The jokes they are aplenty. "Player Valuation" actually fits the rhythm of the song, too.
The basic idea of WAR is to combine batter offensive run production and defensive run prevention and roll them into one complete number based on what a AAA player, a "replacement player", would provide. Our own Jeff Aberle broke down how WAR works for position players as well as for pitchers as part of his Purple Row Academy series.
There are a couple of different permutations of WAR that are available for general consumption:
Rally War, also known as rWAR, was developed by Sean Smith and is available at Baseball-Reference and is the most complete WAR out there - purely based on the fact that it goes back as far as play-by-play data has been kept. Using rWAR, we estimate that Babe Ruth was a 15-WAR player in 1923, when he batted .393/.545/.764, hit 41 HR, and played outstanding OF defense. As far as pitching goes, rWAR takes actual runs allowed and makes an adjustment based on the defense behind that pitcher and makes adjustments based on the defense behind that pitcher.
Fangraphs WAR, shortened to fWAR, utilizes UZR, a defensive metric that's been in use since around 2002. Using fWAR, we see that Todd Helton was an 8.6 WAR player in 2000, where he batted .329/.429/.577, hit 30 bombs and played excellent 1B defense. For pitchers, fWAR removes defense from the equation by basing their pitching WAR entirely around FIP (short for Fielding Independent Pitching). You can read more about FIP in Counting Rocks.
Baseball Prospectus has had a WAR metric for awhile now, known as WARP. I'm always hesitant to try and break down BP's metrics for some reason. Maybe it's because when I first started looking at them I had no idea what I was looking at. Whatever the reasoning, BP's been there for awhile, but somehow Fangraphs and BR took the win-valuation idea and just ran with it.
Fact is though, the big WAR machine catches a lot of flak from all sorts of channels for a variety of reasons:
1. The defensive component of WAR is sketchy.
Both WAR systems use their own defensive valuation. rWAR uses TotalZone, which is based on play-by-play data and has similarities to Bill James' Range Factor. The basis behind this kind of fielding valuation is that players should be valued for the plays they are involved in, whether for better or worse. On the flip side, UZR is what distinguishes fWAR. UZR is pretty complex, as you can see from the Fangraphs UZR Primer. The complaint around Purple Row is the fact that it seems to completely miss the mark on guys like Carlos Gonzalez and Dexter Fowler, who are both scouted as excellent defenders.
2. WAR misses the less tangible aspects of a player's offensive game.
This guy can go 1st to 3rd on a single to RF! This other guy is a high contact hitter, and he makes more productive outs than other guys! This guy here fouls off pitches and runs to first hard! I want these guys on my team! Baseball isn't measured in just home runs and strikeouts, it's who WINS the game! You might not like RBI, but they at least tell me something that WAR doesn't!
We could probably find more faults with WAR, but these ones are biggies, and I have the time to address them.
1. The defensive-metric criticism is pretty valid. I don't have much room to argue with it. Between TZ and UZR, there doesn't seem to be a consensus on any number of players. In 2010, UZR rated Carlos Gonzalez as -2.7 runs below average. TotalZone rated him as 4 runs above average. Similarly, Dexter Fowler was -2.6 runs below average in 2010 per UZR, but TZ saw him as dead average (0 runs).
Thing is, the discrepancy exists, and it's kind of hard to deny an impact to WAR. User RhodeIslandRoxFan had the idea to average the metrics to get a better idea on what a player provided in the field during a given year. It looks like Fangraphs liked the idea enough to run with it, giving us the Aggregate Defensive Ratings. These are found by making a weighted average of TotalZone, UZR, Defensive Runs Saved, and Tom Tango's Fan Scouting Report. For those wondering, TZ, UZR, DRS all had a weighting of 3 as compared to FSR's 1.
2. All those examples listed are potentially positive things. The one that is underrated and analyzed is baserunning, but some stat systems are starting to get on board with better baserunning coverage. Right now, SB and CS are represented, but past that, a lot of the value of smart baserunning (or terrible baserunning) is underrepresented in the WAR universe.
Past that, I have trouble really crediting a guy with a grounder to the right or a couple of extra fouled-off pitches if he doesn't end up on base somehow. These things can be beneficial in the microcosm of a single PA, but their value is just too dependent on the context of the PA. What if the guy grounds out to the right with a runner on 1B? Or nobody on? That groundout suddenly became a lot less valuable. Or if you are fouling off pitches with 2 outs, you still need to get on base or the inning ends.
I could go on with the contextual problems with overvaluing stuff like this, but it'll just end up in a big argument - which is probably inevitable anyhow. Valuing a player should be based on what the batter does, not what the other guys on the team do. If Dexter Fowler hits a double and Chris Iannetta fails to cross home plate, is that Fowler's fault? No, obviously not. Fowler should be credited with the double and that will go toward his batting average, OBP, SLG, OPS, wOBA, etc etc etc.
The thing is, WAR and RBI (or whatever) are trying to tell the same story, just WAR is including more and removing context. WAR is a good tool for comparison and back-of-the-envelope analysis. Anything more in-depth will require one to break out the component stats, the fielding, the batting line, the park, the splits, the everything. I can see how WAR appears to be doing too much at once and as a result doesn't do anything well, but it needs to be viewed in the proper context. WAR isn't a "how did this player fare in this game" type of number, it's a "what contribution did this player make to his team's season" number that's typically looked at during the offseason.
Sure, there are fun intangibles that make some players better than others when the statline doesn't reflect it. But these things are typically reserved for the back end of the bench where a player's impact is pretty minimal over 162 games. No, WAR may not accurately depict every nuance of those replacement players' games, but that's not really what it's there for. WAR removes context from batting numbers (eg, situation, park effects) in an effort to value players on a level plane. If Ryan Howard posts 120 RBI, is he really more valuable than David Wright, who plays for a more offensively-starved team? WAR doesn't care who you play for. It looks at who you are and what you do. No, WAR doesn't measure how you affect your teammates. That's the job of personnel scouting.
If you have more questions about WAR, feel free to ask them in the comments. Myself (and hopefully the staff and other knowledgeable posters) will be more than happy to answer them.