Handicapping the Award Races: MVP
Once baseball’s non-waiver trade deadline passes, you start to see the conversations shift from fantasy to terra firma. Almost all the big-name players who are likely to make Chicxulub-sized impacts on team rosters have already been traded. The focus shifts squarely back to the pennant races, and with them, talk of individual player awards.
It will come as no surprise to most readers that I love working on predictive models. It’s not just about trying to predict the future — though that is inevitably a large part of it — it’s also about dissecting things to see how they work. Awards are something I’ve always found fascinating because they not only deal with truths in baseball but also with the psychology and mindset of the people covering the sport. We talk a lot about baseball writers believing more in stats like OBP and SLG, and eventually WAR, but the proof in the pudding is in the eating. If advanced stats don’t budge how writers are judging the best players in the league, are they truly accepted?
We’ll start this trio of pieces with the current MVP races. I’ve spent a lot of time over the past decade modeling MVP votes, and the truth is that things have, in fact, shifted considerably. Slugging percentage and wins above replacement do have more predictive value than they did in the past, as does on-base percentage (albeit to a lesser extent). The defensive players who get a larger share of the vote than one expects tend to be players who do well in the sabermetric defensive measures. Team quality and the Triple Crown stats still play the largest role, however, and even though the MVP award doesn’t specify hitters over pitchers, pitchers still make far less of a dent than one would expect from their impact.