Rubbing Mud: Overlooking Trout

You can’t get me terribly excited about any award other than the MVP anymore. I have dreams (Kevin Goldstein’s Hall of Fame-level crazy dreams) for the Rookie of the Year, but in its present form, I find it pretty boring. Manager of the Year is preposterous for reasons sufficiently documented elsewhere. The Cy Young Award has only gotten more interesting over the past few years, now that it tends not be about squabbles over win-loss records, but (maybe because of whatever traumatic experience with pitching in my childhood prevents me from appreciating it as much as so many of you do) it leaves me flat.

I still get up for the MVP, though, which is why it’s so strange to me to see no one having much of a conversation about the reveal of the MVPs on Thursday night. In particular, what strikes me is that there’s a hesitant consensus that Josh Donaldson is going to win the AL MVP, and there’s very little resistance or objection to that floating around. The throwaway line copied almost word-for-word from so many columnists’ blog posts and bloggers’ columns is: “There’s no wrong answer, but Donaldson will probably win.”

Well, I think there is a wrong answer, actually, and Donaldson is it. Mike Trout has been the rightful (and fairly clear) MVP of the American League in all four of his full big-league seasons, but he could find out, on Thursday night, that he must settle for one MVP trophy and three runner-up finishes. It’s getting a little ridiculous, and it’s thoroughly bizarre that Donaldson has so much more sabermetric support than Miguel Cabrera ever had. If anything, Cabrera was (especially in 2013) more deserving than Donaldson.

Of course, that’s not the story you’ll read if you pull up the WAR leaderboards on Baseball Reference or FanGraphs. They show Donaldson and Trout more or less neck-and-beefy neck, with less than a win separating them in each formulation of WAR. By Prospectus’ WARP, though, Trout leaves Donaldson in the dust. In fact, WARP pegs Trout for 10.0 wins, and Donaldson for only 7.6. Instead of seeing them as a two-man tier set apart from the league, WARP views Trout as a man alone on a pedestal and Donaldson as just a rounding error better than Manny Machado.

That’s not the only place where WARP zigs and the WARs zag, so instead of simply claiming the superiority of one system over the others, I decided to line up the top 50 players under each value metric and look for interesting cases. Trout and Donaldson are just one cross-section of this.

Top 50 Players by WAR(P), 2015

There’s nothing terribly complicated about the discrepancy between the three models, where Trout and Donaldson are concerned. Here’s how the three systems valued their various skills.

Mike Trout and Josh Donaldson, 2015 Batting Runs

2015 Fielding Runs

2015 Baserunning Runs

There’s definitely a gap between the three metrics in the way they value these two players’ bats. Trout’s .353 True Average easily outstrips Donaldson’s .324. The gap there isn’t proportionate to the gap between them in the other offensive metrics. Trout played his games in tougher parks (his Batter Park Factor was 98; Donaldson’s was 103), and faced tougher opposing pitchers (his opponent RPA+ was 99; Donaldson’s was 101). Still, the jump-off-the-page difference is in the three evaluations of each player’s defense. Because FRAA operates somewhat differently than do UZR and DRS, it frequently diverges from them significantly. In this case, FRAA gives Trout a seven-run edge on Donaldson with the leather, whereas DRS sees Donaldson as six runs better, and UZR has him ahead by nine runs. That’s hard to reconcile, and it’s especially hard to say which model we should most trust.

I will say, though, that I trust TAv’s respective evaluations of the two players’ bats much more than the others sites’ metrics. More broadly, where WARP likes a player significantly better or worse than the other WAR models, it’s often because of some adjustment (like opponent quality and park factors) that Prospectus handles better than the others. For instance, and keep an eye on this when the NL MVP balloting comes out Thursday night: Buster Posey. Since Prospectus has folded catcher framing and blocking into WARP, there are some guys whose full value is now captured there, but not elsewhere. Posey is roughly a six-win player, per Baseball Reference and FanGraphs, and a fringe top-10 MVP candidate. WARP has him at 8.3 wins, though, and places only Bryce Harper and Paul Goldschmidt above him in the NL. Posey will show up on some ballots, but probably won’t come close to claiming third in the balloting (in fact, we know that he won’t, since he’s not even among the finalists for the award).

Obviously, there are other differences worthy of your attention. Notice how FanGraphs’ choice to zero in on the outcomes pitchers can most control changes the makeup of their top-50 WAR list. Look at the number of outliers who leap onto the bWAR list on the strength of defense. (FRAA credited five players with at least 18 runs saved defensively this year. UZR only had three players at that level. DRS had 15.)

There are better and more scientific ways to capture those phenomena, though. Just remember, on Thursday night, that Trout is the rightful AL MVP, even if only one flavor of WAR can see it.