP is useful, just not for determining who actually brought more value to a team.

        Vote -1 Vote +1

      • Jeremiah says:

        Is a pitcher really responsible for home runs? I mean, is he responsible for a home run any more than he is for a double? Sure, sometimes the double is a result of the fielder having poor range or getting a bad jump, but it is often the result of a well struck ball toward the gap.

        I think the distinction should be made that FIP measures the things which aren’t influenced by fielders at all. There are many factors that affect home runs, including atmospheric conditions and the ballpark, which the pitcher can’t control. A ball that is struck in a certain way at Petco might result in a flyout, whereas it would be a home run in GAB. That is not the pitcher’s fault.

        Vote -1 Vote +1

      • Santos says:

        Jeremiah,

        You are right about the HR thing to a degree. That is what xFIP is for. There is data that suggests HRs are a percentage of flyballs given up. xFIP assumes pitcher’s don’t have much control over giving up homeruns, but they do have control over giving up flyballs. The more flyballs, the more homeruns.

        Vote -1 Vote +1

      • CircleChange11 says:

        which are influenced by defense.

        They’re also influenced by the pitcher.

        We just don’t know the exact % of influence.

        Vote -1 Vote +1

      • mb21 says:

        The pitcher most certainly is responsible for the home run. The home run may be a factor of luck, but so are doubles, triples, and home runs for batters. We do not remove the luck in batter WAR because we want an accurate representation of the value that player provided. But when it comes to pitchers we suddenly adjust for luck as if the pitcher didn’t pitch horribly with runners in scoring position or had given up an abnormally high amount of home runs.

        WAR isn’t a true talent metric. There are various metrics that do that. WAR is a metric that combines value in various aspects of the game into one number that represents his total value. If a pitcher pitched horribly with guys on base, ignoring that does not capture how the pitcher performed. It’s only capturing how the pitcher would have performed given normal conditions.

        So people either need to be clearer when it comes to WAR or they need to use something for pitchers that more accurately reflects how the pitcher performed. You do not need to use ERA as some have suggested as the only alternative. FIP measures strikeouts, walks and home runs. That’s it. FIP happens to be a great predictor of future FIP/ERA or whatever you want, but it’s not telling us what happened. It’s only telling us about a subset of a pitcher’s statistics.

        Why not use Run Average and adjust it based on the defense when the pitcher is on the mound? If pitcher A gives up 60 runs in 120 innings, some of that is obviously the result of the defense. If the defense saved 5 runs while he was on the mound we know the pitcher was worth 65 runs in 120 innings.

        This way we’re not ignoring valuable information with regards to how a player performed. If we’re going to use WAR to talk about awards, you have to include the luck. Sequencing matters and while there may be little or no skill, it greatly affects the performance. FIP completely ignores this.

        Vote -1 Vote +1

    • Santos says:

      FIP is not what theoretically happened. It is what actually happened that the pitcher had control over. A hitter has more control over the hits he gets because hitting is a skill. There is data that suggests that pitchers do not have the same control over limiting hits (or, more accurately, controlling BABIP). Think about a pitcher giving up a single on the right side with a runner on second. The ball goes through the hole, the right fielder fields it and makes a poor throw that goes up the third base line. The runner scores. Then on the west coast, at the same time, a pitcher gives up a single on the right side with a runner on second. The right fielder fields the ball and makes a perfect throw and the runner is out at home. Assume that both of these hits are similar enough in nature. The only difference in the first runner scoring and the second runner being thrown out is the right fielder. Both pitchers did exactly the same thing, but because of their fielders they got different results. All FIP does it strip that part out, so runs which are a product of pitching and defense aren’t attributed solely to pitching. And lastly, strikeouts, walks, homeruns, and inning pitched aren’t theoretical. They actually happened.

      Vote -1 Vote +1

      • Patricio says:

        …and David Wright hits a liner to third, where E5 is playing, and it goes down the line for a double/triple. Pujols hits a sharp liner to third, where Zimmerman is playing, and it’s caught for an out. xBABIP is not used for hitters WAR, so why is FIP used for pitcher’s WAR? I’m not saying for or against, I’m asking for consistency. You can’t accurately measure the values of pitching to hitting if you are not, at the very least, consistent.

        Vote -1 Vote +1

      • Santos says:

        I don’t think that’s a 1 to 1 analogy though. WAR is supposed to tell you what happened to an individual player (it’s not 100% accurate, I know), but if you hit a ball that a fielder gets to and makes an out that instance is recorded. Sure your BABIP will go down, and analysts can see that and assume it will regress in to career norms in the future, but WAR will capture it. But having a defender make an out, where another defender may not, is, in my opinion not the same as crediting a pitcher with runs that are in part attributable to his own defense. Playing against the other team is how the game is set up and the stats measure what a player does against what the other team does. We shouldn’t measure what a player does against what his teammates did afterward. At least that’s how I feel about it. I don’t really have a good concrete answer for you. There is a ton of gray area and competing philosophies.

        Vote -1 Vote +1

      • Santos says:

        Also, why can’t I reply directly to a reply? I have to reply to my own original comment everytime.

        Vote -1 Vote +1

      • CircleChange11 says:

        This is exactly why I prefer to average fWAR and brWAR for pitchers (ajnd for batters for that matter).

        We don’t know the exact % that a pitcher influences batted bat success. fWAR (FIP) puts it at zero. Averaging them puts it at 50/50 or so. (i.e., averaging FIP and RA)

        I think it was a Tango comment that really cemented it for me … something like “I’d rather be half right, then all wrong.”

        In the big picture, it’s not going to be drastic for most pitchers, but it can be significant for pitchers that have abnormally high/low BABIPs for the entire season, even if they cannot repeat the success.

        It also matches my preference based on my experience as a pitcher. There wer enot very many times where batters didn;t get hits that I was doing a lot of things well. Likewise, there weren’t very many times where I was giving up lots of hits while doing things well.

        I think pitchers have quite a bit of influence on BABIP, even though i cannot prove the exact numerical % of influence.

        Vote -1 Vote +1

      • Patricio says:

        I understand where you are coming from, and I understand it is not a 1 to 1 analogy, but I think that you can’t have the same stat measure two different positions (hitters/pitchers) and say “We’ll ignore defense for you, as it is your team’s defense. We won’t ignore defense for you, as it was the opposition’s defense.”.

        What an opposing defender does is just as much out of your control as what your teammate does. You making solid contact on a pitch that goes for a hit 95% of the time and getting robbed by Zimmerman, is just as influential as your teammate having a noodle-arm instead of a cannon.

        If you had an incredibly unlucky season hitting, and your line drive rate was higher than ever, but your BABIP fell sharply, you still did not get a hit. You don’t deserve points for value, but you deserve points for expected value. The same goes for a pitcher. If you pitch solidly, and get groundballs all game long, but the opponent racked up hits and runs on you…you still gave up those runs. You weren’t valuable in the game due to your lack of luck. WAR is an advanced metric counting stat. It measures your Wins Above Replacement and not an “expected” variation of it.

        This is all my opinion of course. I think there should be WAR and xWAR. One with an output concerning what actually happened, while the other outputs what was expected to happen.

        Can we get xBABIP on this site? (If I missed it, please tell me where it is….)

        Vote -1 Vote +1

      • Santos says:

        I see what you’re saying. My only point of contention is when you say, ‘you still gave up those runs’. I just don’t see it that way. I don’t see it as “You” (i.e. the pitcher) gave up those runs. I see it as that team gave up those runs. That’s all. We have plenty of stats, and information to use at our disposal to make an informed decision based on the data. And I guess that’s all we can really ask for. Form a conclusion, based on information, and understand why we arrived where we did.

        Vote -1 Vote +1

      • Tangotiger says:

        “I’d rather be half right, then all wrong.”

        I have no recollection of saying it, but I’ll take credit for it anyway.

        Vote -1 Vote +1

      • Patricio says:

        Okay, I think I understand what you are saying now…and I can honestly understand the FIP argument much better. I do have concerns on the influence a pitcher has on balls in play. There are people that consistently out-perform their FIPs, and then there is Javier Vasquez. Also, if I’m in a situation, as a pitcher, that I don’t want the runner to score, and I have a guy with a cannon in right, I might throw something soft and away to a lefty, to ensure it goes that way. This is obviously hypothetical, but it leaves me thinking that pitchers do have some control.

        I still want my WAR and xWAR,

        Vote -1 Vote +1

      • Eric R says:

        “We’ll ignore defense for you, as it is your team’s defense. We won’t ignore defense for you, as it was the opposition’s defense.”

        I think the issue there is that the opposing defense for a batter is unlikely to be too far from average over the course of a season. A particular pitcher, barring be traded a couple times during the season is far more likely to have an extreme defense behind him.

        Vote -1 Vote +1

      • Tangotiger says:

        I agree with Eric R’s basic idea.

        Otherwise, what’s the alternative? Your own personal smushing system? Is that somehow going to give you more confidence than fWAR?

        And if you don’t like UZR, just replace it with whatever you’d use in your smushing system.

        Vote -1 Vote +1

    • Santos says:

      CircleChange,

      Exactly so instead of using a stat that says, ‘this is what some % of the pitcher and some % of the defense did’, they choose to use a stat that says, ‘this is what only the pitcher did’. It’s not perfect, but it isolates individual performance.

      Vote -1 Vote +1

      • Patricio says:

        Then let’s use “this is what only the hitter did” in WAR. There is the xBABIP metric that would do just that.

        I personally believe that WAR should be based on what happened. I don’t want to know how valuable a player would be if they played to their talent level–I’d call that xWAR. WAR should just measure what actually happened. Like everything else, look at it in 3 year measurements and then conclude what you think of the player.

        Vote -1 Vote +1

      • Santos says:

        I guess we aren’t seeing eye to eye. WAR is what happened. for hitters it measures what they did with the bat (wOBA), what they did on the field as best we can measure it right now (UZR and yes I’ve read the thousands of comments on UZR is not actually what happened, but it is as close to it as we have right now), and what they did on the basepath (Bsr). For pitchers it is similar. It’s what they did on the mound (how many strikeouts they had, how many walks they had, and how many homeruns they gave up, with how many innings they pitched). The part where we don’t see eye to eye (at least I think, and correct me if I’m wrong) is that you feel runs should be in there somewhere. I disagree because I think runs are a product of pitching and defense and we simply don’t know how much defense plays a part. The DIPS data does a pretty good job of showing that most pitchers don’t control the amount of hits they give up per balls in play. SIERA does a good job of breaking down the batted ball typed into a stat similar to FIP if that’s more to your liking. But I’ve never been behind the idea of using a stat that is heavily influenced by defense and attributing it all to an individual.

        Vote -1 Vote +1

      • CircleChange11 says:

        I agree. It happened. Even if it cannot be repeated consistently, it happened … and it happened as a result of actions of that player.

        On field value and true talent in a single season are two, sometimes very, different things.

        WAR should not be expected to represent both items simultaneously.

        Vote -1 Vote +1

      • Santos says:

        I agree with that as well. I think where we part ways is isolating the individual. I like it because if we are trying to evaluate one player we should only see what that player did. You don’t like it (in this specific instance, I don’t want to speak for you) because it leaves out some amount of information that is useful in evaluating that player, even if it is muddled with the contributions of other players. That information is still there we just haven’t found a way to distill it. But, we are both in agreement about the what happened part, we just have 2 different ways of getting there. Does that sound right?

        Vote -1 Vote +1

      • Patricio says:

        I think I agree with what you are saying. I posted something above regarding how I want two WARs, one telling what happened, the other telling what was expected to/should have happened.

        I appreciate how you are debating/discussing this. I really completely understand your viewpoint, and I honestly love WAR. I just wish that there were variations on the stat that allowed me to select which ones I deemed appropriate. I understand that’s asking for a lot, but that’s the thinking that led to the creation of WAR in the first place, right?

        Vote -1 Vote +1

      • mattinm says:

        Except IP includes defense. The worse a defense, the more hits will squeak through. Using BF would eliminate the defense entirely. (Yes, I know that the correlation between using BF and IP is exceptionally high, but it’s still wrong to say FIP measures only a pitchers abilities — that’s just what it mostly measures).

        Also, without digging up articles, there has been new theories on pitchers that shows some of them certainly can control BABIP (or at least batted-ball profile). A GB pitcher will inherently have a higher BABIP than a FB pitcher, especially depending on park-factors, but will likely give up fewer HRs and turn more GIDPs, likely evening things out.

        Vote -1 Vote +1

      • Santos says:

        Right. I thought I tried to stick to saying ‘mostly’ what pitchers can control. And I think I mentioned SIERA somewhere, which is like FIP but includes the batted ball types, no?

        Vote -1 Vote +1

    • Barkey Walker says:

      I can’t stand xFIP. If you want to do trajectory, go all the way to tERA, if you don’t want to do trajectory, do FIP. Not that I love FIP, but it at least has some purity to it.

      Vote -1 Vote +1

  11. Cherub Rock says:

    Dave got wasted, woke up, and now has a bad series to abort. I believe life starts in the introduction article. Too late now.

    -15 Vote -1 Vote +1

    • Nik says:

      yeah, I’m sure Dave Cameron is boozing it up right after his chemo treatments.

      Vote -1 Vote +1

      • Cherub Rock says:

        Wow. Was totally joking, does everyone think I actually believe Dave was drinking when he wrote the article? I was poking fun at the fact that he’s ending a series before it ‘matures’.

        Wow.

        Vote -1 Vote +1

  12. TK says:

    The problem with the series is that it was an attempt to intelligently respond to a long series of trolling (and it was worded poorly, but mostly the former).

    Vote -1 Vote +1

    • Telo says:

      Disagree. There are a lot of fundamental yet nuanced details about these stats (WAR, UZR) that people need to understand. This was attempt explaining some of those. Just not very well executed.

      Vote -1 Vote +1

      • CircleChange11 says:

        The big point, IMO, was that the error bars surrounding WAR make it not very effective for how many people use it … as THE final answer.

        I have made that point before … that WAR telling us that matt Holliday’s performance was between 4-6 WAR is not very useful, in the rgeard many people use WAR.

        The author pointing out that Ellsbury’s 8.X WAR was slighlty above Bautistra’s 8.Y fWAR, but really they were “equal” was important.

        It’s just that the cloudy posiiton of “you have to go with different criteria” to settle the MVP discussion (since the fWARs were similar) was not a very good “official position” to take.

        To me, the bigger take was that an inaccurate official position is just as bad (if not worse) as no official position. But, people like official positions in place of critical thinking.

        Vote -1 Vote +1

      • Eric R says:

        “I have made that point before … that WAR telling us that matt Holliday’s performance was between 4-6 WAR is not very useful”

        Well, if we’re assuming +/- 1 win, then it still tells us something. Lets assume that a 5 WAR player is equally likely to be any value between 4 and 6 and that a 5.5 WAR player is equally likely to be any particular value between 4.5 and 6.5.

        Then the first player is about:
        25% 4.0-4.5
        75% 4.5-6.0
        0% 6.0-6.5

        And the second player is about:
        0% 4.0-4.5
        75% 4.5-6.0
        25% 6.0-6.5

        So we then are then pretty sure about a quarter of the time, the 5.5 player is actually better just because sometimes he’s really a 6.0+ player and the other guy is never that good and sometimes the 5.0 player is under 4.5 and the other guy is never that bad.

        Probably not too different from looking at a guy with a .900 OPS and another at .875 and coming up with that the former player is probably alittle better and not likely too much worse than the latter [assuming park, league, position, play-time, baserunning, etc are equal]

        [my guess is that it is more like 70% between 4.5 and 5.5 and 15% each for 4.0-4.5 and 5.5-6.0]

        Vote -1 Vote +1

  13. jake says:

    Dave, your BABIP just went way down.

    Vote -1 Vote +1

  14. Christo says:

    Well…just don’t take the series, split it into two series and call one something like Qwikster. That seems to cause problems and upset people.

    +8 Vote -1 Vote +1

  15. juan pierre's mustache says:

    will there be some sort of article featuring opinions from several writers on the 2011 awards? i don’t so much care who actually wins, but i think it would be really interesting to see some back and forth, see who focuses on which stats and why, and to what extent there is consensus among the staff on various picks. i think this would allow several writers to be a little more direct as to why they think player X is the pick as opposed to one person trying to reach a conclusion without throwing out the other valid arguments against. it’s tough to both write a well-founded statistical article which has to acknowledge that no one player is really the single absolute correct pick while still coming to an official position that isn’t, “well, it’s probably him, but maybe not”.

    Vote -1 Vote +1

    • DD says:

      I agree, this should have been the focus from day 1. I believe Dave said they will have something to this effect on Friday. It would be interesting to see how differently each writer attacks the ballot, considering they all are operating under the same sabermetric umbrella.

      Vote -1 Vote +1

  16. The series was obviously well-meaning, but given that the a sabermetric community already fights the stereotype of nerds who self-righteously believe they are always right, an “Offiicial Position” series on something so subjective as post-season awards was a tough sell from the get go. It would have had to have been perfectly and extremely carefully executed, which is a high bar. I appreciate the idea, and perhaps in a different context it could be recycled effectively, but I am glad you didn’t decide to force it through to conclusion.

    Vote -1 Vote +1

  17. Matt says:

    Anyone take a screenshot or have a browser cache of the original article? I’d like to read it through and formulate thoughts.

    Vote -1 Vote +1

  18. Notrotographs says:

    Are you going to quit trolling everything Dave posts?

    Vote -1 Vote +1

  19. Jason says:

    This is really disappointing.

    Bradley’s piece on UZR was wrong as many of us pointed out in the now nuked thread. The discussion of why he was wrong was really useful.

    In science we have the peer review system, where mistakes like this are usually caught before they go to press. This is a nice buffer that often saves the author from embarrassment and assures the reader of some level of quality. Unfortunately, Bradley’s peer review was done publicly after publication, and was obviously embarrassing.

    But what is the reason for taking the article and ensuing down? Is it to save Bradley from embarrassment? We all saw it, the cat is out of the bag. Just publish a corrigendum explaining why the article was wrong so that it can be a reference in the future and we can all learn from the experience.

    I fear that Bradley’s error is actually a larger and more pervasive error here on FanGraphs. Bradley’s error was obvious because he made his faulty argument for using UZR as a record of what actually happened explicit. However, Bradley’s error is made every time someone uses WAR to assess what actually happened in a season. It is the same problem. Is FanGraphs ready to abandon WAR as a record of what happened in a season? Your discussion of who should be the American League MVP gives no indication of this.

    +6 Vote -1 Vote +1

    • Peerless Priceless says:

      I would just add to Jason’s comment that although in academia articles are peer-reviewed prior to publication, they continue to be peer reviewed after publication. That’s why publishing is important, afterall — to give the community the opportunity to test results and scrutinize the thinking. The process of critique, even when it is negative in the sense of pointing out mistakes or false starts, is how we arrive at progress.

      Vote -1 Vote +1

      • Jason says:

        Absolutely. We progress by continually testing our ideas against alternative explanations.

        Baseball evaluation differs in troubling ways from real science. A lot of “advanced statistics” come from unpublished data. Further, confidence in estimates is never published. As a peer reviewer, I would reject at as all conjecture unless methods, and data are properly made available to scrutiny.

        I understand that this is not science and there are monetary reasons why some things must remain proprietary. But you could at least publish confidence intervals. Where is the harm in that?

        As it is, you publish estimates and ask that we “just trust you”. I am unwilling to do that.

        Vote -1 Vote +1

  20. Richie says:

    So what the heck happened?? In journalistic terms?

    (it seems someone did a pompously poor job on an article, and rather than letting him and/or subsequent authors learn from his mistakes, they’re pulling the whole series; which strikes one as an overreaction; what do I have wrong or missing in that?)

    Vote -1 Vote +1

  21. Tangotiger says:

    “Unfortunately, Bradley’s peer review was done publicly after publication, and was obviously embarrassing. ”

    This is actually a Fortunately, not Unfortunately.

    I live and die on the informal public peer review process. At my blog, I’ve got access to thousands of readers, all willing to put up with style, so they can learn, and *I* can learn. I love those readers.

    At Fangraphs, there’s ten (hundred?) times as many readers. The quick and fierce reaction to the author missing the glaring aspect of measurement error was fantastic. It actually stopped embarrassment before it could fester into something huge.

    ***

    There were two things wrong in that thread:
    1. As Dave noted, taking an “official” stand on something that has multiple and reasonable points of view is akin to taking a political stance. You’ve got a 60/40 decision to make, and the official stance cements it at 100/0.

    For example, my “official” stance on fWAR and rWAR is to split the difference and call it a day. You can make reasonable arguments on either side, and instead of saying there’s only one right answer, I can just acknowledge that everyone has a valid point. Not to mention that there’s an uncertainty level to begin with anyway with any metric.

    Anyway, Dave recognized the path that this series was going to go, and put a stop to it. Again, this is a good thing.

    2. The author did not step in to stop the commenter runaway train by acknowledging that there was a glaring hole with the non-mention of the measurement error.

    (The measurement error for those not following is that we don’t know how many opportunities Tulo actually has in any given year, nor the “quality” of those opportunities. While we can figure out on the batting side what kind of opps he has, facing Halladay, facing Lee, etc, we have a problem figuring out what kind of chance Tulo has on any ball in play. So, we ESTIMATE each and every single ball in play and say, “ok, that one had a 90% chance of making an out, and that one had a 20% chance”. But, we don’t REALLY know that it was 90%. It could have been 80% or 95%. It’s a huge difference. Unlike batter v pitcher where if you face Halladay, you know, plus or minus 1%, the chance of getting on base against Halladay. Measurement error, if random, is of course “solved” by sample size. If systematic, it’s actually made worse by sample size.)

    Not to make this about me, but, what the heck, since I’m here: on my blog, this rarely happens. That’s because I have way too much free time at the office, and so, I respond almost immediately to any comment. And before there’s a chance that a legitimate comment can then be misconstrued into something else which leads to misunderstanding or just plain trolling, I give my answer, and hopefully, we can move forward.

    ***

    What happened with the comments was well-intentioned, and well-done at the beginning, but then it was simply slipping away into summary conclusions. Not that I necessarily blame the commenters, because they don’t have the time or inclination to swim in the deep end like I do. They want some clarity, but instead they got more confusion.

    +29 Vote -1 Vote +1

    • Richie says:

      I thought Fangraphs was doing a good job with their “official” position. That is, ‘we know our metrics are imperfect, but this is exactly where they take us, and how’. That was my impression of what they intended with this series.

      Vote -1 Vote +1

    • JDanger says:

      “…So, we ESTIMATE each and every single ball in play and say, “ok, that one had a 90% chance of making an out, and that one had a 20% chance”. But, we don’t REALLY know that it was 90%. It could have been 80% or 95%. It’s a huge difference….”

      I think this is where the “UZR is what actually happened” vs. “not what actually happened” confusion occurs. UZR measures what happens, but the denominator is an estimate and can’t always be trusted.

      Vote -1 Vote +1

    • I may not have done a good job of it, but I did attempt to “step in to stop the commenter runaway train by acknowledging that there was a glaring hole with the non-mention of the measurement error.”

      Unfortunately, the commenters outnumbered me 100 to 1, so I couldn’t get my responses out quick enough.

      I hadn’t seen MGL’s article on the Book Blog — I wish I had, it would have greatly focused my piece. The problem was — and this was not my decision, but my mistake — I had missed the mark on the series so severely that the piece was taken down as I was even adjusting it to reflect MGL’s article and my failure to give proper thought to the measurement error issue.

      It was a mistake, and I apologize. However, I do not entirely apologize for my stance on single-year UZR. I think the Curtis Granderson analogy — which I would have loved to expand on — well represented my case and requires additional scrutiny. I am not a fan of dismissing UZR just because there’s a risk of measurement error. If there’s a huge discrepancy between the scouts and UZR, doing what Mark Simon did (actually looking at where the differences come from) is much more valuable than just throwing up hands and surrendered to the unknown.

      Anyway, thanks for weighing in Tom. I hope we can continue this discussion in the near future.

      Vote -1 Vote +1

      • Jason says:

        Bradley,

        Because the measurement error in UZR is real, and the estimates it provides are imprecise, it can only provide an estimate of what happened to some degree of confidence. The confidence interval is a function of sample size and of the magnitude of the error (for the portion of the variance solely due to measurement error). Further, if the error is systematic bias and not random, no amount of sampling will help. At any rate, whether one year of UZR is useful for comparing whether one player actually performed better than some other player is something that ought to be quantified, not asserted. We can know the answer to this question. The gatherers of the raw data for UZR have the ability to measure it. They should do so.

        You may claim that one year of data is meaningful for comparing between players, but we won’t know it until the error is quantified and published.

        Vote -1 Vote +1

      • tangotiger says:

        BRadley, I agree that it’s a tough position to be able to respond when the responses came fast and furious.

        I should have been less strong toward you in that regard.

        Vote -1 Vote +1

      • Umpjacked says:

        Could you post links to MGL’s book blog piece and whatever Mark Simon did?

        Vote -1 Vote +1

  22. Matt says:

    Look, it’s fine that the UZR article was taken down, but I think everyone here wants at least a revision or follow-up to the whole discussion. Not an official position or anything; just a better version of the article that maybe responds so some of the more interesting comments in the thread.

    Vote -1 Vote +1

  23. Hurtlocker says:

    I thought the original concept of an “official” position is valuable. Anyone who loves baseball, and I assume everyone that reads this website does, knows that the metrics NEVER tell the whole story. We watch with our eyes and hearts, not with our calculators. All the metrics here do is provide a lens to focus the obscure nuances of a complex game into a tangible “score”. After all, the whole point of playing baseball is to score.

    Vote -1 Vote +1

  24. RMR says:

    It would be nice to see the article and comments reposted with a leading “Update” explaining the situation and directing readers to the lively discussion that followed. As others have mentioned, it is precisely those types of decisions that provide fertile opportunity for learning and developing understanding.

    Vote -1 Vote +1

  25. Sam says:

    I am a fan of Fangraphs. Like most of its fans, I am not qualified to assess the validity or robustness of the various and sundry statistical metrics/algorithms touted by Fangraphs analysts (and commenters) as being THE accurate measurement of whatever is attempting to be quantified.

    All that being said, as a skeptic, I find this comment stream representative of most on this site — it’s an argument about the applicability of some statistical measure that ultimately was generated with the assistance of some subjective component(s). And the pro- and con- arguments, which clearly take time to craft, prove to me (at least) that the baseball stat game is a political one. The fact that arguments inevitably ensue about some writer’s “objective” stat-based perspective demonstrates that there may not be a “best” metric. These debates have become intractable and ad hominem and tiresome.

    More importantly, I think the blossoming of “better” or “more accurate” stats is great, but it kinda undercuts what this website does. Put simply: sometimes it feels like no metric is legit because there will be a better one announced in a few weeks. I get that that is good, and is sorta what science “does.” But if something better is always soon to arrive, why should I buy into/trust what just came down the pipe? And how do I even know that what just came was actually legit and not entirely flawed.

    Has Fangraphs ever thought of organizing it’s narrative analysis into “camps” based on metric allegiance? At least that way we know the assumptions/potential flaws that are relied upon in the author’s conclusions, instead of endlessly questioning the propriety of the raw data used and the algorithm that churns it.

    Vote -1 Vote +1

  26. MightySlacker says:

    With the article title, totally thought this was gonna to be Bumgarner’s late admission into the trade value series. How dare you ambush me with all this reading

    Vote -1 Vote +1

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>