Subscribe to Blog via Email

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

Join 7 other subscribers

MN Vikings Tweets

Bleacher Report – Vikings

Arif’s deep stats dive: Vikings outperforming even optimistic outlooks

By Arif Hasan

The Minnesota Vikings are entering the bye week coming off of two dominant wins that featured a surprising and effective offensive performance to pair with continued profound defensive work and generally good special teams play. We didn’t do a comprehensive statistical review last week of the Vikings’ win against the Giants, but their game against New York and their most recent victory against Houston share so much in common that we may just as well combine the two in our statistical look.

Once again, many of the concepts outlined below were defined in our Week One article.

Team Stats

We’ve been looking at two primary statistics in this column: drive success rate and points per drive. This understands that possessions are a better key unit of analysis for football games more than individual plays (which often have drive-specific strategic goals, like running to set up the pass or creating third-and-manageable) or entire games (which are impacted more by the number of drives than what teams have done “per game” in the past) but it can look a bit different than what people take away from game-level perspectives.

For example, the Vikings dominated the last two weeks in drive success rate—which measures how easily teams create or prevent first downs—both offensively and defensively. Their 0.806 drive success rate offensively would rank first overall this year over the course of a season while the success rate they allowed (0.667) ranks as the ninth-best defensive performance. The net (+0.139) ranks first overall.

That fits all of our perceptions, and the Vikings sit atop a number of power rankings to feed that perception.

But their year-long total is bogged down by their mediocre offensive performances earlier. The defense is excellent—ranking fifth overall (and remember, this measure does not include turnovers, which alone would boost a defense ranking fifth in preventing first downs into first place in terms of total performance).

Their season-long performance offensively ranks 26th overall, at 0.664, which drops their net total to +0.019. Does that mean that the Vikings are the 12th-best team in the NFL? Probably not, because creating and stopping first downs are only part of what teams do on the field, but it does indicate that on a season-long level, the Vikings are swimming among many good teams and not clearly standing out.

But if the last two weeks are better indicators, then the Vikings are far and away the best team in the NFL.

One thing to note: neither the season-long or recent looks account for strength of competition; the Vikings have played some excellent teams in these metrics—they’ve played the eighth, ninth, tenth, 13th and 20th-ranked teams in net drive success rate, per Football Outsiders.

The other drive-specific metric to look at is points per drive. This takes into account the ability of a team to convert field position into field goals as well as red zone opportunities into touchdowns, but is not independent of field position (unlike drive success rate).

The last two weeks have showcased a Vikings team that ranks sixth in the NFL, earning 1.14 more points per drive than their opponents. Offensively, they ranked 11th overall, scoring 2.18 points each drive. Defensively, they ranked second, only allowing 1.05 points per drive.

It might seem odd that a team that looks so good “only” ranks sixth overall, but consider the performances of the teams above them; the Patriots dominated the Browns, Atlanta put up one of the most inspired offensive performances we’ve seen in some time against Carolina, the Cowboys have handily won their last two games and the Eagles have been the Eagles.


The Vikings have seen consistently solid statistical work from Sam Bradford since his debut against Green Bay, but it is possible that despite putting together three elite statistical performances and one very good box score against Carolina, that this last game against Houston was perhaps his best on the film and might be one of his only “elite” all-around performances thus far.

Some advanced statistics recognize this, while others don’t see much reason to distinguish this set of performances from the other two. There’s also one fascinating outlier overall with Bradford that might be interesting to dig into but ultimately is probably meaningless.

His Pro Football Focus scores from the last two weeks somewhat match my casual evaluation of his play—one relatively average game (a 74.5 grade against New York) and one very good game (81.7 against Houston). I think PFF undersells his second performance, but there you go. His passer rating over those two games, 111.5, ranks sixth of all teams in Weeks 4 and 5, and his adjusted net yards per attempt (5.25) ranks seventh among teams in the same time span.

His best mark by far comes from ESPN’s QBR metric, which gave him a pretty good score against the Giants (81.5) and an elite score against the Texans (93.3), which comes out to 86.5 overall—ranking third among all quarterbacks in that two-week stretch.

The outlier comes from implied yards per attempt, which only looks at two things—accuracy rate (completions plus drops over attempts minus spiked balls, throwaways and batted passes) and average depth of target. Because Bradford didn’t throw the ball downfield all that much compared to his competition, his somewhat high accuracy rate doesn’t help him and he ranks 25th among all quarterbacks in implied yards per attempt.

That’s a very surprising result, and I’m guessing some of it stems from the fact that Bradford’s completion rate isn’t too high and that he wasn’t unduly hurt by drops over the course of those two games.

This discrepancy may capture the fact that Bradford isn’t taking as many chances as some other quarterbacks with deep passing and at the same time is only getting a marginal completion rate for his safety, but I’m not sure that matches up with reality.

It could all just be a fluke, of course. Some measures of quarterback play reward certain styles of play and punish others (ANYA seems to punish those who gamble with deep shots a bit too much, for example) and this could be the case here. It does give us some context for his performance though, and may tell us that despite the fact that Bradford is putting up performances commensurate with the top five quarterbacks in the NFL, his true talent could be somewhat below that.

I don’t think the Vikings would mind if they ended up trading for a top ten quarterback instead of a top five quarterback, so the question may ultimately be moot.

His overall scores match his scores from the last two weeks, as his PFF grade ranks him fourth overall, his passer rating places him second among quarterbacks and his ANYA and QBR both rank fifth in the NFL. Only his implied yards per attempt is relatively low, at a dead-average 7.17, which ranks 18th.

As for pass protection, Bradford benefitted from one relatively clean game against the Giants where he was only pressured on 19.5 percent of passing dropbacks (per PFF) but found himself in a somewhat more dire situation against Houston with 37.5 percent of his snaps under pressure. Given that over the course of a season, the best teams in pass protection allow pressure on 20 percent of snaps and the worst on about 45, it seems like Bradford went from relative ease to a war zone.

Against the Giants, Boone, Berger and Fusco didn’t give up a single pressure—meaning some combination of Clemmings, Sirles, McKinnon, Asiata and Bradford himself were responsible for seven total pressures, with likely three of them coming from the left tackle.

While we don’t have all of PFF’s data for the Houston game, we know that Whitney Mercilus earned pressure on 26.1 percent of his snaps, and grabbed a sack or partial sack twice. Almost all of them (save one or two snaps) occurred against Clemmings, which is one of the worst performances any tackle likely put together all week.

It was evidently not as bad as Billy Turner’s, who not only saw PFF’s worst performance at left tackle all week, but got his walking papers from Miami as well.

It is likely that the right side of the line held up very poorly statistically given the fact that Antonio Smith earned pressure on about 26 percent of his snaps as well, but we don’t know how much of it was against Sirles, Kerin, Berger, Fusco, etc. That John Simon had a good game is also not great news for Sirles, but not immediately damning either.


The one receiving statistic we’ll continue to highlight is yards per route run, but it will be useful to look at things like drop rate going forward in order to dissect some of the discrepancies mentioned above.

Stefon Diggs only played in one of the last two games, but because this is a rate statistic, he shouldn’t be too slighted in terms of how it’s measured. It wasn’t a statistically amazing game, however, so he drops significantly against the receivers from the last two weeks. With an average YPRR of 1.25, he ranks 52nd of all receivers and the lowest of all Vikings receivers.

The highest among all Vikings receivers is very surprising: Charles Johnson, who ranked third among all NFL receivers in the past two weeks. The reason? He only ran 17 routes, which is a better indication of what the Vikings think of his talent level than his actual YPRR. Those 70 yards against New York might be easy to forget after a bland performance against Houston, but they do boost his numbers.

As for qualifying receivers who ran a fair number of routes, Adam Thielen ranked 11th in the NFL out of more than 100 qualifying receivers, earning an extremely good 2.75 yards per route run. Cordarrelle Patterson came in with 1.66 yards per route run (ranking 38th), above average among all receivers and very good numbers for the third or fourth option on the depth chart.

Like Diggs, Wright only played one game in the last two weeks with significant action and for his trouble earned 1.39 yards per route run, ranking 46th in the NFL in that time span.

Kyle Rudolph earned 70 yards once more, but instead of it occurring in one game, it happened over two—meaning his YPRR score is pretty low compared to what it was. He ranked 19th of all tight ends in the past two weeks, though was probably hurt by the fact that his routes were less aggressive because of how much blocking help he was forced to provide T.J. Clemmings.

For the year, Diggs still ranks extremely well. He’s fourth in yards per route run with 2.81 yards created every time he goes out into a pass pattern. Adam Thielen is next at 20th, with a mark of 2.14 and Patterson isn’t too far behind at 35th and 1.79 yards per route rune. Jarius Wright is 60th at 1.39 and Charles Johnson finishes with a disappointing 0.82, ranking 78th.

It seems like, based on the snap counts of the last game, that Thielen may have already supplanted Johnson as a starter once Diggs returns, but it will be fascinating to see if that’s actually the case moving forward.

Kyle Rudolph ranks ninth among all tight ends with at least 50 receiving routes run and sixth among those with at least 100 (those that are kicked out once you narrow the list from 50 routes to 100 are all pretty good—Rob Gronkowski, Marcedes Lewis and Jimmy Graham).


The Vikings running game has improved from abysmal to below average, which isn’t great news but does have within it a kernel of hope given how well the running game performs when the game is close.

Consider the fact that over the last two weeks, Jerick McKinnon’s yards over expected (again, “expected” is defined as what the average running back gets in the same down and distance) is -0.6 yards per attempt, or 3.71 adjusted yards per carry. His success rate of 30 percent ranked dead last among running backs over the past two weeks.

This all comes from his yards before contact, which is a measly 0.3 yards—23rd of the 26 running backs who had at least ten carries in the last two weeks. His 1.9 yards after contact, per PFF, isn’t particularly amazing but neither is it as alarming as the inability of either him or the offensive line to prevent contact from arriving immediately.

Matt Asiata’s 2.2 yards before contact over the last two weeks might point to running in more favorable situations late in games, better vision or incidentally better blocking. He hasn’t generated as much after contact (1.5 yards after contact per carry) as you’d expect but his above average run success rate of 47.6 percent fits within his role—even as he earned worse yards against expectation than Asiata, at -1.0 (or, 3.3 adjusted yards per carry).

An interesting note: when the score is within ten points, McKinnon’s yards per carry is 4.34 and Asiata’s is 2.0. Outside of those situations, McKinnon’s yards per carry is 1.4, while Asiata’s is 3.55. I’m not sure how meaningful that is, but it does give us a reason to believe that McKinnon’s low yards-per-carry overall so far isn’t a worry like Peterson’s was because McKinnon is generating yards when the offense needs to score points while Asiata generates successful carries when the offense needs first downs more than points.

Or it could mean nothing.

For the season, not much changes. McKinnon’s success rate (measured somewhat differently by Football Outsiders) is pretty low compared to the rest of the NFL. In fact, it ranks 29th of the 31 qualifying running backs.

Run Defense

Running backs running the ball against the Vikings are almost as bad as the Vikings running game itself. Before the Texans game, there was some talk about the Vikings having a somewhat average run defense. Over the past four weeks, running backs have averaged a 40 percent success rate (interestingly, running backs have had exactly seven successful runs in each of the past four games) and average 0.4 yards per carry below expectation—effectively 3.9 yards per carry.

The most effective running back in the last two weeks was Bobby Rainey, who earned success on three of his four runs. After that, it’s Alfred Blue—a player put in late against the Vikings when the Texans should have passed the ball instead of taking the time to run it out at the end of the game. The primary backs, Orleans Darkwa and Lamar Miller, earned successes on 25% and 37.5% of their runs with -1.41 and -1.78 yards vs. expectation.

This is built on the backs of players like Linval Joseph and Harrison Smith, and it helps that Eric Kendricks produced in a big way last week.

Stop Rate (the rate of tackles that produced an offensive loss), Weeks 4 and 5:

Player Total Sacks Stops Stop Rate
Harrison Smith 15 0 3 20.0%
Captain Munnerlyn 14 0 1 7.1%
Eric Kendricks 13 0 4 30.8%
Terence Newman 4 0 0 0.0%
Andrew Sendejo 10 0 0 0.0%
Anthony Barr 5 0 1 20.0%
Linval Joseph 7 0 5 71.4%
Xavier Rhodes 4 0 0 0.0%
Tom Johnson 4 1 3 75.0%
Danielle Hunter 5 1 3 60.0%
Everson Griffen 3 0 0 0.0%
Shamar Stephen 3 0 1 33.3%
Trae Waynes 1 0 0 0.0%
Brian Robison 2 2 0 0.0%
Chad Greenway 2 0 0 0.0%
Total 92 4 21 22.8%

Linval Joseph has continued to dominate in the run game and stands to be one of the most statistically dominant nose tackles in the past several years of the NFL. Tom Johnson and Danielle Hunter are also putting in several days of good work, while the linebackers seemed to have fallen behind overall, even though Kendricks had a reasonably good game against the Texans.

Kendricks’ effort was felt more in the passing game, of course, but he did some pretty good work in run defense. Still, the best performances came once again from Linval Joseph as well as Tom Johnson and Danielle Hunter.

Ideally, the best way to look at this is not the percentage of tackles that ended up being stops—though that’s a useful dimension of analysis—but rather the percentage of run plays that a player was in on where he produced a stop.

Because most running plays end in a tackle, figuring out how many tackles there are is not a particularly useful way to measure individual run defense efficiency. Figuring out how many of those represent successes for the defense gives us a more complete measure of how important a player is to the run defense of a team.

“More complete” doesn’t mean “absolutely complete,” of course—on specific plays, defenders will be asked to force players into different run lanes or funnel ballcarriers into certain players and so on, and they can be key players without recording a stastistic.

That said, figuring out who is making key tackles to force offenses into unfavorable down and distances will give us an idea of who playmakers are in the run game. Below are the primary defensive linemen, linebackers and safeties playing for the Vikings and their season-long stop rates. Also listed are the average stop rates of players at the same position across the NFL (edge defender, defensive interior, off-ball linebacker and safety) to get a good understanding of how impressive their performance is.

Player Run Stop% Average vs. Average
Danielle Hunter 21.3% 7.6% +13.7%
Linval Joseph 21.3% 9.1% +12.2%
Harrison Smith 9.7% 4.5% +5.2%
Eric Kendricks 12.9% 10.9% +2.0%
Andrew Sendejo 5.8% 4.5% +1.3%
Tom Johnson 10.3% 9.1% +1.2%
Chad Greenway 10.0% 10.9% -0.9%
Shamar Stephen 6.0% 9.1% -3.1%
Anthony Barr 6.8% 10.9% -4.1%
Everson Griffen 2.5% 7.6% -5.1%
Brian Robison 2.4% 7.6% -5.2%

One could even turn their performance against other players at the same position into baseball-style “plus” grades, where one gets a score of 100 for performing like an average player at their position and has 15 points added or subtracted for every standard deviation above or below their performance is against their peers.

So a player with a score of 115 would be creating run stops at one standard deviation better than other players at the same position—better than 84 percent of same-position players. A player with a score of 130 would be rare, and better than 97.7 percent of players at the same position.

Player Grade
Danielle Hunter 171.3
Linval Joseph 145.2
Harrison Smith 126.6
Eric Kendricks 107.7
Andrew Sendejo 106.6
Tom Johnson 104.5
Chad Greenway 96.5
Shamar Stephen 88.6
Anthony Barr 84.1
Everson Griffen 73.3
Brian Robison 72.7

Because Hunter has only been in on 47 running snaps, his score is pretty inflated with his 10 run stops. With three more run snaps and no stops, he’d drop to 164.6. One also has to take into account the fact that some players rotate in against certain personnel packages and down-and-distance situations and that may precipitate something like more outside runs—which would give an end like Hunter some advantage.

Regardless, that’s a pretty fantastic score.

While Barr, Griffen and Robison have disappointing scores, all three of them have likely performed their duties in the run game by setting the edge or filling lanes, and neither Griffen nor Robison miss many tackles, either.

One other important measure in the run game should be figuring out missed tackles. Using Football Outsiders’ charting project, we can determine which players on the defense arrived in time but didn’t finish the job.

Player Pos Missed Tackles Tackles MT Rate Score
Harrison Smith S 1 26 3.7% 122.3
Captain Munnerlyn CB 1 17 5.6% 121.1
Trae Waynes CB 1 17 5.6% 121.1
Terence Newman CB 2 14 12.5% 108.3
Andrew Sendejo S 3 20 13.0% 106.8
Brian Robison Edge 1 6 14.3% 106.7
Linval Joseph DI 2 11 15.4% 103.2
Everson Griffen Edge 2 9 18.2% 99.3
Eric Kendricks LB 6 19 24.0% 89.4
Anthony Barr LB 3 8 27.3% 83.7
Danielle Hunter Edge 4 10 28.6% 79.8

Go Harry! Hunter may be making stops, but he’s also missing a lot of his opportunities. For a young player, I’d say the balance of his play indicates a positive development curve, but he still has a lot to learn.

Pass Pressure

The Vikings are tied for the lead in sacks, but that only tells part of the story. Putting pressure on the quarterback has been a key part of the Vikings’ success against a variety of passing styles and talent. In the last two weeks, their pressure rate has been low, but by forcing Eli Manning and Brock Osweiler to get rid of the ball right away to underneath routes, they effectively achieved the same results.

Here are the pressure rates from players in the last two weeks along the defensive line:

Player Pressure Rate
Brian Robison 13.3
Danielle Hunter 11.7
Everson Griffen 11.3
Tom Johnson 10.9
Linval Joseph 6.5
Shamar Stephen 0
Justin Trattou 0

And for the year:

Player Pressure Rate
Danielle Hunter 13.2
Everson Griffen 12.0
Tom Johnson 11.3
Brian Robison 10.7
Sharrif Floyd 7.7
Linval Joseph 7.1
Shamar Stephen 2.5
Justin Trattou 0.0

Situational pass rushers almost always have significantly higher pressure rates than every-down defensive ends, so Griffen’s 12.0 rate is actually more impressive than Hunter’s 13.2 rate but both are obviously pretty good.

So far the average for defensive interior players in pressure rate is 6.5 percent, while the average for edge players is 10.4 percent.

Tom Johnson is a monster.

Pass Coverage

Weeks 4 and 5 saw a significant change in the number of targets defensive backs saw. In the first several weeks, Trae Waynes saw significant targets whenever he was on the field. In the last two weeks, Waynes only saw targets on 8.1 percent of his coverage snaps. On the other hand, Xavier Rhodes saw the most targets as he shadowed Odell Beckham and DeAndre Hopkins, seeing the ball thrown his way 19.7 percent of his coverage snaps.

It went well:

When Xavier Rhodes has been targeted this season, he’s allowed an NFL passer rating of 0.0. 3 for 10 for 23 yards, 0 TD, 1 INT and 1 PD.

— Nathan Jahnke (@PFF_NateJahnke) October 4, 2016

After that, the coverage target list went Munnerlyn (18.8 percent), Newman (16.5 percent), Sendejo (3.4 percent) and Smith (1.4 percent).

It’s difficult to get game-by-game totals for yards per coverage snap and so on, but knowing that number one receivers—Odell Beckham and DeAndre Hopkins, in this case—did terribly against the Vikings tells us a lot about Rhodes, though he should be held accountable to the ultimately meaningless touchdown at the end of the game if only because it did demonstrate a breakdown of technique.

Here are the year-to-date totals in yards given up per coverage snap, available via Football Outsiders.

Rank Player Targets Targets per Snap Yards per Snap
17 Terence Newman 29 0.142 0.426
29 Captain Munnerlyn 20 0.107 0.492
87 Xavier Rhodes 19 0.176 1.020
107 Trae Waynes 32 0.198 1.442

Newman and Munnerlyn have been underrated and incredible, while Rhodes has probably been even better—though garbage time statistics against him from Hopkins really hurts his overall statistical score.

I don’t think anyone is too upset with Rhodes’ performance thus far and no one would rank him 87th, but it is useful to remember how these statistics were created. Without that garbage time, he would rank 33rd of all cornerbacks, allowing 0.5 yards per snap in coverage (with only 14 targets, or .143 targets per snap in coverage—the same as Newman all while shadowing the opponents’ top receiver).

Overall, this paints a picture of a team that has a few weaknesses—the offensive line and running game could use work—but is becoming a much more complete team than even most optimists could have envisioned.

The post Arif’s deep stats dive: Vikings outperforming even optimistic outlooks 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>