Week 20 DVOA Ratings

As usual after the conference championship games, we're not going to bother with the full 32-team table of weighted DVOA ratings, since there are only two teams left and most teams haven't played for three weeks. We'll just take a quick look at both teams.

Is there a gap between these teams? It depends if you look at our total ratings for the entire year (including the postseason) or the weighted ratings that drop Weeks 1-6 and give less strength to Weeks 7-12. Here's what the offensive/defensive/special teams splits look like for both weighted and total DVOA, through the playoffs:

Just to demonstrate how much better this year's top teams were compared to last year's top teams: the 2018 Patriots and Rams both had weighted DVOA below 20% as of the Super Bowl. This year's Super Bowl teams are significantly above that.

You'll notice that both offenses have improved over the course of the season. For Kansas City, that's a combination of an improved running game and lowering the strength of the games with Matt Moore at quarterback in Weeks 7-9. For San Francisco, it's all about improvement in the running game: the 49ers had -3.5% DVOA on runs in Weeks 1-9, but they have 15.0% DVOA on runs in Weeks 10-20.

The real gap between weighted and total DVOA is on defense. The differences change depending on which weeks you use for splits, but essentially the Kansas City defense has improved a little bit over the course of the season against both the run and the pass while the San Franciso pass defense has seen substantial regression to the mean since the first few weeks of the season. (The run defense, on the other hand, improved over the second half of the year.) The response there from San Francisco fans would be that the pass defense has improved again since the return of injured players such as Dee Ford and Kwon Alexander for the playoffs, and that's true. But it remains to be seen if that's really about the return of specific defensive players or just random variation in quality of play.

Remember, defensive DVOA is better when it is negative:

Which numbers should we be looking at when we judge Super Bowl LIV: total or weighted? As I've noted in the past, weighted DVOA doesn't actually seem to be any more (or less) predictive than total DVOA, even though that was the goal when I created it. However, I do think there are a couple of reasons this year to give the weighted numbers a little more, well, weight. First, the difference in the weight of the Matt Moore games matters, because if Matt Moore is playing in Super Bowl LIV it means something very bad has happened for the Kansas City Chiefs. Second, a lot of research over the last couple years has shown that offensive numbers are more predictive than defensive numbers during a given season, in the same way that offensive numbers are more predictive than defensive numbers when we look from season to season. That research suggested that San Francisco's defense (and New England's defense) would come back to earth a bit after playing at historical levels for the first couple months of the season, and that's in fact what happened. So the lower defensive rating in weighted DVOA is probably a more "true" or at least more predictive rating when it comes to the San Francisco defense. Yes, despite the return of certain players for the postseason.

Our playoff odds simulator uses weighted DVOA and gives Kansas City a 57.4% chance of winning the Super Bowl. This is based solely on top-line numbers rather than specific matchups.

Finally, here are the one-game ratings for the conference championships:

Roster data should be updated for the conference championships later tonight. Also, voting for the 17th annual Football Outsiders reader awards should start later this week, so look for that banner on our front page and make sure to get in your votes!