Major League Baseball will announce the winner of the 2011 American League Most Valuable Player this afternoon. While sabermetric tools such as Wins Above Replacement are very helpful, and perhaps even necessary for sorting out which players have been the most valuable, they are not necessarily sufficient by themselves for deciding such issues. As I discussed in an earlier (no-longer-so-“official”) post on using WAR to help determine the MVP, WAR and its cousins should start conversations about the MVP, not end them. However, this post is less about the general framework and more about why I think, despite the presence of other viable candidates just as Jacoby Ellsbury and maybe Justin Verlander, Jose Bautista is my choice for the American League’s Most Valuable Player of 2011.
As I have discussed before, if you take into account all the uncertainties and other considerations, one can rearrange a ranking of MVP candidates over and over again. I think there are credible cases to be made for players like Curtis Granderson and Miguel Cabrera. However, for the sake of space, I will discuss Bautista’s credentials versus those of two other popular candidates: Verlander and Ellsbury. The WAR leader boards for batters and pitchers make it seem simple enough, but again, they are merely the jumping-off point.
Justin Verlander is a popular candidate with some because of his wins and the Detroit making the playoffs, but he also had a legitimately outstanding season at 7.0 WAR. He already won the Cy Young, although it is worth noting that at least according to FanGraphs WAR, CC Sabathia was equally as good and also pitched against stronger competition. I have no problem with a pitcher winning the MVP, and I think Zack Greinke should have won it in 2009. However, according to WAR, Verlander is really is not in the same league as Bautista (8.3) and Ellsbury (9.4). Verlander’s 7.0 makes him a good down ballot vote-getter on par with, say, Alex Gordon, but not really the MVP. I will discuss two ways people make the case for Verlander: his contribution should count for more since his team made the playoffs, and that FanGraphs WAR underestimates his value because it uses FIP rather than RA.
As for the “playoff issue,” I see MVP as an individual award, not a team award. Going to the playoffs or being a contender involves the whole team. As I said in my earlier post, I define the most valuable player something like this: “if you could guarantee the performance the season ahead of time, which player’s 2011 would you pay the most for without knowing ahead of time what the rest of the team is like?” However, even if you accept that being on a contending team should make a difference in the MVP choice, the Red Sox missed the playoffs on the very last day of the season, and Ellsbury can hardly be blamed for their failure. Combined with his objective performance, that would make Elssbury a better choice than Verlander. One could even turn this in Bautista’s favor. Detroit won their division by 15 games. Toronto did not make the playoffs, but they only missed the wildcard by 10 games. In other words, Toronto was closer to making the playoffs than Detroit was to missing them, so Bautista’s contributions were more “crucial.” I am not saying that is how I would make the argument for the MVP, but the idea that Verlander should be elevated because of his contributions to a contending team itself is not a clear point in his favor.
More relevant is the possibility that FanGraphs WAR for pitchers undervalues Verlander’s 2011 performance because it uses FIP rather than ERA (or RA, I will use them interchangeably). Verlander’s ERA (2.40) was considerably lower than his FIP (2.99). While I do think that FIP is better than ERA (or RA) for measuring the contribution of pitchers in general, it is also true that neither method is on the same level of reliability as, say, using linear weights for hitters. Moreover, it is also the case that some pitchers have the ability to “outpitch” their FIP. I do not think it is the case with Verlander. Phil Birnbaum’s notion of a Bayesian Cy Young (itself inspired by Eric and Dave’s discussion of the 2011 NL Cy Young) is helpful here. We can look to past seasons to get an idea if Verlander’s low BABIP, the main driver of the gap between his FIP and ERA, is likely something he controlled or not. A quick look shows that a .236 BABIP in 2011 was by far the lowest Verlander’s of career, about fifty points below his career average. From the broader perspective of FIP, he does not seem to have the ability to outpitch it. Even including 2011, his career ERA is 3.54, and his career FIP is 3.49. This makes it unlikely that his 2011 BABIP and ERA “overperformance” are the result of something he did, thus it is unlikely that Verlander’s contributions to the Tigers are being undervalued by FanGraphs WAR.
Ellsbury is the stronger competitor for Bautista. When I wrote about Bautista earlier, he and Ellsbury were pretty close. Since then, with updates to the various metrics, Ellsbury is now up 9.4 WAR to 8.3 WAR on the leaderboard. So how can I still think Bautista should be the MVP? He was the most valuable hitter in baseball in 2011, but the difference between Bautista and Miguel Cabrera on that score so small as to is meaningless. Of course, Bautista played a more difficult position than Cabrera (although according to UZR, neither played their position well) and was a better base runner. Bautista was more valuable that Cabrera for those reasons, but, of course, for similar reasons WAR has Ellsbury ahead of Bautista — while Bautista’s base running (as measured by UBR) was (surprisingly) slightly better than Ellsbury’s, center field is a more difficult position than right field.
More significantly in this case, Ultimate Zone Rating had Ellsbury as a very good defender in center, and Bautista as below average overall (Bautista also spend a bit of time at third base). The difference is about twenty runs. Defensive statistics are controversial, but I will not enter that controversy at length here. The gap between Bautista and Ellsbury’s batting runs is much more reliable than that between their defensive statistics. Moreover, for similar “Bayesian” reasons as that given for pitcher BABIP above, if we look just at Ellsbury’s UZR for previous seasons in center field, it seems a less likely he was really a +16 center fielder in 2011. I am not going to get into precise numbers, but given past performance by Ellsbury in center field and the relative uncertainty of defensive metrics compared to offensive metrics (which are not perfect, either, by the way), the approximately one win difference between Ellsbury and Bautista seems much less certain given that Ellsbury benefits from a twenty-run advantage in fielding.
But even if we cut that difference in fielding in half due to both Bayesian and general considerations on reliability, Ellsbury and Bautista are still roughly equal overall at around eight wins. I would not blame anyone for voting for one of the top players in the league for MVP, least of all taking Ellsbury over Bautista. However, I still think Bautista is the best choice when we look at his offense beyond traditional, park-adjusted linear weights. While batting runs (park-adjusted wRAA) are a good general stat for measuring the runs the player would have produced on a average team in average situations, we can also measure who players respond to game situations using WPA/LI. This is similar to “clutch,” but is not quite the same, as it is does not include the leverage of the situation, which the batter has no control over. I have discussed the difference between WPA/LI and traditional linear weights and why it might be something we should give a credit to the hitter for elsewhere. While Bautista’s 6.67 WPA/LI is right in line with his batting runs (6.65 once converted to wins), Ellsbury’s 3.96 WPA/LI is actually about three-quarters of a win worse than his 4.73 batting wins.
I admit that this is far from cut-and-dried. If one simply substitutes WPA/LI for batting runs in WAR, Ellsbury still comes out about half-a-win ahead. However, if one combines this sort of analysis with giving defensive metrics less weight, I think that Bautista comes out ahead. It is close, and I certainly could see how the reasons why someone would still vote for Ellsbury for MVP, but my choice is Jose Bautista. My only regret is that I spent most of the post talking about why Verlander and Ellsbury were not as good (in my estimation) as Bautista in 2011, rather than celebrating Bautista’s awesome season. I will leave that for you, if you need anything beyond simply reciting his numbers: .302/.447/.608, 54 home runs, 181 wRC+.
NB: While I live in the Greater Toronto Area, I am not a Blue Jays fan. The Royals are my team. So please, base your angry accusations of bias on something else.
Matt Klaassen reads and writes obituaries in the Greater Toronto Area. If you can't get enough of him, follow him on Twitter.