The Minus Stats

Last week, the big boss around here rolled out the newest stats to hit the site; ERA-, FIP-, and xFIP-. I know, I know, the last thing FanGraphs needs is more numbers with weird acronyms, and now we’re introducing stats that have acronyms followed by a mark that often indicates that subtraction is on the way? Who wants to be caught in conversation saying “Cliff Lee had a FIP minus of 62 last year”? Not me, that’s for sure.

But, that’s the beauty of these new numbers – they actually give us the ability to describe advanced pitching metrics in plain english. You can use them without ever actually using the acronym; for example, “Cliff Lee was 38 percent better than average at things pitchers have the most control over.” You’ve just explained Lee’s FIP- in one not-overly-confusing sentence, and you’ve done it without causing the person you’re conversing with to yell “NERD!” and walk away.

While our goal isn’t to proselytize the masses, statistical analysis is gaining a foothold in the mainstream, and more and more people are becoming interested in the conclusions that these kinds of metrics can offer. However, I’ll be the first to admit that our community has an alphabet soup problem – there are so many numbers with varying acronyms that it can be dizzying to try and keep up with. The FanGraphs Library that Steve Slowinski has created (and is always adding to) is a fantastic resource, but if you’re just talking with someone in conversation, odds are good that you’re not going to pull out your iPad and show the other person what Steve wrote about FIP-. I mean, that’s okay, but it seems awkward to me. I don’t know. Try it and let me know how it works – now I’m curious.

If you don’t want to go that route, though, these three metrics (along with wRC+, which is the offensive equivalent of this kind of statistic) give you the option of quoting a nerdy tool without sounding like one. Personally, I think the concepts that are revealed by the numbers are far more important than the actual number itself; I don’t really care what James Loney‘s wOBA against LHPs is, but I do want to be able to tell Dodger fans that he probably shouldn’t hit clean-up against southpaws.

This is where these three numbers will shine. There are a lot of people out there who don’t want to hear us try to pronounce BABIP while talking about Francisco Liriano, xFIP- gives you the chance to say that Liriano was “30 percent better than the average pitcher last year” without having to throw an acronym out there. From there, assuming the person you’re talking to doesn’t think you’re totally insane, you might have a chance to explain what goes into xFIP- (without even calling it that if you don’t want to), noting how Liriano’s walk rate, strikeout rate, and groundball rate were all excellent last year, and these are the kinds of things that are most consistent from year to year.

Not everyone’s going to want to hear it, and that’s fine. But I think there are people out there who would be open to “sabermetric ideas” but have no interest in trying to discern the difference between a FIP, a VORP, or an RE/24. There can be a legitimate barrier to entry created by the terms that we’re all familiar with, and that’s unfortunate, as I think some people who would enjoy the types of discussions we end up having can walk away before it ever gets started because they simply don’t understand the language we’re speaking.

That doesn’t mean we’re going to stop using FIP or wOBA around here, of course. However, I do think there’s room for measures that can help us converse more regularly with the casual fan, and that’s one of the reasons I’m excited to have things like FIP- and xFIP- on the site. They might look like more horribly named esoteric stats, but they’re actually an opportunity to use these concepts in a regular conversation. I look forward to using something like xFIP- without ever having to pronounce it.




Print This Post



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


44 Responses to “The Minus Stats”

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

    “30 percent better than the average pitcher last year”
    edit*

    Vote -1 Vote +1

  2. Jeffrey Gross says:

    This article was 30% better than i expected

    +20 Vote -1 Vote +1

  3. jaywrong says:

    i think fangraphs just got 30% better than any other sabermetric site out there.

    awesome.

    Vote -1 Vote +1

  4. neuter_your_dogma says:

    Great stat. My only question is if you (and we) don’t want to get “caught in conversation saying “Cliff Lee had a FIP minus of 62 last year,”, then why use the dreaded sign at all?

    Why not use something like “Goodness Factor?” :)

    Vote -1 Vote +1

    • Ben Duronio says:

      This is a topic that rarely gets discussed enough. The names of these numbers do matter, and the more acronyms and lower case letters followed by capitalized letters or acronyms followed by a +/- sign pushes the general fans away from this stuff. It really does segregate the population while the intention should be to get everyone on pace.

      The same people who get upset at managers and ex-players for not understanding these numbers are the ones who aren’t trying to speak to the masses, necessarily. That’s my opinion though, but I think it has some relevance in this discussion.

      Vote -1 Vote +1

      • Wraithpk says:

        The +’s weren’t created by baseball statisticians, they use those to mean indexed statistics in economics too.

        Vote -1 Vote +1

      • Ben Duronio says:

        Obviously true, which is why the average fan wants it to have nothing to do with baseball. Lots of my close friends, as well as many around the internet, get turned off very easily by words used in the business world.

        Vote -1 Vote +1

  5. JOEiq says:

    You don’t need to open by defending it! We appreciate all stats you have. Except for maybe …(tried for some comic relief I can’t think of any that aren’t important)

    Vote -1 Vote +1

  6. mattybobo says:

    My only issue here is… where’s the argument for making it “minus” instead of “plus”? I know that a lower number for the raw pitching stats is better, so I understand the idea of making it minus, because negative is lower than positive. Duh.
    But I have a couple problems. First, there are already a lot of stats with pluses on the end. So stat people are familiar with representing it that way, and making anything more than 100 above average. It just makes more sense this way. It’s the way IQ is presented. Further, it’s similar to how tERA uses the earned runs allowed scale instead of the runs allowed scale–it’s just more familiar and makes things that much easier to compare.
    Lastly, if you want to say that Cliff Lee was 38 percent better than average, doesn’t an FIP+ of 138 do a much cleaner job of this than an FIP- of 62? 100 + 38 = 138. Simple, elegant. I don’t see the need to flip things around.

    Vote -1 Vote +1

    • Wraithpk says:

      The problem is that the original formulation of ERA+ was done wrong. It’s hard at this point to change it, because people are used to ERA+ of over 100 to mean above average, so it’s easier to make a new statistic with a slightly different name.

      Vote -1 Vote +1

    • fredsbank says:

      in the offensive side of the game, you want higher numbers, hence OPS+, wRC+ etc; its easy to see that the higher of those a batter has the better he was offensively.

      the opposite is then true of the primary pitching stats, ERA- (whether it should be or not…) FIP-, xFIP-; the smaller the number, the better a pitcher did.

      that doesnt seem like that hard of a concept to grasp…

      Vote -1 Vote +1

      • Shane says:

        They did acknowledge batters want higher numbers, pitchers want lower. It’s just weird that you’re representing a % better than something by showing it in the – direction. Nothing wrong with it, but I think it would be more general public friendly if it was scaled above 100.

        Vote -1 Vote +1

  7. Nick44 says:

    But wouldn’t a FIP+ of 125 equal a FIP- of 80?

    FIP+ would say that the league average FIP was 25% higher than the pitcher’s FIP. FIP- says that the pitcher’s FIP was 20% lower than the league average.

    The number are not symmetrical. Because you can have a FIP+ greater than 200, but not a FIP- less than zero.

    Am I thinking about this correctly?

    Vote -1 Vote +1

    • Wraithpk says:

      Yes you are. A FIP+ of 200 would mean your FIP is half that of average. For FIP- it would be 50. FIP- is more intuitive, and as was mentioned in the initial article, averaging with the – stats works correctly, but doesn’t work for the + stats.

      Vote -1 Vote +1

  8. Kazinski says:

    Liriano was “30 percent better than the average pitcher last year”

    That still doesn’t tell me much without knowing what the standard deviation is. With an even distribution then Liriano would just be in the 80th percentile.

    Maybe we need something like xFIPp where 50 is average(median) and 99 is 99th percentile.

    Vote -1 Vote +1

    • AJS says:

      This suggestion is far more confusing to the average fan.

      Vote -1 Vote +1

      • harry says:

        I disagree. With SATs + all the standardized tests these days, percentiles are well known.

        Vote -1 Vote +1

      • Shane says:

        As someone who teaches introductory physics lab, regardless of the SAT, students do not seem know percentiles well. Standard deviation would be nice, but to the average person it is unknown.

        Vote -1 Vote +1

      • Llewdor says:

        This is true, but appalling. How can people think they understand trends or averages without knowing the standard deviation?

        Vote -1 Vote +1

  9. Devern Hansack says:

    I feel like it would be more intuitive as follows:
    +30 –> 30% better than average
    0 –> average
    -30 –> 30% worse than average
    It’d be a lot easier for people to grasp, IMO.

    Vote -1 Vote +1

  10. Dan says:

    My wife won’t let me name our new dog FIP or BABIP, maybe xFIP-?

    +9 Vote -1 Vote +1

  11. lester bangs says:

    This reminds me of when Prince changed his name to a symbol that was supposedly unpronounceable.

    Vote -1 Vote +1

    • Shane says:

      “supposedly unpronounceable” How did you pronounce it?

      Going off topic slightly, but I saw Prince in concert two weeks ago. He got on the PA system to say hello during the show set-up and 10,000 women went bat shit crazy. Never saw anything like that in my life before. Prince is 90% better at live performances than other artistI’ve seen.

      Vote -1 Vote +1

  12. If the concern is bringing sabrmetric analysis to the “average fan” (whatever that means), all the naming conventions in the world won’t mean ess. Sabrmetrically impoverished stats like RBIs and wins have the benefit of being extremely easy to comprehend; you can point to something on the field and immediately understand the correlation to the stat. Most advanced stats just don’t have this benefit, if that’s the kind of thing you value, which I think under this framework most “average fans” do. Even translating it into the language of “Francisco Liriano was 30% better than the average pitcher at doing the things he could control last year” doesn’t point anything specific that happens on the field. As the conversation is currently constructed, the “average fan” just doesn’t have the conceptual apparatus to apprehend and appreciate the meaning of a sentence like that. And I’m not so sure that’s such a bad thing. What’s wrong with leaving the sabrmetric analysis for the sabrmetricians?

    Vote -1 Vote +1

  13. pft says:

    The library is great but you really need to update the old glossary and make it easy for folks to get a quick summary of the stats.

    Vote -1 Vote +1

  14. ang says:

    I don’t really like this “38% better” notation, since you really mean “38% fewer expected runs”. It’s like saying something that’s “100% off” is “twice as good” (100% better). Actually it’s infinitely better.

    To get something like Devern Hansack’s suggestion above, you should use FIP- -100 for those worse than average and FIP+ -100 for those better than average. In mathematical notation, this is
    FIP* = [FIP+ -100]^+ – [FIP- -100]^+,
    where ^+ is the ramp function.

    Vote -1 Vote +1

  15. Heathcliff says:

    I think one good way to open the minds of non-sabers is to just point out the obviously major flaws of the basic stats (in a non-douchey way). Explain in basic terms why Wins is a poorly named stat that doesn’t actually measure what it’s name implies. If you have a computer handy during the conversation, give some basic supporting data for this like how Phil Hughes won 18 games last year due to high run support, and actually go through basic game logs to show it. Talk about how RBIs are very context-dependent and show some of BP’s stats related to RBI chances and RBI% for each player. Talk about how Juan Pierre led the league in outs in some of the years where he actually had a decent batting average. Point out to them every time a guy who drew a walk happened to score later in the inning and that’s a major part of why walks are valuable.

    And don’t ever mention any of the saber stats when discussing the flaws of the basic stats. Just get them to see the flaws of the old stats. There is no need to use FIP and WAR and the wRC+ to get the point across. That could just overwhelm or confuse them. The nice thing is that the flaws of the basic stats are easily explained in normal language without any need for newer stats.

    Before they can begin to embrace the new stuff, you need to convince them to question the old stuff. If they accept that the old stuff doesn’t work so well, then perhaps they will want to know what works better.

    It is understandable that someone would be hesitant to accept a new approach to solving an old problem. But if you convince them the the old method of solving the problem doesn’t actually work so well, well then maybe they will want to seek out a better solution.

    And if they have no interest in accepting the flaws of the old stuff, well then your friend is probably named Murray Chass.

    Vote -1 Vote +1

  16. Heathcliff says:

    Also, once you get them to kinda-sorta question the old stuff, perhaps you should introduce them to the archives of FJM. It has lost some of its luster since it is outdated and some of the things they have criticized have been proven wrong in recent years, but overall those archives still do a pretty good job of getting people to question conventional baseball wisdom/cliches and laugh at the same time.

    Vote -1 Vote +1

  17. Russ says:

    These new stats are great and much appreciated. I do have a few questions. Is there a reason why we don’t have a tERA-? Is there any chance that we’ll see this added to Fangraphs?

    Vote -1 Vote +1

  18. CircleChange11 says:

    I agree with others that have stated that it’s overkill. Why not just have a stat called “percentile”? Cut through all the acronyms, and math signs, and put it in simple terms.

    I don’t see how this would be appealing to anyone outside of the hardcore stats geek (said with affection).

    I do, however, appreciate and admire the continual work you guys do in this regard. Always trying to find the best way.

    Vote -1 Vote +1

  19. mike wants wins says:

    Yup, I HATE the names for these stats. Just brutal and not full of meaning. Y’all should find some fluff (non-stats person) to name your stats. The Twins can’t even figure out what these acronyms mean, how do you expect fans to?

    Vote -1 Vote +1

  20. Mr wOBAto says:

    Since every park has different weights for things like BB% is FIP- going to be weighted by park factor, Also will their be a statistic like sOPS where a players performance is weighted against the league average in that situation?

    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>

Current ye@r *