Subscribe to Blog via Email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.

Join 6 other subscribers

MN Vikings Tweets

Bleacher Report – Vikings

Arif’s deep stats dive: Putting loss to Eagles in perspective

By Arif Hasan

After faltering in full for the first time this season, the Vikings loss has turned them from the sole undefeated team to one of a few one-loss teams. Losing the luster of the undefeated season is disappointing but no one is surprised that it seemingly comes at the hands of the offensive line. How exactly did the various units of the Vikings perform and how did those performances inform each other?

We’ll take a look at a number of advanced statistics to begin to answer those questions, acknowledging that these advanced statistics will only ever tell us part of the story. What went wrong and needs fixing and what went right that needs emphasis?

As a reminder, anyone needing a primer on the nitty-gritty on a lot of these statistics should check out the Week One article. If you see an unfamiliar term, there’s a good chance it’s explained in that article or an accompanying link.

Team Level Statistics

First, let’s look at what we can tell about the state of the team overall. We’ve largely gone over two—points per drive and drive success rate, acknowledging that drives (possessions) are the best unit of football analysis, because slower teams can score better than other teams without necessarily leading the NFL in points per game. We’ll look at a few others as well this week.

In points per drive, the Vikings offensively ranked 23rd of the 28 teams who played this week, grabbing 0.91 points on every offensive possession. That’s the worst offensive performance of the season—despite the fact that they were in some very sticky situations early on. 0.91 is frankly awful; the worst team in the NFL in this statistic over the course of the season (the Chicago Bears) scored 1.33 points per drive.

Defensively, things look a little better; they ranked 11th this week—only allowing 1.67 points per drive, which over the course of the season would rank ninth. What’s even more interesting is that this respectable performance is actually their worst defensive performance in this metric this season.

Still, that resulted in the worst net points per drive of the year at -0.76. There were seven other teams this week who did worse, so it wasn’t the national embarrassment that the Bears game was for Chicago, but it’s a score that would rank 29th in the NFL this year over the season.

From a year-to-date perspective, the Vikings rank eighth in the NFL. The teams above them are almost all clearly quality teams (Philadelphia, New England, Dallas, Buffalo, Arizona and Seattle) and a team that has played very well for about 50 minutes a game (San Diego).

In drive success rate (a measure that mostly looks at , the Vikings were unsurprisingly tepid on offense. Their rate of 0.655 would rank 30th over the course of this season and demonstrates the fact that their inability to create first downs was fatal. Obviously, specifics might come into play—the uncalled pass interference on Stefon Diggs and a potentially poor spot on 4th and 1 would change things, but ultimately the offense just didn’t do its job.

They were against one of the top defenses in the NFL in this respect, but the offense generally needs to do better—not great, but better—in order to be more competitive.

Defensively, they held down an otherwise strong Eagles offense (ninth in drive success rate this season) to an even worse rate than the Vikings offense: 0.593.

That net drive success rate is positive and though the Vikings were net negative in points per drive, it might be the case that the teams were closer than the final score led us to believe—if one takes away the eight points scored on special teams, the evidence points to a game that played out like a 10-13 contest.

While that probably gives the Vikings too much credit—the final touchdown occurred on a drive that felt more like garbage time than anything else—it does speak to some of the key differences and how close this game might have been; had Andrew Sendejo turned his interception into a touchdown, or had the Vikings added seven from that amazing field position, the story would have been much different.

Speaking of garbage time, we do have a way to account for it. We can also better capture the “feel” of the game, so if it feels like one team is dominating there’s a small way to account for that. Game scripts basically take the score at every second of the game and average it out. So if a team scores 24 points early and doesn’t give up any points until they give up ten late in the fourth quarter, you’d think that team was dominant throughout the game and only let up a little in garbage time—24-10 does not accurately represent the difference in quality.

This accounts for that.

The game script score for the Vikings is -7.4, which means the average point differential throughout the game was slightly more than a touchdown. Given that the Eagles didn’t get over a touchdown’s worth of a lead until seven seconds in the first half, this makes sense despite the 18-point differential they held for six minutes near the end.

Naturally, the average game script score for the NFL is 0.0, and the most extreme scores this year so far have been 24.5 (Pittsburgh over Kansas City in Week 4), 20.0 (San Diego over Jacksonville in Week 2), 18.9 (Seattle over San Francisco Week 3), 17.3 (Dallas over Cincinnati Week 5) and 17.0 (Minnesota over Houston Week 5).

Obviously, that -7.4 isn’t good but it’s worse than most teams will put together. Only 14 games of the 92 games played from Weeks 1-6 featured a game script score of 7.4 or larger, so showing at the bottom 15 percent of single-game performances from this perspective is disappointing.

The significant difference between the game script score and the relatively neutral net scores in points per drive and drive success rate really bring home the significance of the special teams touchdown; the play from scrimmage was not too far apart, but the early return touchdown defined the mood of the contest.

While the coaching staff will need to drill in on what allowed the score, Vikings fans should feel somewhat emboldened that the biggest difference between the two teams in total performance was an unlikely, unsustainable score.


It may be obvious to say that Bradford had a poor night, but contextualizing that performance can be a big part of our reaction. For example, a passer rating of 71.6 is below-average and speaks to a passing performance that will need to be better in the future.

Throwing in sack numbers makes that point even better: an adjusted net yards per attempt of 3.5 is extremely bad; no quarterback will put up numbers that bad over the course of a season (by three quarters of a yard) and that ends up as 184th-ranked performance of the 200 single-game quarterback performances we have this year.

Bradford’s performance in the various advanced statistics don’t really disagree:

Metric This Week Rank vs. 2015 Total 2016 Rank
PFF 18 26 7
Implied YPA 26 24 22
Passer Rating 22 33 8
ANYA 25 26 10
ESPN QBR 28 18 13

PFF is friendliest to him, ranking him 18th for the week, while ESPN’s QBR is the harshest and gave him the lowest score of any of the quarterbacks who played this week (except for those who played on Monday night, who don’t have a score yet).

Perhaps one reason that QBR ranks him lower than the other statistics has to do with his amount of yards-after-catch in this game, over 50 percent of the total passing yards. We know that PFF ranked him just a little bit higher than the other measures in part because of the amount of pressure Bradford was under.

In this game, he was under pressure on over 42 percent of his dropbacks—among the highest rate in the NFL, but still about average for a Vikings game. It may have felt like more, however, as Bradford was hit on almost every play he was pressured—with 13 hits and six sacks comprising 95 percent of the pressures he faced.

Against pressure, Bradford played closer to his career average than he has against pressure so far this season, where he was throwing with a passer rating of over 100 when there were defenders bearing down on him. In this game, his passer rating was 16.4.

So Bradford was under pressure too often, that pressure converted too often, and he did a poor job against pressure when it appeared.

For what it’s worth we have some season-long work on a new metric by Bryan Frye, called total adjusted yards per play. His adjustments also give bonuses to first downs, which is significant because consistent quarterbacks tend to help teams win more often than inconsistent quarterbacks with similar rate statistics.

This measure also credits all quarterbacks for their running plays and incorporates rushing yards and fumbles. Here are the year-to-date numbers, adjusted for opponent.

Top teamwide QB performance (TAY/P using SRS methodology)

— Bryan Frye (@LaverneusDingle) October 24, 2016

The opponent adjustments are “SRS”-style, which refers to “simple rating system”—it looks at how well one performs against the average performance of opponents… and then how those opponents’ opponents have done and how those opponents’ opponents’ opponents have done and so forth. The iterative ranking system changes less and less each time with each cycle until we get to the final set of rankings.

After adjusting for opponent, the Vikings’ passing offense ranks sixth of all teams for the season.


The Vikings running game was very good. Naturally, the final yards per carry (3.4) looks abysmal but that doesn’t tell the whole story from an evaluative perspective. Zach Line had two short yardage carries, Bradford’s QB kneel at the end counts as a negative rushing attempt and Ronnie Hillman’s six-yard loss really brings down an RB average that’s much closer to 4.3 yards per attempt.

The success rate of the two primary Minnesota backs were 54.5 (Jerick McKinnon) and 50 percent (Matt Asiata). For context, only three running backs with over 100 carries have a success rate above 50 percent and the average running back success rate this year is 45.9 percent. Only one running back has a success rate higher than 54.5 for the year (Ezekiel Elliott).

Both Asiata and McKinnon finished with a decent yards over expectation mark as well. With McKinnon earning 0.1 yards over expected based on down and distance, his effective yards-per-carry comes out to 4.4.

This doesn’t match his actual yards-per-carry of 3.9, but once one accounts for the fact that his three lowest-yardage runs happen to come on with two or fewer yards to go, you can see why his YPC numbers don’t match his actual level of play.

Asiata did even better in yards over expectation with an adjusted YPC of 5.7! Without that big 3rd and 13 run, it drops to 3.4, but the whole point of the metric is to give partial credit for favorable down-and-distances; accounting instead of discounting anomalous runs is a better way of going about things, though this does demonstrate that Asiata on a run-by-run basis seemed to have a rough go of it.

In terms of yards after and before contact, McKinnon did well again. With 2.5 yards after contact per attempt, McKinnon exceeded his 2015 average (2.3) and the NFL average (also 2.3) in this game. For the week, he ranked ninth of the 30 running backs with at least ten carries.

His yards before contact (1.4) was below the NFL average (1.8) and significantly below his own average from last year (2.9). This may have to do with the run blocking in the game and that likely plays a role, but there’s a very good chance that his decisionmaking, patience or agility plays a role.

Asiata averaged only 1.8 yards after contact, relatively low in the NFL. His yards before contact, however, were great—2.8. But if you eliminate his big run, he ends up with 0.4 yards before contact—lower than McKinnon by an enormous amount. Again, Asiata deserves credit for his run but in terms of determining what kind of back he was or what he did to contribute to the offense, it makes sense to look at outliers and see if they materially impact the outputs.


Receiving stats look a lot different when the quarterback himself performs poorly. Once again, we’re looking at yards per route run but the numbers will look pretty bad regardless of the skill level of the receiver in question.

Rank Player Receptions Yards YPRR
25 Cordarrelle Patterson 7 67 1.91
11 Kyle Rudolph 5 55 1.38
47 Adam Thielen 2 52 1.24
80 Stefon Diggs 2 18 0.36
85 Charles Johnson 0 0 0

Those are Week 7 ranks compared to others of the same position, which is why Rudolph has a higher rank with lower yards per route run. Given how the offense has spread the ball between multiple receivers over the last couple of weeks, Diggs’ season totals are significantly lower than they were to begin the season.

Rank Player Receptions Yards YPRR
17 Stefon Diggs 27 390 2.19
26 Adam Thielen 22 324 1.96
14 Kyle Rudolph 26 291 1.61
34 Cordarrelle Patterson 18 158 1.82
98 Charles Johnson 6 92 0.81

Diggs still shows up well, though largely on the strength of his incredible opening to the season. Hopefully he can continue it when healthy. Patterson is moving up the ranks and is performing like a somewhat consistent and reliable contributor. There are now several games in his 2016 portfolio that speak to his ability to be a number two receiver and that’s big for the Vikings.

Pass Defense

One thing that Bryan Frye at Gridfe has done is create adjusted yard per play rankings for both the offense and the defense, so before checking out some of the Week 7 advanced stats, we can look at the aggregate score for the defense for the season. Again, these are adjusted for opponent.

Top defenses against quarterbacks (TAY/P using SRS methodology)

— Bryan Frye (@LaverneusDingle) October 24, 2016

Unsurprisingly, the Vikings rank first. Not only that, they rank first by a significant, healthy margin. This also gives us the opportunity to measure the differential, functionally how good the Vikings quarterbacks play against how well they stop opposing quarterbacks. That by itself is a good proxy for team success, though of course the running game will have an impact as well.

Top teams by quarterback differential (offensive and defensive TAY/P SRS)

— Bryan Frye (@LaverneusDingle) October 24, 2016

The Vikings are first by a significant amount, not that that surprises anyone.

As for this game, Carson Wentz had the worst game of his young career and it shows. It was his lowest ESPN QBR, PFF grade adjusted net yards per attempt and so on.

For individual performances, we’ll start with the defensive line, who had a somewhat mediocre performance given their usual standards within their roles.

Much of this might have to do with a shifting role against Carson Wentz, who is a threat as a runner—though that didn’t mean anything against Cam Newton, where most of the individual linemen posted season-highs in pressure rate.

Only three linemen generated any pressure at all—with Shamar Stephen, Linval Joseph and Brian Robison generating none at all. Everson Griffen was well below his standard at 3.6 percent (his average last year was pass pressure on 14.3 percent of passing plays) and that’s a low number for a defensive tackle much less a defensive end. Even after accounting for how often one pulls back on the pass-rush to account for edge contain, that’s disappointing.

After that is Tom Johnson, who had an excellent game in terms of pressure rate, but he didn’t play too many snaps. Still, 7.7 percent is OK for a defensive tackle and he was one of the only players who could bring pressure to bear on Wentz—important because Wentz’s yards per attempt gets cut in half under pressure.

Only Danielle Hunter produced pressure at an above-average rate—13.3 percent—and only marginally above that.

As for pass coverage, we don’t always get complete data, but the amount of targets a player gets can be a very useful proxy. Here are the things we know about targets and amounts of routes in coverage:

Team Targets Cvg Snaps Tgt%
Xavier Rhodes 5 22 22.7
Terence Newman 3 21 14.3
Harrison Smith 3 27 11.1
Captain Munnerlyn 2 18 11.1
Trae Waynes 1 13 7.7
Jayron Kearse 1 24 4.4
Andrew Sendejo 0 4 0.0
Chad Greenway 6 14 42.9
Eric Kendricks 8 25 32.0
Anthony Barr 0 23 0.0

We also know that Andrew Sendejo has an interception (on a pass he wasn’t targeted on, according to PFF) and that Xavier Rhodes brought in a pick. Rhodes only allowed two completions for 18 yards, or one completion for 28 yards, using the numbers PFF gave us but generally speaking acquitted himself extremely well.

It’s surprising that Barr wasn’t targeted at all, and that’s probably a combination of his assignment and skill level. Greenway was targeted a number of times with Sproles as the primary receiver, but didn’t give up too many completions mostly because of Sproles’ inability to hold on to the ball. Greenway had a very poor day in coverage as a result.

Kendricks allowed seven completions on his eight targets, but kept it to 63 yards, or 7.8 yards per target—honestly not too bad as a linebacker.

Carson Wentz only had 67 passing yards when targeting a defensive back—five completions on 14 attempts. Phenomenal work by the secondary.

Run Defense

The Vikings run defense played out a bit better than the initial numbers show—with Eagles running backs averaging 4.5 yards a carry. But those running backs had a success rate of only 36.3 percent, an abysmal number.

Mathews had a great run that boosted the yards-per-carry total, but most of the runs were failures from the perspective of keeping the offense on schedule. When adding in Carson Wentz’s runs, things don’t improve much, as he only had two runs (and only one successful one).

Sproles was the only runner who had any positive yards above expectation (9.0 per carry), but it was all built on one run of 17 yards, as he had three failures in the running game to go with his successful run. The fact that Sproles had an eight-yard run on third and nine helps his yards per carry but didn’t help the offense.

So the Vikings actually did a very good job defending the running game overall, but made two mistakes. The primary playmakers are the ones who created the most “stops” in the running game—or tackles that constitute losses for the offense. Here are the stops created in the running game from the defenders:

Player Total Sacks Stops Stop Rate
Chad Greenway 2 0 2 100.0%
Everson Griffen 1 0 1 100.0%
Shamar Stephen 1 0 1 100.0%
Harrison Smith 6 0 3 50.0%
Xavier Rhodes 4 0 2 50.0%
Jayron Kearse 2 0 1 50.0%
Eric Kendricks 9 0 3 33.3%
Danielle Hunter 3 0 1 33.3%
Anthony Barr 5 0 1 20.0%
Brian Robison 2 0 0 0.0%
Terence Newman 1 0 0 0.0%
Captain Munnerlyn 1 0 0 0.0%
Linval Joseph 1 0 0 0.0%
Trae Waynes 1 0 0 0.0%
Total 39 0 15 38.5%

Those are of course the percentage of tackles that a defender had a run stop in, not the percentage of running snaps that they were able to stop the offense. Here’s the season-long total from the perspective of total running snaps instead of total tackles:

Player Running Snaps Stops Stop Rate
Danielle Hunter 57 8 14.0%
Linval Joseph 107 11 10.3%
Eric Kendricks 129 13 10.1%
Chad Greenway 55 5 9.1%
Harrison Smith 131 10 7.6%
Tom Johnson 50 3 6.0%
Anthony Barr 131 7 5.3%
Captain Munnerlyn 72 2 2.8%
Shamar Stephen 84 3 3.6%
Andrew Sendejo 105 3 2.9%
Xavier Rhodes 70 2 2.9%
Brian Robison 106 1 0.9%
Everson Griffen 107 2 1.9%
Trae Waynes 68 1 1.5%
Terence Newman 122 1 0.8%

Harrison Smith’s run stop rate is unusually high for a safety, which makes sense given how talented he is. Danielle Hunter’s stop rate is incredibly good, especially for a player designed to be a rotational pass rusher.

Greenway has had a couple of good games before this most recent game where he struggled and in his limited role has been an effective run defender. Tom Johnson is also unusually solid in run defense given what he’s supposed to be doing as a defensive tackle in on nickel downs.

Kendricks is putting up a good rate of run stops, but Barr has been struggling and that’s been reflected in the big runs that have come up through his gap assignments for the past several games. Beyond that, the Vikings need more from the defensive ends—what they get from the cornerbacks is mostly just gravy… though Xavier Rhodes is doing a very good job.

Overall the Vikings should be proud of the defensive performance they put together against the Eagles because it kept the Vikings in it until very late—remember, they only gave up 13 points. The secondary put together one of the best unit performances I’ve seen from a Vikings unit, especially because the other two units weren’t performing to their usual standard.

It’s not unusual (which is an excellent thing to say) to see quarterbacks underperform so much against the Vikings defense. But to see them perform so poorly despite having weak coverage defenders at linebacker and little pressure from the defensive line is pretty extraordinary. Anybody that needs to be convinced of the Pro Bowl potential of the players in this unit simply needs to watch this game.

The post Arif’s deep stats dive: Putting loss to Eagles in perspective appeared first on 1500 ESPN Twin Cities.

Source:: 1500 ESPN Sportswire

Leave a Reply

You can use these HTML tags

<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>