More on WAR

WAR is really designed to look at individuals. It is an attempt to evaluate a player holistically by including offense, defense, and baserunning. For pitchers it is all about run prevention. We'll look at pitcher-WAR later. Here's a chart (apropos our discussion in the comments) showing the distribution of WAR among position players:

It is from 2010--a really good year! The chart is from a Justin Bopp article on Beyond the Boxscore. Elite players are just that: elite, meaning the choice or the best.

Here's another visual (from the excellent FanGraphs primer on WAR):

I will take a look at specific players and individual seasons later. WAR includes a positional adjustment as well as things like park factors. There are a lot of components in the model, and a lot of tweaking goes on as well as yearly calibrations. WAR has generated much healthy debate about how to value a player's contribution to his team's wins. I don't get too granular with WAR. A 5-WAR player may really be a 4-6 WAR player (or 4.5-5.5) but I'm not debating MVP or HoF votes so I don't care. A 4-6 WAR player is a damn good guy to have on your team!

As with any model, one should not think of WAR as true or false or right or wrong. A good or bad model perhaps, or an in-between one, something that exists on a continuum between "more effective" and "less effective." We should expect a model to have varying performance, and with WAR the individual pieces of the model can be re-assessed and thus the model can be improved.

I've learned two things from WAR. First, it helped me appreciate players that did not appear to be positive contributors because the traditional stats weren't revealing enough. Second, it reinforced how hard it is to be an MLB player, and how good you have to be just to be considered "average."

--M.C.