Leave Kirk Alone!

Midway through the season — with Russell Wilson dragging a flawed team into contention for the No. 1 seed in the NFC, Dak Prescott leading the league’s best offenses and Lamar Jackson making his doubters look silly — this was shaping up to be one of the most exciting MVP races in recent memory. But a couple of months later, there’s really only one option. We’ll get to that.

Last year was a two-person race with Patrick Mahomes emerging as a clear winner over Drew Brees in the end. But this year, there are a lot of players having great seasons. Let’s narrow down the field before casting a top-five ballot.

The numbers

First of all, we’re only considering quarterbacks for obvious reasons, and since this is “Inside the numbers,” let’s look at some per-play efficiency numbers so we have a concrete starting point. This isn’t to say efficiency is all that matters — as we’ll discuss later, these numbers are certainly influenced by coaching and teammates — but if we’re going to put someone in the MVP race, he should be the leader of an efficient offense. In addition, numbers are invariably going to be used when weighing MVP candidates against each other. So if we do use numbers, let’s make sure to use ones that isolate the performance of a quarterback to some extent, and not use bad ones.

For our pool of candidates, let’s start with the six Pro Bowl quarterbacks and swap out Drew Brees for Kirk Cousins, as Brees missed nearly six full games. Here are the six players, ordered alphabetically, with ranks among players with at least 500 plays.



This is all somewhat subjective, but these are the numbers I’ve been keeping my eye on throughout the season. Here’s a description of each of the numbers shown above (feel free to skip to the next section if you don’t find this part interesting):

The efficiency measures: Each of these measures the success of the plays involving the quarterback in some way.

*Expected Points Added (EPA) per play: the raw efficiency of plays a QB was involved in. Note that I cap negative-EPA plays at the average value of a turnover (-4.5) because whether plays are more negative has little to do with the quarterback (for example, following an interception, whether the defender is tackled or runs it back for a touchdown). See also the interception returns section in the QBR explainer.

*Completion Percentage Over Expectation (CPOE): depth-adjusted completion percentage, a measure developed by FiveThirtyEight’s Josh Hermsmeyer.

*EPA + CPOE index: a predictive composite of EPA/play (how effective plays were) and CPOE (how well a QB completes passes given where he throws them) designed to best get at the underlying play of a quarterback, which I first introduced here.

*PFF Offense grade: ProFootballFocus’ charting-based grades.

The team-support measures: Each of these attempts to measure how much support a quarterback has received from his teammates.

*Pass Block Win Rate: ESPN’s measure of pass blocking, obtained from Next Gen Stats. Note that I’m using numbers through Week 16 since some of the QBs under consideration sat out in Week 17 (and season-long numbers won’t change much with the update anyway).

*Defense EPA/play: It’s harder to win games with a bad defense.

*Rush EPA/play: It’s harder to win games with a bad rushing offense.

Note that a weakness of the team support measures is that we don’t yet have a good way to quantify wide receiver play, which means we know which quarterbacks overcame especially poor pass protection (most notably Russell Wilson), but not which players elevated his receivers.

5. Kirk Cousins

Wait, really? Kirk Cousins?

Let’s tackle some of the arguments against him. First, he hasn’t been as involved in the offense as some other quarterbacks, with the Vikings running the ball at a high rate on early downs. Cousins was involved in fewer plays than Mahomes despite the latter missing nearly three full games (this was true even before Cousins sat out Week 17). But I don’t think holding Cousins’ low volume against him is reasonable. With a conservative defense-minded head coach, his team’s run-pass preferences aren’t his fault, and he’s been available to drop back whenever his team wants. Could he have added more value if his team had dialed up more pass plays? Sure. But that doesn’t change how he performed on the plays he executed when his number was called.

Second, Cousins benefits from a well-designed offense that plays to his strengths. As outlined by Zone Coverage’s Nick Olson here, Cousins has made hay on the Vikings’ play-action shot plays: The Vikings were No. 4 in the percentage of dropbacks from play action, per Football Outsiders, and their 8.9 yards per play on play-action dropbacks ranked No. 6 in the league.

But Cousins playing in a well-designed offense shouldn’t be held against him, either. The job of NFL coaches is to put their players in the best positions to succeed, and Cousins has succeeded, with a few very notable exceptions.

I’m also skeptical about the degree to which Cousins is in as great a situation as is often portrayed. The Vikings ranked No. 18 in Pass Block Win Rate and their offensive line was demolished by Za’Darius Smith and the Packers in Week 16. Stefon Diggs is really good, but a lot of teams have at least one good receiver. Adam Thielen has been injured for most of the season and didn’t exceed four targets after early October. Kyle Rudolph spent the offseason being dogged by trade rumors. Dalvin Cook is a running back. But Cousins has been one of the most efficient quarterbacks this year while rating favorably by PFF as well. He’s had a good season.

Cousins gets docked for two dismal performances against the Packers that removed any shot for his team to win the division, but I can’t think of anyone with a better case for this No. 5 spot.