DVOA Week 3: Schedule Will Get Harder for Steelers, Cardinals, and Giants
Powered by the No. 1 defense through the early part of the 2021 season, the Carolina Panthers sit atop a tightly packed set of DVOA ratings for the second straight week. Less than three percentage points separate the top four teams: Carolina, Cleveland, the Los Angeles Rams, and Arizona. Denver ranks fifth, a couple more percentage points behind the top four.
But if you look at it a different way, the Arizona Cardinals have moved to No. 1 in the DVOA ratings after a 31-19 win over the Jacksonville Jaguars. Denver is close behind in second place. The rest of the top five goes Cleveland, Los Angeles, and then Carolina.
The difference between the two lists? Opponent adjustments, and that relates to a change I decided to make for this season.
Historically, Football Outsiders has not applied opponent adjustments in DVOA until Week 4. But every couple of years, we run into a situation where the DVOA ratings just look kind of hokey after three weeks because the opponent adjustments aren't being applied yet. This year, for example, some of the teams at the top of the ratings have clearly picked on the have-nots of the league. Denver's three opponents, for example, are a combined 0-6 in their other games, although DVOA is surprisingly sanguine about the New York Giants. (They rank 18th!) Buffalo's three opponents have at least won games, but DVOA sure doesn't like them, ranking the three teams 23rd (Washington), 24th (Pittsburgh), and 25th (Miami).
So I went yesterday and did some research and experimenting, running Week 3 for the past few years as if it had opponent adjustments and comparing it to the actual Week 3 numbers that didn't include opponent adjustments. And what I found was... it didn't matter. The correlation between Week 3 DVOA with opponent adjustments and DVOA at the end of the season was the same as the correlation for DVOA without opponent adjustments.
So, I was left with a dilemma: include the adjustments or not? In the end, I felt that the ratings with the opponent adjustments passed the eye test better than the ratings without, and since the correlation was the same, I could reasonably go with either. And so, this year we've got opponent adjustments a little earlier than usual. However, I only did the adjustments at 30% strength so they fall in line with what we'll do in future weeks. Next week, the adjustments will go to 40% like they always do, they'll be 50% after Week 5, and so on until we are at 100% after Week 10. (Note that the individual stats do not have the adjustment for annoying code reasons; the adjustments on individual stats will begin next week like usual.)
If I told you that I was applying opponent adjustments early without showing you the table of the top teams in DVOA, you probably would guess that Carolina would be one of the teams to take a big hit. Surprisingly, the Panthers stay almost exactly the same. Sure, the Panthers have played the Jets, and the Jets have been very bad. They've also played the Texans, but the Texans haven't been horrendous in the early going. They currently rank 19th in DVOA. The Panthers also dismantled a Saints team that played phenomenal in its other two games, and now ranks eighth in DVOA. So there isn't much opponent adjustment for Carolina, and the Panthers stay in the No. 1 spot.
Here's a look at how the top ten change once we add in opponent adjustments at 30% strength:
|DVOA Through Week 3, 2021|
|Original, No Opp. Adjustments||New, 30% Opp. Adjustments|
Seattle is one of the teams that drops now that we're applying opponent adjustments, but you may be wondering what on earth the 1-2 Seahawks are doing this high in DVOA anyway. Seattle has positive DVOA for all three of its games this year, including both losses. In fact, Seattle had 25.3% DVOA for this week's loss to the Vikings, despite losing by double digits. Look closer at the box score and this does make some sense. The Seahawks outgained Minnesota on a per-play basis, 7.5 yards to 6.2 yards, and had no turnovers.
However, there's more involved than this. One of the weaknesses of DVOA is how it treats drives at the end of a half. I've never figured out a good way to get rid of these "meaningless yards" and improve DVOA's predictive ability. And the Seahawks had a lot of these meaningless yards in this week's loss to Minnesota. They had three passes for 54 yards at the end of the first half, including a 31-yard gain with one second left that didn't score and couldn't stop the clock for a field goal. They had another four passes for 39 yards at the end of the second half, going from their own 1 to their own 40.
How much do these "meaningless yards" change Seattle's DVOA on the season? The rating changes a lot, but the rankings don't change that much. Overall, Seattle's offensive DVOA falls 5.7% DVOA without considering these seven plays. But this would drop the Seahawks just one place in overall DVOA, from 10th to 11th, and just one place in offensive DVOA, from second to third. Over a 17-game season, the effect of these plays will become smaller and smaller and mostly wash out to where it won't make much difference by the time we get to the end of the year. And I'll keep looking in the offseason for the best way to disqualify plays based on the game situation -- or whether I should be disqualifying those plays at all. Perhaps it is more accurate to include them even if they seem meaningless. That's the result I've gotten in the past.
Meaningless yards don't play any role in how well we rate the Seattle defense, though, and the ineptitude of the Seattle defense does seem to be overstated by the Seattle fan base right now. This has not been the worst defense in the league, or close to it. The Seahawks rank 24th right now in defensive DVOA. They rank 26th in success rate allowed. Seattle ranks No. 1 in yards allowed because they rank No. 1 in opposing plays from scrimmage. Their 5.8 yards allowed per play ranks 17th in the NFL. (Compare that to Seattle's 7.4 yards per play, which ranks first in the league by more than half a yard.) It's hard to figure out why opponents have so many plays against Seattle. It's not like the Seahawks can't get off the field on third down -- they actually have a very average -1.4% DVOA on third and fourth downs! It's a bad defense, but it doesn't look anywhere near as bad as people are making it out to be.
Yes, as you can probably imagine, the Seahawks have a crazy halftime split in their performance. On offense, they go from 62.6% DVOA before halftime to -10.2% DVOA after halftime. There's no such change on defense, though, as their defensive DVOA is slightly better in the second half (8.7%) than in the first half (15.2%).
What team is the opposite of the Seahawks? Probably the Los Angeles Chargers. You're probably wondering how on earth the Chargers can be just 22nd in DVOA. Didn't they just beat the unbeatable Kansas City Chiefs? Yet the Chargers have negative DVOA in all three games this year!
OK, it's not that negative in two of the games. The Chargers currently get -1.4% DVOA for their close Week 1 win over Washington, which DVOA sees as essentially a tie. They currently get -7.6% DVOA for their win over the Chiefs, as the Chiefs have the slightly higher DVOA despite losing. Los Angeles will see that rating go up as the opponent adjustments get stronger, especially if the Chiefs are as good as we all think, but right now there's basically no adjustment for playing the Chiefs because Kansas City ranks just 15th overall. The surprising rating might be -24.1% for the Week 2 loss to Dallas, which seems low since that one was close as well, but the Cowboys did outgain the Chargers (7.0 to 6.6) and win the turnover battle.
Really, DVOA isn't that far off from other efficiency stats when it comes to the Chargers. Los Angeles is currently 13th in yards per play and 24th in yards per play allowed. They are 11th in success rate but 31st in success rate allowed at 56%. Only Kansas City (59%) has allowed success on a higher rate of plays based on our definition of a successful play.
Another team that will surprise you in DVOA is the 0-3 New York Giants, who somehow rank 18th, ahead of Green Bay and the Chargers. Well, you know how bad that Week 1 Green Bay loss was, so that part may not be a shock, but still, the Giants are rather high for an 0-3 team. In fact, they are the 10th-highest 0-3 team in DVOA history. (Before applying opponent adjustments, they would have been the eighth-highest 0-3 team.)
|Best 0-3 Teams by DVOA, 1983-2021|
|1995||DET||-1.8%||17||10-6||13.2%||7||Won Wild Card|
|1998||BUF||-5.5%||17||10-6||19.3%||7||Won Wild Card|
This is another team where DVOA isn't far off from standard stats. The Giants have played two very close games, albeit against teams we think are not very good. They currently rank in the middle of the league in yards per play: 17th on offense and 19th on defense.
Unfortunately for the Giants, their schedule is about to get a lot harder. That's true whether we look only at DVOA through three weeks or whether we look at schedule through the lens of DAVE, our rating that combines early-season results with preseason projections to get a more accurate look at what to expect from teams the rest of the way. DAVE, for example, knows that the Kansas City Chiefs aren't really an average team, and it knows that the Carolina Panthers are probably not actually the best team in the NFL this season.
This is something I usually like to do after three weeks, look at how schedule strength for the rest of the season is different if you look at DVOA and DAVE. For example, if you look at Chicago by DVOA, they have an average remaining schedule. DAVE, which is more accurate, shows Chicago with the second-toughest remaining schedule. They play a number of teams that we think are better than they have shown so far this season, led by Green Bay twice but also Pittsburgh and Baltimore. Pittsburgh, incidentally, has the hardest remaining schedule based on DAVE ratings. Chicago, Arizona, Washington, and Green Bay round out the top five. The Giants are sixth after playing one of the five easiest schedules so far based on opposing DAVE ratings.
The Buffalo Bills have the easiest remaining schedule, whether we look at DVOA or DAVE. Other teams with particularly easy schedules the rest of the way include Miami, Indianapolis, Tennessee, and Tampa Bay.
Here's a full table showing schedule strength both so far and over the rest of the season. Both current DVOA and DAVE are listed.
* * * * *
A reminder that all our free stats pages, including DVOA and player position stats, now require registration to view. This is not a paywall! You only need to register (for free) and then log in to the site to view these pages. While you're at it, you can get a seven-day trial of FO+ and check out the FO+ features like a deeper DVOA database, weekly fantasy projections, fantasy football research tools, and picks against the spread.
* * * * *
These is the Football Outsiders Top 16 through three weeks of 2021, measured by our proprietary Defense-adjusted Value Over Average (DVOA) system that 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.)
OFFENSE and DEFENSE VOA are adjusted for opponent and performance indoors and consider all fumbles, kept or lost, as equal value. Because it is early in the season, opponent adjustments are currently at 30% strength. 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 78% preseason forecast and 22% actual performance. It is not currently adjusted for any backup quarterbacks.
To save people some time, please use the following format for all complaints:
<team> is clearly ranked <too high/too low> because <reason unrelated to DVOA>. <subjective ranking system> is way better than this. <unrelated team-supporting or -denigrating comment, preferably with poor spelling and/or chat-acceptable spelling>
Click here for the full table.