WAR has its other flaws to me. As Voros McCracken has pointed out, the DIPS numbers and BABiP does not apply to knuckleballers the same as anyone else, and there are decades of data supporting that. Dickey was largely sold should by his FIP, and even FanGraphs posted that WAR should be ignored for Dickey, projecting that he was likely worth two more wins than his WAR indicated.

My bottom line with the AL MVP was this. If the margin is negligible, then yes, you must look at the full overview of accomplishments, although you should do this regardless. With the NL MVP, however, there were a handful or more of candidates within the margin of error. You cannot use WAR to say that Braun, Posey, Wright, Molina, McCutchen or Headley are better/worse than their peers when they’re separated by a win or less. When the margin is three wins, however, you can look at the full picture and see Trout’s full value was just flat out better than Cabrera. Look at it this way. On the days when both look human and finish 0-4, who would you rather have on your team? I don’t think its tough decision stretch to say Trout at his worst still provides elite defense, whereas Cabrera becomes completely worthless that day.

WAR was designed as a summary stat, not an end-all-discussions stat. It’s a handy tool but is not all encompassing, and is not meant to be. There is no flawless stat.

WAR is designed to be the end-all stat. If it’s not, let’s adjust it rather than dismissing it.

I just came here to make a snarky political comment and ran into an interesting (sort of) baseball discussion.

The thing to keep in mind is that WAR (wins above replacement) is not the same as raw value as WAR is scaled for each position. So….if Howard had 3.0 WAR and Schumaker had 3.0 WAR that just means that both were about 3 wins better than a replacment 1B or 2B, respectively….NOT that Howard was less productive than Schumaker in a raw performance perspective.

Doesn’t matter that there is no one recognized way of calculating the stat

Because WAR isn’t a stat any more than hitting is. fWAR is a stat, rWAR is a stat, WARP is a stat and they each are calculated the same way, the sum of their components /10. Now each uses different proprietary components involved, but that’s not an issue.

(r/f)WAR(P) is actually incredibly easy to calculate. Calculating the components is harder, but no one ever argues that. In fact, you can create your own version of WAR if you want. Create a metric to determine offensive performance and apply it to everyone objectively. Then do the same for defense, and position, and playing time, and base running. Add them all together and voila, you have WAR…

Again, neither analogy is appropriate. Bloodletting did work, but not well? Well the collection of traditional stats has worked, and worked well, which is why they are still used today, after over 100 years of baseball and countless advanced statistics and sabermetrics invented. Do you realize how long WAR has actually been around? It isn’t new, and there is a reason that it hasn’t replaced other traditional stats, because the collection of those traditional stats paints a better picture of overall importance than a single number of WAR could ever do. All stats are flawed, but the more complex the formulas to create those stats become, the more flawed those stats become.

As far as using a “collection,” my point was that WAR is not an all-encompassing stat, and should not be the basis of evaluation, but instead the collection of ALL stats should be looked at. I think when you try to combine and formulate the simple stats into more complex ones, you lose a lot in translation. To me, it’s impossible to quantify something as complex as a baseball players performance into a single number, in this case, the WAR statistic.

Second, WAR shows that Jeter is a far better SS than Brendan Ryan.

Feel free to search online for some 2012 AL MVP discussions and good luck finding any saber-oriented writers that used WAR as the entire or primary basis for their opinion.

I’m not saying it’s flawless, by definition it includes flaws. There is bais built in depending on who’s WAR you’re using and how THEY value certain things. That’s also the reason it’s not a collection of traditional stats. They’re adjusted (weighted) by opinion.

WAR is a frothy mix of stats and opinion.