Readers of these fine pages know that I've been all wrapped up in a little thing I like to call wWAR (or, Weighted WAR) lately. Reader Chris Levy sent me an email asking about wWAR, so I figured I should probably step back and explain it a bit. Here's Chris' email:
I was wondering if I could ask you a couple of questions about your wWAR construct.
When WAR entered the discussion I felt as if the lights in the room had been turned on, no longer having to weight the merits of a great slugger versus a great hitter or a great base stealer. However, like you I was a bit troubled by the effect the dreaded compilers have had whenever using career stats to 'rank' someone.
I think yours is a novel approach. Which WAR data did you use? Sean Scott's from BaseballProjection/Baseball-Reference? The version from FanGraphs? Somewhere else or your own custom construct?
I notice you use Wins Above Excellence (3.0+) and Wins Above MVP (6.0+). I've always been working with the assumption that 5.0+ was AS and 8.0+ was MVP. Where did WAE and WAM come from?
I'm always trying to stay up to date with WAR developments. I've tried counting only 8.0+/5.0+ seasons myself.
First of all, thanks to Chris for writing in. I've gotten a decent amount of email about wWAR and it's very exciting. People seem to dig it. And I dig that.
So, where did it come from?It all does come from Sean Smith's WAR model that started off at BaseballProjection.com and then moved on to Baseball-Reference.com. Why use Rally's WAR (or rWAR)? Well, this was the first WAR implementation to go all the way back to 1871. Since my research spans throughout all of baseball history, this was a very compelling reason to choose rWAR. I'm also a big fan of Rally's model for pitchers when analyzing the past. I still think FanGraphs WAR is the tool to use when predicting pitcher performance. But I'm interested in the Hall of Fame—past performance. You don't build a Hall case based on how good you should have been.
What's Wins Above Excellence? As if Sean Smith hadn't completely blown my mind already, he casually introduced Wins Above Excellence in a post over at The Hardball Times.
Time to introduce a new junk stat. For this measure, I'm looking at how many wins a player has above three in a season, though his season total can never be below zero. This gives a player credit for great seasons, and ignores anything where a player is average or below, it neither adds nor hurts a player's case for greatness. A great player should not be penalized if he hangs around past his peak contributing a only little bit to his teams.
I love this idea. This takes just the player's years as a productive player and cuts out the false starts as a 19 year old or the few seasons of part time play in their 40s where they tried to hang on too long. I mean, if Ted Simmons retired five years earlier, he would have been worth 53.2 wins instead of 50.4. WAE won't penalize him for that.
How about Wins Above MVP? WAM is something I came up with after playing with WAE a little bit further. 3.0 seemed like a low baseline for "excellence", so I wanted to see what a 6.0+ version of WAE would look like. Wins Above MVP seemed like a nifty name. I wrote about it on this very site.
Wow, so the average WAR for all MVP winners thoughout the history of the National League is 7.62. The average for all MVP winners in American League history also happens to be 7.62. My quick math tells me the average of the two of those is 7.62.
As that list shows, there are plenty of MVP winners that don't reach 7.62 WAR. Still, the fact that Willie Stargell won and MVP award with a 2.3 WAR season doesn't mean I should use that as my baseline either. So what baseline should I use for finding MVP-type seasons? I hesitate to use the average because, by definition, that means 50% of the players who did win the MVP award didn't even reach that level.
So, let's go back to my original baseline of 6.0 WAR. Of the 181 players who have won MVP awards in MLB history, 133 of them achieved a WAR of 6.0 or above. That's 73%. That, to me, sounds like a pretty good baseline.
And that's how Wins Above MVP was born.
And why wWAR? As Chris references, WAR can be rather kind to "compilers". Hall of Fame voters, of course, look for peak performance in addition to longevity. The idea behind wWAR is to give extra credit to excellent seasons (WAE) and then even more credit for MVP-type seasons (WAM). The formula is ridiculously easy and seems to work well. It is simply:
wWAR = WAR + WAE + WAM
Let's look at a couple players who have similar WAR. George Sisler posted 50.4 WAR in his career and Tony Perez posted 50.5. Do they have equally good Hall of Fame cases? Heck no. Sisler's WAE was 24.6 and his WAM was 6.0. This gives Sisler a wWAR of 81.0. Meanwhile, Perez posted WAE and WAM numbers of 15.6 and 0.7, respectively. That comes out to a wWAR of 66.8. wWAR recognizes that George Sisler has a better Hall of Fame case than Tony Perez. Straight up WAR does not.
On the pitching side, take Don Sutton and Sandy Koufax. Sutton's 70.8 WAR bests Koufax's 54.5 by a substantial amount. Of course, Koufax's value was condensed into just a few seasons. As a result, his wWAR of 97.7 beats Sutton's 87.9 easily.
I really only see wWAR as useful in Hall of Fame discussions. WAR actually means something — how many wins the player was worth above replacement level. wWAR just morphs that to show players who condensed their value into peak years better. That's not particularly more valuable. It's just more Hall-worthy.
Update: (3/15/2011) I just posted the last of a series of articles where I created the "Hall of wWAR". I re-populated the Hall of Fame based purely on wWAR. Have a look!