Packers, Colts Among Teams with Surprising DVOA Ratings
NFL Week 8 - The Buffalo Bills remain No. 1 in the Football Outsiders DVOA ratings after this week's win over the Miami Dolphins. In fact, the top eight teams remain the same with only one change, Tampa Bay hopping the Los Angeles Rams to rank third overall behind Buffalo and Arizona. Under that top eight, however, there's some surprising movement and teams that aren't quite where you expect them to be in our rankings.
You may remember back at the start of the season, I introduced a formula called Post-Game Win Expectancy (PGWE). This formula asks the question, "Given how each team played in this game, based on VOA splits without opponent adjustments, how often should we expect each team to win the game?" Some games are pretty easy. The Philadelphia Eagles had 99.97% PGWE after beating the Lions 44-6. Not a surprise. But a couple of games each week have a PGWE that does not agree with the actual winner of the game. These "unexpected results" help explain movement in the DVOA ratings because usually it means the winning team had a lower DVOA rating than the losing team. There were three games this week with "reverse PGWE," and let's go through those three games to see how they help explain some of the unexpected movement in the DVOA rankings this week.
Tampa Bay: 79% to Beat New Orleans, Lost 36-27
Tampa Bay outgained New Orleans in this game, 7.3 net yards per play to 5.0 net yards. They did lose the turnover battle, 3-1, but an even bigger issue was penalties. Most penalties are not included in DVOA, because they aren't very predictive. The Buccaneers were penalized 11 times for 99 yards. The Saints were only penalized twice for 10 yards. As a result, the Buccaneers lost despite coming out with the higher DVOA for this game, 27.4% to 17.4% for the Saints.
This helps explain how Tampa Bay passed the Rams for third place in DVOA even though the Rams dominated the Houston Texans. The other issue is that the Rams' DVOA for their win over Houston isn't anywhere near as high as you probably expect. After heavy opponent adjustments for playing the Texans, the Rams only end up with 18.8% DVOA for the game, which is lower than their DVOA on the season. Of course, the issue here is the fourth quarter. The Rams destroyed the Texans in the first three quarters of this game. Then they sat a bunch of their starters, and the Texans came back with 22 fourth-quarter points. Most of these plays in the fourth quarter are down-weighted in DVOA because the game was so out of reach, but even at half weight, they end up lowering the Rams' rating for the game (and thus for the season in general). Here's a look at offensive DVOA for each team in the game:
|Offensive DVOA, LAR-HOU Week 8|
It's tough to know what to do about a team giving up on the game like the Rams did, because it is so rare. Most teams, even when they pull starters in a blowout, don't give up three touchdowns in a quarter. So we're comparing the Rams to the average performance for other teams with big leads, and they are found wanting. Should this be downweighted more based on the idea that this won't have much impact on how the Rams play for the rest of the year? Perhaps, but we're always doing new resesarch in the offseason to figure out where to draw the line on such things. Past research has found that keeping some of the results of blowouts in DVOA (at half weight) still leads to a more predictive rating.
Arizona: 75% to Beat Green Bay, Lost 24-21
Here's another game where you have yardage on one hand and turnovers on the other. The Cardinals outgained the Packers, 6.1 net yards per play to 4.7. The Packers won the turnover battle, 3-0. However, there were four fumbles in this game and the Packers recovered three of them, a little regression of the superb fumble luck the Cardinals had early in the season. Overall, the Cardinals ended up with the better DVOA in this game despite the loss. The Cardinals were at 21.1% and the Packers at 9.4%.
And yet, even after reading that the Packers had the lower DVOA in this one win, you probably are not expecting to see them drop from 11th to 14th in DVOA this week. Why did the Packers drop this week despite having a positive DVOA against the Cardinals and why is a 7-1 Packers team so low in the first place?
The answer to the first question is "schedule strength." A lot of the teams that the Packers played earlier in the season had terrible games in Week 8. The three worst single-game DVOA ratings of Week 8 belonged to the Lions, Bears, and Bengals, all teams that have been on the Packers' schedule. (We'll talk about the Bengals and their horrible performance against the Jets before we're done today.)
There are a couple of answers to the second question. The biggest issue here is Week 1, as we covered last week when we ran ratings that ignored the first week of the season. The Packers were horrible in a 38-3 loss to New Orleans to open the season. They have a positive DVOA for all seven of their wins since that opening game, but none of their wins are even half as strong as that one loss. Let's bring out a week-to-week graph; this one shows the difference between Green Bay's DVOA and VOA for each game for those who are curious.
Take out Week 1, and the Packers would be sixth in DVOA so far. That's a lot closer to public perception. Are we giving too much weight to one game early in the season? Well, we're not giving that game any more weight than any other game. Sometimes one blowout has a major impact on the DVOA ratings, although that tends to become less of an issue the more of a sample we have. It's part of the reason we have the DAVE ratings, to prevent weird early-season results from having too much influence on how we perceive teams. And Green Bay is seventh right now in DAVE. Again, that's closer to public perception.
Week 1 isn't the only issue, however. The Packers don't have any really dominating wins on their resume, with no games over 40% in DVOA. The Detroit, Chicago and Washington wins are knocked down by opponent adjustments. The Bengals win in Week 5 is a strange one. You expect overtime wins to have DVOA close to zero, but the Packers outgained the Bengals by the average of 7.4 yards to 5.6 yards. They also won the turnover battle, 2-1. However, the Bengals were the much more consistent offense in that game, with a 49% success rate compared to 41% for the Packers. The Packers also had terrible special teams in that game, with Mason Crosby missing three different field goals and an extra point plus Cincinnati's Brandon Wilson averaging 30 yards per kick return. So the Packers are just barely over 0% for that win.
The Packers rank third among the lowest-rated 7-1 teams in DVOA history.
|Worst 7-1 Teams by DVOA, 1983-2021|
|*Does not include strikebreaker games|
Most of the teams on this list went 4-4 or 5-3 over the remainder of the season. All of these teams made the playoffs except the 1988 New Orleans Saints back when there were only five playoff teams per conference. The 2015 Broncos, of course, went on to win the Super Bowl, but they do not really resemble the 2021 Packers in any way.
Indianapolis: 68% to Beat Tennessee, Lost 34-31 (OT)
Here's another weird one. How did Indianapolis end up with a better DVOA rating than Tennessee when the Titans actually outgained the Colts, 5.0 to 4.3 average yards per play? Some of the issue is consistency from play to play, as the Colts had 47% success rate and the Titans were at only 38%. There's also the issue of the interception Ryan Tannehill threw which the Colts fumbled back to the Titans. That was effectively a "free possession" for the Titans in terms of DVOA, because we don't give them any credit for causing a fumble on an interception return. That's one of those "non-predictive" plays. It takes skill, but it doesn't really tell us anything about the teams going forward.
With that big gap in success rate, DVOA for this game ends up at 30.2% for the Colts despite the loss, and -2.5% for the Titans despite the win. So the Colts actually move up in the DVOA rankings, from 15th to 10th, while hte Titans stay the same at 20th.
I know, I know. I say "Indianapolis Colts" and the first thing you picture is Carson Wentz throwing a tremendously stupid interception. How on earth is this team 10th overall in DVOA after eight weeks? Well, let's look at the individual components of their DVOA rating.
The Colts are 19th in offense. They don't have a terrible offense. They have a slightly below-average offense. That sounds about right, doesn't it? Take out kneels and spikes, and the Colts are 18th in yards per play so far this season. They are 23rd in success rate. Wentz actually has only four interceptions, which is lower than the NFL average of 5.9 per team. Those stats add up to 19th in DVOA. Makes sense.
The Colts are fifth in defense. This is probably the surprise, although you knew the Colts had some good defensive players such as Darius Leonard and DeForest Buckner. The run defense stands out, currently No. 1 in the league in DVOA. The pass defense is mediocre, currently 20th. This is the side of the ball where the parts don't properly add up to the whole. The Colts rank 22nd in yards allowed per play and 18th in success rate allowed. They rank second in the league in turnovers per drive, but still, that doesn't seem like it should combine to rank the Colts fifth in DVOA. The Colts are particularly strong on third downs (fourth in DVOA) although they are particularly weak in the red zone (28th in DVOA). Third downs and turnovers are likely to regress a bit, but that red zone performance should regress positively. Overall, the Colts are probably not as good as DVOA thinks they are on defense -- they would drop to 13th in the league if we looked at defense without turnovers included -- but DVOA is probably recognizing something in their situational play that makes them better than their raw yardage allowed.
The Colts' schedule is weird. The started with three games below zero, all losses. However, the Colts have positive DVOA in each of their last five games even though two of those were overtime losses. Here's the week-to-week graph:
As for the Titans, I covered them in depth last week. It will be interesting to see where they go from here with Derrick Henry injured, a fascinating test case about the value of a running back. The Titans are ninth in run offense DVOA so far this year, which doesn't seem strong enough to make an adjustment for the loss of Henry in our playoff odds simulation. Like the Packers and 7-1 teams, the Titans come out on a list of the worst 6-2 teams in the history of DVOA.
|Worst 6-2 Teams by DVOA, 1983-2021|
The 2007 Lions are the real cautionary tale for a team that builds an early winning record out of close wins, but the Titans aren't anything close to that. Three of these teams missed the playoffs, the three listed as finishing 7-9 of 8-8. On the other hand, the 2007 Giants won the Super Bowl and three other teams on this list made it to the Super Bowl and lost, so there's nothing that says that a low DVOA means the Titans don't have any chance to make noise in the playoffs. Of course, it helps if you dramatically improve in the second half of the season the way the 1998 Atlanta Falcons did.
Let's cover one more game, even though it didn't have reverse PGWE:
New York Jets: 98% to Beat Cincinnati, Won 34-31
This is the opposite of reverse PGWE, actually. This is when DVOA suggests a win is much stronger than the final score indicates. We cover this game further in Any Given Sunday this week, but what's important for DVOA purposes is the big gap in success rate between the two teams. The Jets' short-passing offense and strong running game ended up with a 59% success rate, the best for any offense in Week 8. The Bengals were at just 45%. Combine that with very strong opponent adjustments for playing the Jets and the Bengals end up with -59.9% DVOA for this game.
As a result, the Bengals' overall DVOA for the season plummeted this week. They dropped from 11.6% DVOA (ninth) all the way to -1.1% DVOA (21st). This week's game isn't the only issue. Like we noted above about Green Bay, the Bengals fall a little bit because past opponents now look weaker, particularly Chicago and Detroit. Cincinnati ranks 31st in past schedule and the opponent adjustments are getting a little stronger in the DVOA system every week until we hit Week 10. Also, there is a very tight pack of teams between 16th (Chargers, 1.2% DVOA) and 21st (Bengals, -1.1% DVOA) and the Bengals just happen to be at the bottom of that pack. There's really not much difference between those six teams.
* * * * *
Football Outsiders playoff odds, snap counts, and the FO+ database are now all updated through Week 8. A reminder that all our free stats pages, including DVOA and player position stats, 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. FO+ is now on sale for just 99 cents per week if you buy an annual subscription, so make sure to check it out!
* * * * *
Here is the Football Outsiders Top 16 through eight 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 DVOA are adjusted for opponent and performance indoors and consider all fumbles, kept or lost, as equal value. Opponent adjustments are currently at 80% strength and will increase 10% each week until Week 10. 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 50% preseason forecast and 50% actual performance for teams with seven games played, and 45% preseason forecast and 55% actual performance for teams with eight games played. 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.