Sports Reference Blog

WAR 2.0: WAR Upgraded on Baseball-Reference.com

Posted by sean on May 4, 2012

Baseball-Reference.com WAR Explained - Baseball-Reference.com

As many have noticed, we have not been updating the 2012 WAR data in-season. The reason for this is that we have been working on a major upgrade to our WAR framework. The link above gives a very detailed rundown including: how to compute each part of our WAR formula, our reasoning on making the changes, downloads for historical WAR data, charts of the various constants and factors we use, and a chart of differences between the new B-R WAR, the old B-R WAR and FanGraphs WAR.

We also have pages describing the details of the components as well.
WAR for Position Players
WAR for Pitchers
How we compute wRAA for WAR

All in all, these are major improvements to the system and while there are dozens of details, here are the main points of difference between our old framework and the new one.

  • Switch from BaseRuns for batting to an advanced wRAA metric.
  • Folding ROE, infield singles, SO vs. Non-SO into wRAA.
  • Excluding pitchers' hitting and averaging by league rather than year from the league averages for wOBA and wRAA.
  • Estimation of CS numbers for leagues they are missing.
  • Use of Baseball Info Solutions Defensive Runs Saved from 2003-present (in our view the most advanced defensive metric).
  • Use of a player-influenced runs to win conversion for both batters and pitchers based on PythagenPat.
  • Use of a player-specific park factor for pitchers weighted by actual appearances in each park.
  • After a preliminary WAR calculation, we fine-tune the replacement level on a playing time basis, so the total WAR in each league is very consistent year-to-year.
  • I've changed the oWAR and dWAR formulations. oWAR is now called ndWAR for (no-defense WAR), but is the same otherwise. dWAR now contains the position component of WAR, so the Career Leaderboard is now dominated by SS, C and other great defensive players.

As always our goal is accuracy and we think this gives us a more accurate system. As suggestions, tweaks, and more data comes in we'll continue to adjust the system.

One other thing we are doing is providing a daily updated download of our WAR totals for every major league player. Batters and Pitchers. Feel free to use it how you wish so long as you cite us. If you prefer the old numbers they are available for download from About index.

I owe a big thanks to many people, but wanted to single out the following.
Sean Smith, who provided us with the original WAR incarnation on the site and provided much of the methodology used including, but not limited to TZR (pre-2003), GIDP runs, Baserunning Runs, Replacement Level, and Pitcher WAR.
Baseball Info Solutions licenses up their tremendous defensive metrics.
Tom Tango, who provided us with a great deal of feedback related to modifying and improving on wRAA and runs-to-wins calculations.
Neil Paine, who walked through each step of this rethinking with me.
RetroSheet and Pete Palmer for the raw batting, pitching, and fielding numbers that go into the system.

16 Responses to “WAR 2.0: WAR Upgraded on Baseball-Reference.com”

  1. Dr. Doom Says:

    Wow! That's a lot of change! I look forward to seeing what comes of it.

  2. admin Says:

    Hopefully all for the better. Our offensive numbers should now look a lot mroe like FanGraphs, which some what we think are improvements. Pitching is always going to look very different due to the RA vs. FIP differences.

  3. Dr. Doom Says:

    I'm also intrigued that Matt Kemp now led Ryan Braun by only .1 WAR last year... that's a HUGE difference from what the old calculation did... which I believe had them separated by 2.5 WAR.

    Also, poor Adam Darowski, who will now have to go back and recalculate his Hall of wWAR!

  4. admin Says:

    I know, changes like this can be painful. The main issue there are the defensive numbers. Switching from TZR to DRS for 2003 on is going to change a lot of post 2003 numbers and it will also affect the pitchers as well. Halladay got a big boost because DRS saw the Phillies as terrible defensively which makes some sense given their age.

  5. Sky Says:

    Love it. Love it.

    Especially like the new definition of dWAR. But a nitpick: you use the term "defense" differently in dWAR and ndWAR. The former is pos+fielding, the latter is just fielding. Maybe dWAR and nfWAR?

  6. KJOK Says:

    dWAR now contains the position component of WAR, so the Career Leaderboard is now dominated by SS, C and other great defensive players.

    This alone would have been great. Good job as always.

  7. Chris J. Says:

    Impressive change.

    WIsh I hadn't just finished filing away team dWAR, oWAR, and pWAR.

  8. abe Says:

    I find it interesting how the new "version" of WAR has Mays ranked ahead of Cobb now, and that's with no subjective league quality adjustments included, and with out giving Mays war credit.

  9. buddy Says:

    Okay, here's something I don't quite understand. Shouldn't the WAR number be the sum of ndWAR and dWAR? But, as an example, Miguel Cabrera has 43.8 ndWAR and -11.0 dWAR over his career. But his total WAR is 37.9. Huh?

  10. admin Says:

    I'm going to change ndWAR to afWAR for average fielding. Both afWAR and dWAR will include the position adjustment, so adding them will double count that.

  11. admin Says:

    Chris,

    Shoot me an e-mail and we'll work something out.

  12. kds Says:

    Wonderful! Both the technical changes and the explanations thereof.

    Thanks for the redefination of dWAR.

    For batters' DP runs you use as opportunities ground balls when runners on 1B with <2 outs. I don't think you should restrict it to times when a ground ball has been hit. Part of the ability to avoid GiDP is the skill to not hit the ball on the ground, by hitting air balls and by taking pitches. A K is better than a GiDP. (Well maybe not tie game, bottom of the 9th, no outs, runners on 1st and 3rd.) I think DP opportunities should be any runner on 1B <2 outs minus IBB and HBP.

  13. Steve Says:

    Excellent work, and much appreciated. Agree wholeheartedly that BIS's defensive metric is best. Looking forward to spending many happy hours learning surprising things.

  14. Bip Says:

    I thought fangraphs was harsh on Matt Kemp's 2010, which had him at 0.4 WAR while b-r had him at about 2. Now he's at -1.4! He doesn't look too good this year either. I guess Matt Kemp has to hit .400 with 80 homers in order to overcome his defense.

  15. John Says:

    It certainly seems like pitchers (as a whole, and especially at the elite level) are getting major boosts from these changes.

    Last year there was something of an uproar when Justin Verlander won the MVP as a pitcher last year. According to these new calculations, a pitcher should almost always win at least one league's MVP...and last year, Halladay was #1 overall in the NL, followed immediately by Cliff Lee.

    It's also going to be fun when people see that Ben Zobrist led the AL in WAR (among position players) last season. Back in 2009, people went a little crazy when he was ranked #1 in fWAR. He's an outstanding player, no doubt about it, but it seems like he gets credit for being about as good defensively as both Bobby Grich and Roberto Clemente at their respective positions.

    By how much have these changes improved WAR's correlation with actual winning?

  16. admin Says:

    Here is a post outlining the issue with pitchers and how we've addressed it. It took a lot of work, but it seems to be much better now.

    http://www.sports-reference.com/blog/2012/05/baseball-reference-com-war-explained-converting-runs-to-wins-baseball-reference-com/