Week 4 DVOA: Undefeated Vikings to No. 1

Two undefeated teams remain in the NFL after four weeks, and they are in very different places in our DVOA ratings. The Minnesota Vikings have outscored opponents by 57 points and stand on top of the table at 41.5% DVOA. This is the first time the Vikings have been at No. 1 since the end of the 2017 season. The Kansas City Chiefs have only outscored opponents by 20 points and thus rank just ninth despite their undefeated record.

Both teams get a boost from the arrival of opponent adjustments, which we install for the first time after Week 4. The Chiefs have played the third-toughest schedule so far based on the average DVOA of their four opponents. The Vikings’ schedule ranks eighth.

Other teams that have played particularly tough schedules so far include the Eagles, the Falcons, the Patriots, and the Broncos. Teams that have played particularly easy schedules include the Browns, the Titans, the Jaguars, the Bills, and the Seahawks.

The Vikings are driven by their defense, which has a DVOA that is almost double that of the second-best defense, New Orleans. The Vikings just barely sneak on to a list of the top dozen defenses ever measured by DVOA through four games:

Best 2-2 Teams Ever

Even more interesting than the Vikings at No. 1 might be the teams ranked second and third: the Baltimore Ravens and the New Orleans Saints. Both of these teams are just 2-2 but have been outstanding by DVOA. The Ravens lost to the Raiders in Week 2 despite a 96% PGWE (Post-Game Win Expectancy) while the Saints lost to the Falcons this week despite a 74% PGWE. A muffed punt recovered for a touchdown is not a very predictive play for future performance! The Saints are probably a bit overrated by DVOA because the opponent adjustment for beating Carolina is half Bryce Young games and half Andy Dalton games, but even if we discount the Saints somewhat for that, they still would come out as one of this year’s top three teams in DVOA so far. Both the Ravens and the Saints appear high on a list of the top 2-2 teams in DVOA history. None of these teams won a Super Bowl, but most of them ended up with a winning record (the 2003 Buccaneers collapsed down the stretch) and I think the strong early DVOA is a good indicator for both teams’ seasons.

None of the teams on this list won the Super Bowl, but that doesn’t mean no team has ever started at 2-2 and won the Super Bowl. It’s just that those teams were lower in DVOA, generally struggling early, such as the 1981 49ers and the 2014 Patriots.

The Seahawks are fourth after falling to Detroit on Monday night, with the Lions climbing up to fifth. The Bills plummet from first to sixth after their blowout loss to the Ravens, plus they take a hit from the addition of the opponent adjustments. San Francisco is up to seventh after a big win over New England. Pittsburgh drops to eighth. Kansas City, as I noted earlier, is ninth, and Green Bay drops a little bit to round out the top 10.

Washington Commanders: DVOA vs. EPA

One thing I wanted to address is the difference between publicly available EPA (Expected Points Added) stats and DVOA for the Washington Commanders. You may have seen notes all over the Internet that the Commanders are lapping the field in offensive EPA this season, one of the best offenses ever rated by EPA through four games. Yet in DVOA, Washington isn’t even No. 1! Baltimore is on top, with Washington as the No. 2 offense so far. I went to investigate the differences a little bit.

There are three things that seem to be going on here.

1) There seems to be a really big difference in how EPA and DVOA are judging the Washington Commanders on first downs. Both stats agree that the Commanders are better on second and third downs than on first downs. The Commanders rank No. 1 in both stats for third downs. On second downs, the Commanders are second in DVOA and first in EPA. However, there’s a huge difference on first downs, where the Commanders rank only 25th in DVOA but 11th in EPA.

What’s interesting is that both stats come out with a 45% success rate on first down, using different definitions. But the overall rank is much difference. It looks to be mostly about passing plays, as the Commanders are ninth in pass EPA but 17th in pass DVOA on first downs. It also may be connected to the difference in how EPA and DVOA treat plays based on time remaining and score in the second half of games. In the first half of games, the Commanders are 18th in EPA on first downs but 25th in DVOA. The gap is much larger in the second half of games, with the Commanders ranking ninth in EPA on first downs but 21st in DVOA.

2) Opponent adjustments are playing a role, as the Commanders’ schedule of opposing defenses ranks 27th so far this season.

3) EPA stats consider the change in EPA on every play, which means they only penalize fumbles that are lost to the defense and they incorporate the specific results of interception returns. DVOA stats, on the other hand, try to filter out the effects of luck on these plays. Fumbles are all penalized based on the type of fumble, no matter which team makes the recovery. Interceptions are penalized based on the average return of an interception considering the line of scrimmage and the length of the pass. And it turns out that the Commanders have recovered all four of their own fumbles this year, and Jayden Daniels’ only interception happened to have no return at all. So these plays will all be penalized much more in DVOA than in EPA. And it also turns out that four of these five plays were on first downs, so that helps explain that difference a little bit.

Some DAVE Changes

Speaking of the Commanders, I’ve made two changes in the projection part of the DAVE ratings this week. I’m calling a manual audible. First of all, based on the research I did yesterday showing that rookies who play this well early on generally continue to play well all season, I’ve moved the Washington offensive projection up. I’m trying to still be conservative about things, and I don’t want to suddenly project them as the best offense in the league for the rest of the year, but they’re clearly a lot better than we projected before the season.

The other change I made was for the Denver defense. In general, it is better to project teams using the entire previous season rather than assuming that improvement in the second half of the year will carry over to the next season. However, the Broncos defense was so incredibly bad in the first five weeks of last year, and so much better after that, that I think it incorrectly led to a projection that had the Broncos as the worst defense in the NFL this season. They’re pretty clearly not going to be the worst defense in the league. Again, I tried to be conservative about this change, projecting them instead as something like the No. 25 defense, but that overall will improve their defensive DAVE rating to be something more realistic. 

Week 4 DVOA Ratings

These are the FTN DVOA and DAVE ratings through four weeks of 2024. Our Defense-adjusted Value Over Average (DVOA) system breaks down every single play and compares a team’s performance to the league average based on situation in order to determine value over average. (Explained further here.) Opponent adjustments are currently at 40% of their final strength.

OFFENSE and DEFENSE DVOA are adjusted for performance indoors and consider all fumbles, kept or lost, as equal value. SPECIAL TEAMS DVOA is adjusted for type of stadium (warm, cold, dome, Denver) and week of season. As always, positive numbers represent more points so DEFENSE is better when it is NEGATIVE.

DAVE is a formula which combines our preseason forecast with current DVOA to get a more accurate projection of how a team will play the rest of the season. DAVE is currently 70% preseason forecast and 30% actual performance.