I think that because personnel interplay and interdependencies are more tightly coupled in football than in baseball, it's fair to weight team success more heavily in deciding which player was the most valuable. We can separate out the individual achievements more readily in baseball and have been able to move past the argument of "sure Trout had the best year, but the Angels finished 3rd, how valuable could he have possibly been?", as if players can (or ought to) do everything for the team.
But in football, we're still sort of in the statistical stone age in evaluating how players add value on every play, in both rate and counting stats. If we grant that we don't need deep stats to know that QBs are almost always going to be the most valuable players by far (at least in today's game), then the MVP question becomes much simpler: what QB added the most value relative to expectations? To me, that's a counting stat, and "Expected yards" and/or DYAR are about the single best numbers we can attach, but we also need to account for scheme, health, clutch performance (e.g. 4QC / GWD), etc. One stat I wish we had is the fraction of a team's total WPA added by the defense, offense, and special teams respectively. That might mirror the DVOA stats closely, but unlike DVOA, it would account for the clock and game situation more heavily.
As a start in that direction, Ben Morris's articles on
Gunslinger of the Week have shown us that sometimes taking unreasonable risks (= higher INT%) to go for a higher chance of winning the game (and correspondingly higher chance of losing spectacularly, as opposed to closely) can be justified. So, I don't think INT% alone - and the resultant outsize influence on Passer Rating - tells the most important part of the story. There's a narrative element to QB decisions at crunch time, making snap decisions (ha!) that dramatically shift win probability and manage the clock and team focus/morale at the most important times. You can easily get into "calm eyes" territory there, but I think it's fair to look at all the close games that Newton played this year and the way he led decisive drives (@SEA, @NO, @NYG), or salted away the win to prevent a comeback (HOU, @JAX), and conclude that he stood out above Palmer or Wilson. I would disagree with that conclusion, but the point is, it's not like rate stats tell the whole story.
On the third hand, if we focus on crunch-time analysis, Carolina's defense plainly and obviously won the games vs NO, @TB, IND (sudden-death INT!), GB, and arguably the PHI and WAS games too. And some games were curb-stompings from the get-go, of course. But barring a solo defensive standout like JJ Watt or Lawrence Taylor '86, you can't really argue that Newton wasn't his team's most valuable player, even if he wasn't the best player
relative to his position's average on his roster (I'd still vote Kuechly).
