New Pitcher Value Stats

The Value (WAR) section for pitchers has been updated to help further break out a pitcher’s value. These stats are now available in the leaderboards and player pages.

We’ve added the following stats:

RA9-Wins – Wins above replacement calculated with RA9. This is both park adjusted and league adjusted.

BIP-Wins – Wins above average based on BABIP. This is both park and league adjusted.

LOB-Wins – Sequencing (and miscellaneous) wins above average calculated as RA9Wins – WAR – BIPWins.

FDP-Wins (Fielder Dependent Wins) – This is the full difference between RA9Wins and WAR, or BIPWins + LOBWins.

WAR – Wins Above Replacement remains unchanged.

Dave Cameron will have two posts with more details about these new stats at 9:00am and 10:00am and will be answering questions during his 12:00pm chat.

The full formula for BIP-Wins is:

((((H-HR)*(w1B * p1B + w2B * pxBH) ) / (TBF – HR – BB – HBP – SO)) – lgwBABIP) * (TBF – HR – BB – HBP – SO) / PF / RtW * -1

p1B = singles as a percentage of hits that are not home runs
pxBH = doubles and triples as a percentage of hits that are not home runs
w1B = linear weight value of a single
w2B = linear weight value of a double
RtW = runs to wins converter
PF = park factor
lgwBABIP = league average: ((H-HR)*(w1B * p1B + w2B * pxBH)) / (TBF – HR – BB – HBP – SO)

Print This Post

David Appelman is the creator of FanGraphs.

28 Responses to “New Pitcher Value Stats”

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

    head explode

    +10 Vote -1 Vote +1

  2. Graham says:

    Wouldn’t “percentage of singles that are not home runs” and “percentage of doubles and triples that are not home runs” be 100% in every case?

    Are we to assume those two definitions should mean “percentage of hits that are not home runs” and “percentage of extra base hits that are not home runs?”

    Vote -1 Vote +1

  3. odditie says:

    Any update on Quality Start data being added :)

    Vote -1 Vote +1

  4. JDanger says:

    This is the day your life will surely change.

    Vote -1 Vote +1

  5. andys says:

    Great work, but why not allow other versions of WAR too? Why not have, say, SIERAWAR, xFIPWAR, etc.? I think these metrics are VASTLY overdue.

    Vote -1 Vote +1

    • JDanger says:

      I think the idea is to keep WAR strictly to measuring things that actually occurred, rather than involving regressions.

      Vote -1 Vote +1

      • andys says:

        xFIP is based strictly to things that occurred. There are only three variables – BB, K, and FB. The difference is you’re using FB instead of HR.

        Vote -1 Vote +1

      • JDanger says:

        OK sure, but it’s intended to be a regressed HR rate. You are assigning run values of HR (that have not necessarily occurred) to 11% of the pitcher’s FB%, rather than actual HR allowed.

        Vote -1 Vote +1

      • andys says:

        But there’s still no regression happening. It’s just treating all fly balls the same, in the same way that FIP treats all non-HR balls the same.

        FIP: HR balls, non HR balls each are being treated as the only two groups of contact.
        xFIP: Fly balls, non-fly balls each are being treated as the only two groups of contact.

        I fail to see the difference.

        Vote -1 Vote +1

      • Tangotiger says:

        FIP is agnostic on what happens to fieldable balls. That’s because FIP only cares about non-fieldable PA. Just like OBP is agnostic on HR and 3B being different from 1B.

        xFIP is agnostic on the actual outcome of a flyball. If you find that valuable, then fine. I don’t.

        Vote -1 Vote +1

  6. themiddle54 says:

    Since workload informs WAR to a large degree, would be great if the Value tab included IP. I find it handy to look at that stat when putting into context in my brain how/why a certain pitcher is valuable.

    Vote -1 Vote +1

  7. paranoiaagent says:

    any idea when these will be available in the dashboard stat customizer?

    Vote -1 Vote +1

  8. Anon says:

    Excellent additions.

    My vote for the next stats on Fangraphs is for stats that consider quality of opposition.

    Vote -1 Vote +1

  9. Will says:

    For RA/9, don’t you think it would make sense to somehow adjust for runners left on base after the pitcher leaves that subsequently score? RA seems to give credit for pitchers on teams with good bullpens.

    Vote -1 Vote +1

    • Tangotiger says:

      You are totally correct.

      Better than using RA9 is RE24 (already on Fangraphs, and already park-adjusted). For a pitcher that completes his own innings, then RE24 and RA9 are identical.

      As best as I understand it, since RE24 is only available for a good deal, but not all, of baseball history, then RA9 was chosen.

      But, I think it’s a matter of time until David will update his code to use RE24 when available.

      Vote -1 Vote +1

  10. Nathan Nathan says:

    Thanks for the formula. Might we see a MathJax formatted equation as well? It would make it a lot easier to read, I think.

    Vote -1 Vote +1

  11. AustinRHL says:

    Ooh. I like these! I assume that in due time, these statistics will be available to add to our customized dashboards?

    Vote -1 Vote +1

  12. Snowblind says:

    I have no idea what a “good” number for any of these are. Is that going to vary wildly pitcher to pitcher? Year to year? Something else?

    For example, I have a pretty good idea of the talent level and contribution of a 0 WAR player, or a 2 WAR player, or a 5 WAR player. I have a rough idea of the skill range of a guy with a 3.00 FIP, or a .320 OBP, or a RC+ of 100.

    What the heck is a “good” number for FDP, RA9-Wins, or any of the other shiny new stats?

    Vote -1 Vote +1

    • TFINY says:

      I believe that 0 is the baseline; at 0 FDP, the pitcher is performing precisely as his WAR dictates, when taking into account BABIP and LOB. Any positive number will make him better than his WAR declares.

      Vote -1 Vote +1

    • Tangotiger says:

      RA9-wins is on the same scale as WAR.

      FDP and its subcomponents (LOB-wins and BIP-wins) are centered around 0 being average.

      Vote -1 Vote +1

  13. noseeum says:

    I’m sure cursing is not generally appreciated on the boards here, so I apologize in advance. But it must be said:
    This is fucking awesome.

    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>