PS: I imagine Matthew will come along shortly with pitcher hitting WAR for everyone not named Zambrano.
How right you are.
I modified Graham's numbers just a bit, calculating the average level from 2008 data only, which raised the wOBA up to .172.
Please do not fall into the easy trap of thinking of this in terms of replacement level. We attempted to proxy the replacement level for a pitcher's hitting ability in order to find a proper positional adjustment. After that, it's the standard hitting WAR calculations used for anyone else. To re-state something Tango has mentioned:
WAR is offensive wins above AVERAGE, defensive wins above AVERAGE, a positional adjustment, and then we apply wins above replacement (at a league level).
The end values [Wins added] presented here are the results of the pitcher's offensive wins (above average [BRAA] ) added to their positional adjustment + replacement level ( [Pos. Adj.] which comes from their average wOBA compared to the league as a whole, an implied ranking which I think does a decent enough job of estimating replacement level).*
Does that make sense? To me it does. There is a great comment made by Dave Cameron (reprinted here) that talks about positional adjustments in terms of selecting a baseball team out of, say 100, people trying out. It went along the lines of saying that the first thing you would do would be determine who could handle catching, and that might be some 10 people. Then you'd separate out those that had the skills to play shortstop and that might be 15 people, and so on and so forth. In reality though, the actual first thing you would do is separate the groups into two; those that can pitch, and those that cannot.
A pitcher's hitting has been, to my knowledge, overlooked so far and the spread in values between say Carlos Zambrano and Ben Sheets (two whole wins!) is significant enough to call this a gross negligence on our part. Of course, this wouldn't be an issue at all if the National League adopted the DH.
Point of interest: Felix Hernandez was the 15th best hitting pitcher by value last year, generating an extra 3.2 runs with his two plate appearances.
*The more clever among you might note that we are still missing a defensive value for pitchers' fielding prowess.