The AL MVP race just got more interesting, to me at least.
What the BP guys did was look at every PA by every player, and see how much the team's probability of winning the game changed as a result of that PA's outcome. This is the opposite extreme from Linear Weights or VORP or MLV -- it's the fully-clutch-loaded measure of how much this particular player changed his team's immediate chances of winning games, PA by PA. For those of you who think that late-and-close PAs "matter more" than other PA's, this stat's for you.
Cumulative Delta Win Probability, all major league players:
Looks like we have a race after all, due to Ortiz's massively clutch season. The context-sensitive offensive difference between Ortiz and A-Rod is 2.55 wins. Quality of pitchers faced is almost identical; A-Rod gains a few points back for baserunning, but that's in the noise. Defensively, A-Rod has been about 11 runs above replacement (9 below average) according to Clay Davenport's metrics, so that's certainly not going to make up the difference.
What we're left with is a very difficult question: namely, is the opportunity cost of having to put Ortiz at DH (and thus play Manny Ramirez in left, among other things) high enough to offset the win expectation difference between Ortiz and A-Rod this year? Manny's roughly a replacement-level defensive player, 1 to 1.5 wins below average in left. I don't think that quite gets you there.
So, if you believe that the MVP is the player who, in part because of the luck of who his teammates are and the situations he has come to the plate in, has done the most to increase his team's chances of winning games, then I think you have to vote for Ortiz in the AL this year.
In the NL, Derrek Lee drops from "no brainer" to "OK, let's use the brain after all". Here, though, I think Lee's defensive and baserunning advantage over the other contenders is enough to keep him in the lead. Jones is a poor fielder, and Delgado and Clark have been even worse this year. I'd still vote for Lee, if I were going by win prob.
(Of course, we don't have the defensive data to do the same thing with the leather as with the bat. Who knows whether those extra defensive outs were in clutch situations or in blowouts?)
FWIW, Derek Jeter shows as a very un-clutch hitter this year. No kewpie doll for him.