Last week, my colleague Michael Baumann had the audacity to suggest that Orioles closer Zach Britton could be a defensible selection as the American League’s Cy Young Award winner. If you squint, you can kind of see Michael’s logic. Britton’s 0.54 ERA is challenging Fernando Rodney’s record-low 0.60 ERA from 2012 (when runs were rarer than they are in 2016). And yes, he’s threatening to become the first pitcher to post a full-season ground-ball rate above 80 percent (which would break his own record). Moreover, he’s yet to blow a save in 37 attempts, and he’s still extending a major league record for consecutive appearances without an earned run allowed. Combine those flashy firsts with the Orioles’ surprising 66–51 record and the weak field for the AL Cy Young, and the case for the lefty looks clearer. But Britton has thrown all of 50 regular-season innings thus far, and will most likely finish with around 65 or 70. It’s crazy to think he’s anywhere near as deserving as the league’s best starters, who’ll throw three times as many frames, right?
Actually, that is right, but not in the way you were thinking. Britton is so not near those starters that he deserves not only to blow them away in the AL Cy Young voting, but also become the first reliever to win an MVP Award since Dennis Eckersley in 1992 — depending on your opinion of what an MVP is.
To make the (quite convincing!) case for Britton, you have to come down strongly on the context-sensitive side of the annual “how to define an MVP” debate. If you think the league’s most talented player (or at least the player who’s played the best) should win the MVP Award, then there’s no real argument about the appropriate pick: It’s Mike Trout. Whatever your concerns about sample size or methodology, Trout’s skill satisfies them; he’s on track to lead the American League in Baseball-Reference WAR, FanGraphs WAR, and Baseball Prospectus WARP for the fifth consecutive season. As we know, though, Trout won the award in only one of his previous four league-leading years — not coincidentally, the only one of those years in which the Angels won the AL West. And the 2016 Angels are having a tough time winning their 50th game, let alone their division.
Trout still has some space in his trophy room because voters have penalized him for being on teams that fell out of contention. It’s a variant of the “tree falls in the forest” debate: If Trout is the best player in baseball, but the Angels lose 90 games and finish in last place, then what were Trout’s wins worth, really? That’s not fair to Trout, but it’s also not really “wrong.” As long as the award goes to the most valuable player, and as long as that value isn’t strictly defined, it’s reasonable to consider the circumstances in which a player’s contributions came.
If we are going to pay any attention to when and where a player did his damage, though, then we should try to pay close attention. We can go beyond the binary “playoff team/nonplayoff team” distinction and start slicing and dicing more deeply, assessing whether a player produced his stats at the most crucial moments and in the biggest games.
As Michael mentioned, Britton leads all pitchers in Win Probability Added, a stat that sums all the credits and debits in win expectancy that a player’s actions produce over the course of a season. Unlike WAR, WPA is somewhat context-sensitive, in that it weights events differently depending on when they occurred in a game; a run-scoring single at the tail end of a blowout counts a lot less toward a player’s positive total than a run-scoring single that breaks a 10th-inning tie. However, WPA considers only some of the context we need to account for if we want to take “most valuable” to its logical conclusion. Notably, it tells us a limited amount about a player’s sway over the standings, because it doesn’t distinguish between the achievements of a player whose team has already been mathematically eliminated and the exploits of a player whose team is in the thick of the race. Nor does it place any additional emphasis on games against a division or wild-card rival, which have an outsize impact on a team’s playoff picture. Case in point: The major league leader in WPA is — yep, you guessed it — Mike Trout.
Last year, with the help of Dan Hirsch, creator of historical stats and analysis site The Baseball Gauge, I introduced/revived a stat called Championship Probability Added, which addresses the aforementioned factors that fall outside the scope of WPA. Built on the bones of similar stats that hadn’t been tweaked to reflect the current, 10-team playoff format, Championship Probability Added answers a simple but compelling question: How much has a given player helped or hurt his team’s chances of winning the World Series?
In 2015, for instance, cWPA said Anthony Rizzo was the wisest choice for NL MVP, even though Bryce Harper had the superior surface stats (and eventually won the award). Rizzo’s Cubs made the postseason by winning the second wild card, finishing third in a top-heavy NL Central, while Harper’s Nats missed the playoffs completely. In that sense, the wins Rizzo added made more of a difference. What’s more, Rizzo racked up hits at the most meaningful times of individual ballgames, piling on the production when the score was close and when runners were in scoring position. That’s the kind of nuance cWPA can capture.
To calculate cWPA, Dan determines each team’s division-title and wild-card odds on every day of the season by simulating the rest of its schedule 100,000 times (factoring in home-field advantage). He then runs additional sims to determine how much those odds would change after a win, as opposed to after a loss. Based on the size of that win-loss swing, he assigns each game a Championship Leverage Index (or CLI) score. The average game on Opening Day has a CLI of 1; any game after a team has been eliminated from the race has a CLI of 0, because neither a win nor a loss would produce any change in that team’s playoff outlook. Important games and must-win games, by contrast, have CLIs far higher than 1. By multiplying a player’s WPA in a given game by the CLI while he was playing, Dan can compute that player’s playoff footprint on a seasonal scale.
As one would expect, teams in close races tend to have high CLIs. The Orioles have been separated from their closest competitor by fewer than two games on every day since July 27, and they haven’t had three games of breathing room since July 3. This graph of each AL East team’s CLI by day shows that the three top contenders have played increasingly tense, high-stakes games as the season has unspooled, while the out-of-it Rays’ games have flatlined.
For the O’s, almost every game has been a fight for their spot in the standings. Sure enough, their average CLI leads the majors:
That brings us back to Britton. Put the player with the highest Win Probability Added on the Team with the highest Championship Leverage Index, and you have a formula for a runaway winner in the Championship Probability Added race:
According to cWPA, Britton, all by himself, has increased the Orioles’ odds of winning a championship by 4 percentage points. No other player comes close: The difference in cWPA between Britton and second-ranked Daniel Murphy is as great as the difference between Murphy and 33rd-ranked Carlos Correa. (Trout, who’s having another fantastic season on his worst roster yet, ranks 66th in the majors by cWPA, which probably presages another MVP snub.) One of Britton’s highest-WPA outings of the season, in which he pitched a scoreless ninth and 10th, coincided with the Orioles’ highest-CLI game of the season thus far: a July 31 showdown with Toronto when the teams were tied in the loss column. It’s been that type of utopian season for both Britton and Baltimore, which has also had huge assists from Manny Machado (no. 12) and Chris Tillman (no. 16).
With several weeks of the season remaining, there’s still time for Britton to run into a rare rough patch, or for the Orioles to stop spiting their preseason projections and finally fall out of the race. But barring a complete collapse down the stretch or a Britton blow-up at a season-altering time — the ninth inning of a Game 163, for instance — Britton will be the best MVP pick for people who don’t draw an equivalence between “best” and “most valuable.” Britton, who insists he isn’t interesting, isn’t better than Trout, reigning AL MVP Josh Donaldson, or even Cy Young contenders Corey Kluber or Aaron Sanchez. But because of his usage and place in the pennant race, coupled with his impressive performance, Britton probably will have had a greater positive influence on his team’s probability of winning a World Series than any other player in 2016. Trout is an unparalleled player, but even he can’t say that.
The statistics in this piece reflect action through Sunday’s games.
