The Minnesota Vikings defense clearly led the way for the Vikings near-playoff berth after the team went through a string of injuries on the offensive side of the ball, and at one point were the top defense in the NFL.
Last week, we went over how different organizations and statistics ranked the players on the Vikings offense. We can do the same thing for the Vikings defense, in part as a means to see how the defense did what it did, but also find areas for improvement.
The numbers we’re listing are not their actual grades for those groups, but their percentile rank among those groups; a player win the 10th percentile of his grade for Pro Football Focus didn’t receive a 10.0 score from PFF, but rather was only better than 10 percent of his peers, with 90 percent of his peers with higher scores.
Because Football Outsiders doesn’t really provide DVOA for defensive players, I used the DVOA of WR1, WR2, and Other WR for the cornerbacks and combined their defensive line scores (for run defense and adjusted sack creation) to create a general defensive line metric from them.
As for the statistical scores used in the final column, I combined some advanced statistics metrics to provide some clue as to their defensive output. They aren’t perfect, especially because they’re defensive statistics, but they at least give us some clue as to how the defenders performed.
For defensive linemen, I used run stop rate (how often a tackle in the run game resulted in an offensive failure) and pressure rate. Edge defenders were compared to other edge defenders, nose tackles to nose tackles and three-technique tackles compared to three-technique tackles.
Linebackers were more difficult to figure out, in part because coverage numbers for cornerbacks are easier to come by than linebackers or safeties. I ended up choosing a combination of run stop rate, missed tackle rate and target rate (how many passing targets a player received in a defender’s coverage vs. the number of coverage snaps they took).
For cornerbacks, I used coverage success rate and target rate. That wasn’t too difficult.
For safeties, I used the same statistics as linebackers, with a reduced weight for run stop rate.
And in chart form:
There’s a lot more variability among the defensive scores than there are for offensive scores—even after excluding Anthony Harris’ statistical production. The differences between the PFF and the Bleacher Report scores are massive for quite a few players: Brian Robison, Anthony Barr, Eric Kendricks, Xavier Rhodes, Terence Newman, Captain Munnerlyn, Trae Waynes and Andrew Sendejo.
By contrast, the offensive only had five players (Sam Bradford, Matt Asiata, Jeremiah Sirles, Rhett Ellison and Adam Thielen) with those large differences.
Alternatively, those most agreed-upon players are ones that would be pretty intuitive. Harrison Smith had nearly identical grades from every grading method, and Linval Joseph wasn’t far behind that consensus. Unfortunately, Shamar Stephen found himself in the same spot for a bad reason, and he was the third-most agreed-upon player among the Vikings defensive group.
Tom Johnson, Danielle Hunter and Chad Greenway round out the group of players who forced a consensus from radically different methods.
Defensively, I’d argue that the strength for the Vikings is the defensive line. While the secondary’s starting five had a higher average rank than the defensive line’s starting four, play flowed out from the success of the defensive line.
And, in either case, the top scores among the positions (Everson Griffen, Danielle Hunter, Tom Johnson and Linval Joseph) do beat out the top scores among the secondary (starters Harrison Smith, Andrew Sendejo, Captain Munnerlyn, Xavier Rhodes and Terence Newman) despite half of the top defensive linemen playing a rotational role.
There was essentially no disagreement for the top performer. Linval Joseph was, at minimum, a 94th percentile nose tackle. His statistical output topped all nose tackles, including the somewhat more well-regarded Damon Harrison.
That’s not a huge surprise; we’ve been singing his praises all year. Linval Joseph is a force to be reckoned with, and the graders seem to agree.
After Joseph, the next-best scoring starter is Everson Griffen, whose film and technical grades were higher than his statistical output. It might help that BR gave a specific score for explosion off the snap, where Griffen shined relative to his peers (he had the highest such score among the overall top five 4-3 DEs).
Both BR and PFF graded his pass-rush ability overall higher than his actual pressure rate, which might be the result of playing against a series of quick-strike offenses in the middle of the season. For example, he abused Ereck Flowers against the Giants, but Eli Manning got rid of the ball extremely quickly. Following that, they played games against the short-passing Eagles offense and a lightning-fast Detroit.
Brian Robison had issues producing, and Pro Football Focus punished him for it, but Bleacher Report thought his approach was sound enough—if they gave him a bit of a boost because of a run-specific assignment in the defense that would make sense, but his run defense score was not particularly higher than his peers, with his pass pressure score pulling most of his weight.
Danielle Hunter had astounding statistical output, but his presence on pass-rushing downs likely inflated his score. As it is, I think PFF may have underrated him, but it makes sense that within the context of the plays he was in on, production was much more likely. His over 85.2 score is extremely good for such a young player and bodes well for the future.
Other than that, the evaluations agreed on Shamar Stephen and Tom Johnson, which seems fitting.
Linebackers were incredible fodder for disagreement. It’s probably not relevant that every group disagreed on Chad Greenway to some extent, because they all essentially agreed that he was a lower-tier player at his position. The more interesting disagreements were between Eric Kendricks and Anthony Barr.
Relative to the other measures, Bleacher Report massively overrates the impact and play level of Anthony Barr and Eric Kendricks. There are distinct differences between the “scores” constructed from chosen statistical measures and PFF grades, but in different directions. PFF thinks Barr’s impact was smaller or worse than his statistical output, while they favored Eric Kendricks.
There could be a couple of reasons for this, but fundamentally it may be because of the poor sideways construction of those statistical scores. Target rate is a decent-at-best measure of coverage ability that loses its focus the more coverage responsibility a player has.
Good cornerbacks deter targets. Great cornerbacks thrive on them.
The best example might be in the 2015 play of Johnathan Joseph and Marcus Peters. Joseph was not targeted often, didn’t allow many yards or have many touchdowns. He also only grabbed one interception. Peters, on the other hand, had seven interceptions but allowed eight touchdowns and twice as many yards. He was targeted the fifth-most of any cornerback that year.
PFF used this argument for why Peters was overrated, but it may be that they undervalued turnovers. Peters ended up with a lower passer rating allowed and adjusted yards per attempt allowed. More importantly, he rendered quarterbacks impotent. Joseph, a very good corner, rendered receivers impotent.
Peters’ high target rate resulted in 64 failed passing attempts in his direction; an eighth of a team’s passing attempts over a year. Joseph’s relatively average target rate but solid coverage resulted in 40 failed passing attempts in his direction.
That same phenomena could be responsible for the iffy statistical score for Eric Kendricks, though that’s probably partially true at best. With one interception and nine pass deflections, he ranks 13th of 85 linebackers in pass deflection rate.
That’s pretty good, and forcing incompletions or interceptions is good play. Beyond that, Kendricks gets marked for targets that are screen passes, which if you’ll recall, he was pretty good at stopping—especially in the second half of the season.
So in this case, one third of the formula marks him negatively for what could be considered a positive.
As for Barr, I think the scores don’t capture how much he was missing. A low stop rate does get to the issue from one perspective, but his inability to get through blocks created significant problems aside from not being able to generate tackles for loss. Not only that, the Vikings often hid him in coverage, so his target rate was fairly low.
In either case, Pro Football Focus was probably closer to the true grades of the linebackers than Bleacher Report or the statistical scores—though they were probably too harsh on Anthony Barr.
If the defensive line was the foundation for the Vikings defense, the secondary provided the skeletal framework—the line’s play set up the secondary to create the obvious splash plays that undergirded what we saw of the defense: interceptions, incompletions and frustrated receivers.
There’s an interesting inversion between Bleacher Report and PFF scores for the two outside cornerbacks. Terence Newman is as well-regarded by Pro Football Focus as Xavier Rhodes is by Bleacher Report. On the other hand, Newman seems not to get his due from BR.
Pro Football Focus explicitly does not take into account the strength of opposition, and we’ve argued in the past that this means they tend to overvalue Newman and undervalue Rhodes.
At the same time, BR’s relatively average rating for Newman needs to answer the question of why Newman is merely above-average in their rankings while allowing the fewest yards per coverage snap in the NFL. Even if it’s against number two receivers, holding an opponent’s second-best receiver to 259 yards over the course of an entire season deserves significant credit.
No fanbase would be happy if their second-best receiver couldn’t crack 300 yards, or even get close to it.
PFF and the statistical scores found Captain Munnerlyn to be an above average cornerback, but Bleacher Report didn’t like what they saw; Munnerlyn is somewhat polarizing within the Vikings fanbase, but fans should consider that Munnerlyn was targeted just as often as Xavier Rhodes and had the same success rate in coverage.
That’s not to say he’s as good as Rhodes; the argument for the Vikings draftee comes from who he played and his ability to create splash plays this year—something that doesn’t translate for Munnerlyn. But Munnerlyn in the slot isn’t playing against mere third options; the slot has had some devastating players.
Often, when number one receivers try to hide from Rhodes, Munnerlyn has to pick up the slack—forcing him into matchups with Kelvin Benjamin, Odell Beckham and Jordy Nelson. That’s not to mention the threats meant to come from the slot: Larry Fitzgerald, Sterling Shepard, Golden Tate, Randall Cobb, Cole Beasley and more.
Entering Week 12, he had only allowed one touchdown, per PFF and generally held up his end of the bargain.
Harrison Smith’s consensus entry at the top isn’t a surprise; Luke Inman broke down his game a few months ago and it all holds up.
As for Andrew Sendejo and Anthony Harris, that’s a bit of a tougher nut to crack. Harris likely didn’t play well and incidentally grabbed good rate statistics from limited time on the field (the cutoff for any player, offense or defense, was 200 snaps—but that may be too few for the players furthest away from the ball); it is more likely that the small-sample statistics are more misleading than an agreement between PFF and BR are wrong.
In the case of the statistics, he had a very poor coverage grade but an excellent run defense grade—with an unusually high run stop rate for safeties (7.1%) and a very low broken tackle rate. But one fewer run stop, and he drops to the 68th percentile overall instead of staying near the top at the 78th percentile.
Sendejo, on the other hand, may deserve more respect from Vikings fans. In what is undoubtedly a career year for him, he posted excellent statistical grades, and Bleacher Report thought he had acquitted himself well. Sendejo was targeted less often than nearly all safeties and produced a play on the ball a surprisingly high 15.8 percent of the time.
Still, above average-broken tackle rates and below-average run stop rates means that his run defense stood out as a weakness for the Vikings this year. If he was hidden in coverage, it means that his statistical case is probably even weaker than Harris’.
That said, his improvement overall as a safety is a surprising asset for the Vikings. Minnesota fans have wanted an investment in safety for some time—or development from one of their late-round projects to push for a starting spot.
But if the other safety position next to Smith is being held down adequately, that investment priority may not matter as much as other clear needs.
This piece ran a bit longer than its companion on the offense, but hopefully still held enough insight to provide us with the proper context for evaluating the Vikings and the organizations that make their hay evaluating the NFL.