Toward an MLB MVP-Voting Rubric

[The following is an introduction to a more thorough study I intend to publish at Banished to the Pen following the conclusion of the current MLB season the purpose of which is to suggest an approach to harmonizing traditional ways of thinking about the sport’s annual MVP award with available sabermetric principles. -ed.]

Last week’s Zach Britton trade reminded me about the idea of championship win probability added (cWPA). In 2016, Ben Lindbergh made the case that Britton, then serving as the closer for the Baltimore Orioles, should win the award for the most valuable player in the American League based on the idea that Britton, at least during the regular season, had done more to help his team win the World Series than any other player had helped his own team win the championship. Britton– again, a relief pitcher who threw sixty-seven innings– finished eleventh in MVP voting. Mike Trout, the WAR leader, claimed the award, his second, and I don’t suspect many people have thought much about cWPA since then.

What should make cWPA (defined: cWPA “takes individual game win probability added (WPA) and increases the scope from winning a game to winning the World Series. Where a player’s WPA is the number of percentage points that player increased or decreased their team’s probability of winning a single game, their cWPA is the number of percentage points the player increased or decreased their team’s chances of winning the World Series.”) attractive to a broad swath of the MVP electorate is that it accounts for the traditional notion that the individual award-winner ought to have been on a winning team. When handing out performance awards for a given season, it makes sense to reward players based on what they actually accomplished, as opposed to what they should have accomplished but for bad luck, sequencing, weak teammates, strong opponents, environmental variations, and other contextual and extrinsic factors. After all, these factors work, to some extent, on all players, and just as we determine team season standings based on actual win percentage (and not a sabermetrically adjusted winning percentage), so too should we determine individual season awards based on actual results.

This is not to say that the MVP should go to the player with the best batting average or the Cy Young to the hurler with the most pitcher wins, obviously. While using cWPA as a guide will steer us toward players who are, in the conventional sense, winners, it uses accepted sabermetric principles to maneuver in that direction. Still, I think there may be some concern that cWPA, when used alone as a player-valuation measure, might be too context-dependent and inclined to reward disproportionately an otherwise unremarkable player who happened to find himself in one or two of the right places at one or two of the right times. Certainly, though, it seems likely that a player who accumulated performances in those situations often enough to find himself near the top of the cWPA leaderboard also would be a well-rounded and highly productive player in all situations.

Still, I was curious how well cWPA correlated with WAR, something it ought to do reasonably well if it’s to be available as a useful informant for MVP voters. I pulled the cWPA numbers for the current season and used FanGraphs’ version of WAR mainly because their combined WAR leaderboard made data manipulation easier.

(click image to enlarge)

I don’t have a statistics background, so I won’t comment on the significance of the correlation between cWPA and fWAR except to say that it seems sufficiently strong. If you’re curious about who’s who on this plot, here are the current top-ten players by cWPA:

There is a lot of significant, championship-relevant baseball yet to be played in 2018, and a more significant study of the above would have involved prior, completed seasons, but I think there’s something here and wanted to share what I had compiled on the current season as it moves into August and teams begin their playoff charges in earnest. I anticipate updating this information after the conclusion of the regular season and supplementing it with historical data to create an even more robust analysis. In the meantime, I welcome any input on win probability added and seasonal awards.

Advertisements

A narrowly focused update on Zach Britton, new New York Yankee

blern

Last night, the New York Yankees completed what to this point constitutes the second-most significant trade of the month when they sent three prospects to Baltimore in exchange for a few months of closer Zach Britton’s services.

The Orioles drafted Britton out of high school in 2006, and Britton debuted five years later as a full-time starter for Baltimore in 2011. By 2014, he had transitioned to a full-time bullpen role, and my earliest memories of him date to two years after that.

Britton was a key part of the 2016 Baltimore team that finished second in the AL East and made the postseason as a wild card. That was the Orioles’ last playoff appearance, and manager Buck Showalter’s decision not to use Britton as the win-or-go-home contest went into extra innings granted the game an air of infamy.

Prior to that, writer Ben Lindbergh memorably made the case that Britton, a closer who would pitch sixty-seven innings that season, merited serious consideration for the AL MVP award. As recorded contemporaneously in these digital pages, Lindbergh’s argument was based on a modification to the notion of Win Probability Added (WPA):

Earlier today, Ben Lindbergh argued that Baltimore reliever Zach Britton has a claim to the 2016 AL MVP award. To make that case, Lindbergh demonstrated that Britton had done more than any other player to help his team win games that mattered. Lindbergh did this by placing Britton’s performance in the context of the individual games in which Britton pitched– did Britton’s actions help or hurt his team’s chances of winning that game, and to what degree did they do so?– and then placing those games in the context of his team’s position in the playoff hunt. Viewed this way, Britton (excellent contributions to a good team in close contention) is more valuable than, for example, Mike Trout (superlative contributions to a bad team far out of contention). The metric that captures this contextual performance concept is called Championship Probability Added (cWPA), and Britton currently holds a commanding lead atop that leaderboard.

(Emphasis added.)

The road has been a bit rough for Britton since that 2016 season, however, as the trade article linked above summarized:

After consecutive two-win seasons in 2015 and 2016, he has missed time with the following injuries:

  • April 16, 2017 – Hits the disabled list with a strained left forearm and misses a little over two weeks.
  • May 6, 2017 – Almost immediately after return from disabled list, goes back on it with same injury.
  • August 25, 2017 – Injures his left knee and is shut down in September.
  • December 2017 – Hurts his right Achilles in an offseason workout requiring surgery.

The lefty returned to action on June 12 but hasn’t been lights out like he was before 2017, with a 4.43 FIP and 3.45 ERA thus far. He’s been a bit better of late, tossing eight straight scoreless outings, but has still produced just six strikeouts against four walks in that span. Perhaps more encouragingly, his velocity is up over his last few outings, getting closer to the 97 mph sinker he used to throw. If the velocity return is here to stay, better results might follow.

By a clear margin, Britton led all pitchers in WPA in 2016. This year, however, he’s nowhere to be found atop that list. That no team has done as little winning as Baltimore (record: 29-73) likely contributed to that shift. Still, the fact that he has a negative WPA (-0.13) for the first time since he moved into the bullpen seems worth noting in light of the foregoing.

As the block quotation immediately preceding the immediately preceding paragraph indicates, there are a number of red flags that suggest that the version of Britton the Yankees acquired (insert reference about Redcoat POWs) last night may be meaningfully different from the one who presented an intriguingly compelling case for consideration as the most valuable player in the American League in 2016.

As a closing addendum, the current leaders in pitcher cWPA for 2018 are Justin Verlander (.023) and Josh Hader (.020).

________________________________________

Related
Baltimore Closer Zach Britton Isn’t Just a Surprise Cy Young Contender — He’s the AL MVP