...absolutely everything, unless you don't understand how it works.
The more I have read up on WAR for pitchers today I have made some very interesting discoveries. WAR is a very tricky stat to use for relief pitchers because while it does in fact factor in innings pitched, one has to consider that so does WAR for starting pitchers or at bats for position players. For example the more innings pitched or at bats you have the more opportunity you have to add value. For a relief pitcher your stats are constrained to how few innings you pitch, and thus your WAR is relative.
Take for example last year. Carlos Marmol led all relievers with 3.1 WAR. Cliff Lee led all starters last season with 7.1 WAR. Yet Cliff Lee logged 212 innings pitched versus Marmol's 77 innings pitched. If you were to extrapolate out Marmol's number of innings pitched to that of Lee's, his WAR would come in somewhere around 9.
Now here is the tricky part. Relievers have an inherent advantage over starters in posting Value because they face far fewer batters in a game and thus have the ability to utilize their arm in a way a starter simply can't. Starters are inherently more valuable than relief pitchers because while they may provide lower quality innings, they provide them in a far higher quantity.
So when looking at the value of relief pitchers versus starter, while we can use WAR as a very useful stat to compare the two, or even the value between a starter and a relief pitcher, one must consider what the value of a player is relevant to his position. Where 0 WAR is replacement level for position players, starters and relievers, the league average for both position players and starters is 2 wins above replacement, whereas the league average for relievers is a paltry 0.3 wins above replacement. When assessing the value of a player you must consider it relative.
"it ain't you all, its yall"