Why the Vikings Were So Bad in DVOA Last Season- And Why They’ll Be Better This Season

The Vikings had a wild and unconventional season last year. They won ten straight one-score games and they went 13-4- but with a negative three point-differential. They engineered the biggest comeback in NFL history. They had a top ten offense, which was roughly top ten in key metrics like 3rd down conversions and red zone TD%. They were masters of situational football, producing key plays when they needed them most. Kirk Cousins tied the all-time record of eight game winning drives in a season. They also finished with the fourth-highest overall team PFF grade at 92.5, behind only the Eagles, Chiefs, and 49ers.

But there were also signs of stress.

The greatest comeback in NFL history was against the Colts- one of the worst teams in the league last year. The Vikings’ defense was one of the worst units in the league. And that situational mastery often masked poor play through the middle of many games. Their negative three point-differential was also a sign of stress, as was giving up nearly 500 more yards than they gained over the course of the season.

And so while the Vikings were tied for the second-most regular season wins last season, they were the worst 13-game winning team in NFL history by many metrics. They ranked well down the list in many advanced metrics as well, including one called DVOA. In fact, they had the worst DVOA of any 13-win team in the history of the metric going back to the mid-1980s.

What is DVOA and Why Does It Matter?

DVOA stands for Defense-adjusted Value Over Average and the ultra-short definition is that it measures a team’s efficiency by comparing success on every single play to a league average based on situation and opponent. It is meant to add context to a three-yard run on first-and-ten compared to a three-yard run on 3rd-and-two, for example. And based on the context of the play, it endeavors to assign whether a play was successful or not. For example, on first down a play is considered successful if it gains 45% of the yards needed for a first down (or touchdown in goal-to-go); on second down if it gains 60% of the yards needed; and on third or fourth down only if it gains 100% of the yards needed.

A point value is then assigned to each play that reflects how successful or unsuccessful the play was, given the context, and those point totals are accumulated over the course of a game. The exact point formula is proprietary, but from what Football Outsiders, who publishes DVOA statistics, has provided online:

We then expand upon that basic idea with a more complicated system of “success points,” improved over the past few years with a lot of mathematics and a bit of trial and error. A successful play is worth one point; an unsuccessful play, zero points with fractional points in between (e.g., eight yards on third-and-10 is worth 0.54 “success points”). Extra points are awarded for big plays, gradually increasing to three points for 10 yards (assuming those yards result in a first down), four points for 20 yards, and five points for 40 yards or more. Losing three or more yards is -1 point. Interceptions occurring on fourth down during the last two minutes of a game incur no penalty whatsoever, but all others average -6 points, with an adjustment for the length of the pass and the location of the interception (since an interception tipped at the line is more likely to produce a long return than an interception on a 40-yard pass). A fumble is worth anywhere from -1.7 to -4.0 points depending on how often a fumble in that situation is lost to the defense — no matter who actually recovers the fumble. Red zone plays get a bonus: 20 percent for team offense, five percent for team defense, and 10 percent for individual players. There is a bonus given for a touchdown, which acknowledges that the goal line is significantly more difficult to cross than the previous 99 yards (although this bonus is nowhere near as large as the one used in fantasy football).

Those point totals are then adjusted based on the success rate of the opposing defense in stopping those plays over the course of a season. So, against a better defense, a premium is awarded, while against a worse defense a discount is applied. The adjusted result is then compared to “league average” and given as a percentage above or below zero. In measuring offense, the more positive number the better. In measuring defense, the more negative number the better, and for special teams the more positive number the better. The three results, reversing the sign (+ or -) for the defensive score is then totaled to provide the overall DVOA for the team.

Without getting too far into the statistical weeds- which we can’t really do anyway because the exact DVOA formula is proprietary- the scoring system for DVOA rewards smaller but more frequent successful plays over less frequent big plays that achieve the same result. For example, a team that has a seven play, seventy-yard drive for a touchdown consisting of seven plays each gaining ten yards, will accrue 21 success points, plus the redzone and TD bonus, while a one-play, seventy-yard touchdown will accrue just 5 points plus the TD bonus. So, while both drives achieved the same result, the former would score over 400% more points than the latter. That’s an extreme example, but it drives home the point that this efficiency model values frequent singles over occasional home runs, to use baseball parlance. And more frequent unsuccessful or negative plays that occur in a touchdown scoring drive are likely to produce a more modestly positive DVOA- or maybe even a negative one. A TD scoring drive that started on the opponent’s 30-yard line and consisted of a strip-sack recovered by the QB for a loss of five yards followed by a 35-yard touchdown pass could result in only marginally positive or even negative DVOA despite it resulting in a touchdown. Again, an extreme example, but it’s meant to highlight the impact negative plays can have on DVOA calculation despite a successful result- a touchdown scoring drive.

A real-world example of how some more extreme examples in DVOA calculation can play out over the course of a game was the Vikings-Bills game last season.

As you may recall, the Vikings won that game 33-30 in overtime in the NFL game of the season. It was the game of the season because it was highly dramatic with many exceptional plays. But it was those exceptional plays that caused DVOA to be wildly off from a correlation-to-winning standpoint. In that game, the losing Bills had a DVOA of +35.9% - a very high number- while the winning Vikings had a DVOA of -19.0% - a very low number. Even the Vikings offense had a negative DVOA in that game (-10.6%) despite outstanding plays by Justin Jefferson, Dalvin Cook, and Kirk Cousins.

But breaking that game down, the Bills had a 10-play, 70-yard TD scoring drive, a 4-play, 47-yard TD scoring drive, and a 5-play, 71-yard scoring drive. On the other side, the Vikings had a one-play, 81-yard TD scoring drive, a 6-play, 74-yard TD scoring drive featuring 3 unsuccessful or negative plays, a 13-play, 66-yard TD scoring drive featuring 7 unsuccessful plays, and a defensive touchdown due to an aborted snap- which defenses do not get any credit or success points for according to the DVOA formula. There were also four other turnovers that may have been accounted for differently according to the proprietary DVOA formula, which we don’t know precise details about.

Thinking again about the aspects of the DVOA formula that Football Outsiders has revealed, and the examples mentioned above, we can see how the Bills’ TD drives resulted in much greater positive DVOA than the Vikings’ TD scoring drives did. It stands to reason also that the Bills’ defense, having given up fewer points and generating more negative or unsuccessful plays, fared better than the Vikings defense, which gave up more frequent successful plays and more points from a DVOA perspective.

And while that was an unusual game to be sure, it was a microcosm of the season for both teams from a DVOA perspective. The Bills led the league in DVOA last season at +35.1%, while the Vikings ranked a miserable 27th at -13.6%. This of course implies that the Bills were a much better team than the Vikings last season, despite the fact that the Vikings beat them at Buffalo, and that both teams won 13 games during the regular season. Even the playoffs produced a similar result for the both teams, with the Vikings losing to the Giants while the Bills narrowly beat the Tua-less Dolphins at home (the Vikings beat the Tua-less Dolphins on the road) before falling decisively to the Bengals 27-10 in the divisional round.

All that suggests that DVOA is a flawed metric that doesn’t accurately represent how good (or bad) a team is, or corresponds much overall to winning, or even assigns a representative value that corresponds to which team won and which team lost on a per game basis. Indeed, DVOA is a flawed metric, just like every other advanced metric attempting to make sense of the game of football. This is football being measured here after all- not physics.

And yet it’s far from worthless. It can be insightful, and when looked at in larger sample sizes, over several seasons, it becomes much more representative and correlates at a high rate to winning on a seasonal basis- very similar to point differential in that regard.

It’s also somewhat intuitive: it makes sense that teams with more frequent successful plays and less frequent negative plays, on both sides of the ball and on special teams, are more likely to win more consistently than teams that have more negative plays but occasional more high impact plays that are more difficult to achieve.

The counterexample to this, which isn’t uncommon, is that good teams somehow manage to win games even when they’re not playing their best. That’s true enough, and if that happens a few times over the course of a season, it holds up. But if a majority or nearly every game in the season fits that description, then that axiom becomes more suspect. And such was the case with the Vikings last season.

2022 Vikings Season DVOA

Football Outsiders

Above is the Vikings 2022 season by DVOA broken down on a per-game basis, and by offense and defense DVOA, each in turn broken down into pass and rush DVOA, and special teams DVOA. The last column is Post-Game Winning Expectancy or PGWE, which is the probability the team will win the game based on the play-by-play and game situations encountered.

Let me draw your attention to a couple key takeaways from this chart:

• Offensively, only Pass DVOA was generally positive- and occasionally highly positive-consistently over the course of the season, Dallas and Philadelphia excepted.
• Defensive Pass DVOA was frequently very bad (high double-digit positive %) and only three games significantly positive (double-digit negative DVOA %).
• Offense Rushing DVOA had a few good games early in the season but was a major liability the entire second half of the season, apart from the Jets game.
• Special teams had a significant negative impact on the game more often than it had a significant positive impact and was a net negative for the season.
• Defensive Rush DVOA was modestly positive for the season (-4.2% DVOA) with seven significantly positive (double digit negative DVOA %) games.

Overall, it’s easy to conclude that the major liabilities for the Vikings last season were:

1) Pass Defense

2) Rushing Offense, particularly the second half of the season

Looking Ahead to This Season

The most meaningful change the Vikings made this off-season was to replace defensive coordinator Ed Donatell with Brian Flores. The change in coordinators is likely to be more impactful than in other like cases due to the dramatic difference in the defensive scheme and philosophies of the two coordinators. In fact, I’m not sure there are two more diametrically opposed approaches to defense by any two defensive coordinators in NFL history, at least from what we’ve seen from Ed Donatell last season with the Vikings and from Brian Flores while he was in Miami.

From Ed Donatell...

Ed Donatell’s approach last season was a passive, prevent defense- rushing four with a below league average blitz rate, defensive backs giving 8-10-yard cushions in zone defense, which he played a league-high 85% of the time. Philosophically, Donatell was willing to give up underneath routes all afternoon, assuming eventually the offense wouldn’t be able to execute and convert a third down and the drive would stall, or they’d fail to score a touchdown- which was also considered a win.

From a DVOA perspective, this approach was philosophically and schematically designed to do poorly in DVOA, particularly Pass DVOA. Which it did. Allowing those modest, but more frequent, successful pass plays in hope of preventing a big play or a touchdown, is a recipe for bad defensive DVOA. And not only was Donatell’s defense bad from a DVOA perspective, it was also the most consistently bad. It had the least amount of DVOA variance from game-to-game of any defense in the league- which also supports the ‘bad-by-design’ aspect of the scheme.

... To Brian Flores

By contrast, Brian Flores’ approach in Miami was an aggressive, attacking defense- rushing five or more 33% of the time with an above average blitz rate, defensive backs in man coverage, high rates of Cover-0 (no deep safeties) with a focus on pressuring the quarterback to generate negative plays. Flores was willing to risk giving up the occasional big play to the offense if he could generate more negative plays to force stalled drives.

From a DVOA perspective, this is philosophically and schematically designed to do well in DVOA, particularly Pass DVOA, barring significant incompetence in executing the scheme.

After Flores’ first season as head coach in Miami, an injury-riddled season defensively, Flores’ defense went from at or near the bottom of the league in DVOA the previous two seasons, to 11th in 2020- a huge 26.3% improvement in total defensive DVOA from the previous year and an even more massive 47.5% improvement in Pass DVOA. And without much of a defensive roster either. It was ranked 10th in DVOA in 2021 under Flores.

Interestingly, as Flores brings his scheme to the NFC North, one quarterback in particular has had two nightmare games against Brian Flores’ defense: Jared Goff- current quarterback of the Lions, the early favorite to win the division this season. Goff had four turnovers- 2 INTs, 2 lost fumbles- and a 65.7 passer rating against Miami in 2020 while with the Rams, and a 57.9 passer rating with the Rams in the Super Bowl following the 2018 season with 1 INT, 4 sacks, and having completed only 50% of his pass attempts.

Young quarterbacks often find Flores’ attacking, deceptive, blitz-heavy scheme difficult to navigate. With two young QBs in the NFC North this season, Jordan Love and Justin Fields, it will be interesting to see how they fare.

In terms of year-over-year changes in DVOA, defensive DVOA has proven to be more volatile according to Football Outsiders, while offensive DVOA has been more stable. The bi-polar changes the Vikings have made in terms of defensive scheme and philosophy would tend to support a more dramatic change in DVOA this season. And given where the Vikings finished last season in defensive DVOA (27th), greater year-over-year volatility is more likely to benefit than hurt them this season.

Offensive Changes Too

Kevin O’Connell has said that what you focus on, or what you emphasize, tends to be what you get when it comes to coaching. This off-season, O’Connell has talked about becoming more effective running the ball on offense. Undoubtedly, he’s familiar with the Vikings’ DVOA in running the ball last season and wants to improve in that area.

Breaking it down, the Vikings offensive line generally gets good PFF grades in run blocking, although their run block win rate is near league average at 71% - same as the 49ers and Titans last season. But 72% would’ve put them tied for 8th best, as there isn’t as much variance across the league in run block win rate as in pass block win rate. So, despite average or above average performance in run blocking, the Vikings’ offense was near the bottom in Rushing DVOA last season. Part of that was likely due to the decline of Dalvin Cook, who took the lion’s share of the running back rush attempts last season. Sometimes a running back can be blamed for not getting the yards blocked for him by not hitting the hole quick enough, leading to unsuccessful run plays that hurt DVOA. But perhaps some of the poor rushing DVOA was due to poor run blocking at the tight-end position, given all of the offensive linemen graded relatively well in run blocking (they were the top 5 best run blockers on the team), which can often be an important blocker in both mid-zone and some gap/power runs which the Vikings run. Johnny Mundt was among the worst run-blocking tight-ends in the league last year with 50% of the run blocking snaps- and TJ Hockenson wasn’t that much better.

In order to improve their running game efficiency, and Kevin O’Connell particularly mentioned being able to run better on early downs to setup more favorable third-down conversions, the Vikings are moving on from Johnny Mundt as the primary blocking tight end, in favor of new free agent acquisition Josh Oliver- who was the highest graded run blocking tight-end in the league last season. He was also one of the best pass-blocking tight ends, particularly in true pass sets (which exclude plays with less than 4 rushers, play action, screens, short dropbacks and time-to-throws under 2 seconds) where he ranked sixth best in the league.

The Vikings also appear to be moving on from Dalvin Cook, who at age 28 has exited the prime for running backs, in favor of Alexander Mattison (24) and likely Ty Chandler as well. Mattison graded higher than Cook last year as a rusher, and while it makes sense to go with a younger back (and definitely from a salary cap perspective), it’s unclear if Mattison can deliver better results with a higher volume of snaps. Mattison also did better as a pass blocker than Cook last year, which also benefits the offense, although not the running game.

The addition of Josh Oliver also suggests the Vikings will use more two-tight-end sets (12 personnel) than they did last season. This could lead to benefits in both the run and pass game, as it could force defenses to use their base defense more often, which has benefits to the pass game, or gives an advantage to the run game if opposing defenses maintain a nickel defense and/or two deep safeties. In any case, Oliver’s better blocking ability in both the run and pass game can only improve offensive efficiency.

The only other significant change offensively will be the replacement of Adam Thielen with Jordan Addison. It remains to be seen how fast Addison can get acclimated to the NFL and if he can earn the WR2 position that Thielen held last season, but even more than Dalvin Cook, Adam Thielen is well into the back-nine of his career. His ability to separate from defenders last season declined noticeably, as did his production. Among wide receivers with at least 88 targets last season, Thielen ranked 41st of 45 in PFF receiving grade. KJ Osborn ranked 43rd. So, while we don’t know how well Addison will do this season as a rookie, the bar to clear is not as high as you may think.

Bottom Line

The Vikings were one of the worst teams in team DVOA last season, ranking 27th, and were historically bad for a 13-win team. Most of that had to do with poor defensive DVOA, particularly Pass DVOA, but also Rushing Offense DVOA.

The Vikings have made a major philosophical and schematic change in their approach to defense this off-season by replacing Ed Donatell with Brian Flores. In fact, there is unlikely a more diametrically opposite approach to defense between two defensive coordinators in NFL history, contrasting Ed Donatell’s approach with the Vikings last season and that of Brian Flores’ with the Dolphins. Flores has previously made dramatic improvements in defensive DVOA during his time at Miami, and he could do the same in Minnesota.

Offensively, Kevin O’Connell his indicated he is focused on becoming more effective in running the ball and has made some key personnel and likely schematic changes regarding use of double-tight-end sets to help him accomplish that goal- which may also result in benefits to the passing game. And while the offense was the strength of the team last season, Passing DVOA was only about league average (15th), while Rushing DVOA ranked 28th.

Overall, the changes and focus of the coaching staff this off-season are likely to lead to significant improvement in DVOA this season. It may not lead to more wins- after all the Vikings were tied for the second-most wins last season with 13- but it could increase the margin of victory, point differential, and other advanced metrics for the Vikings this season- perhaps even leading to a deeper run in the post-season.

Stay tuned.