The Wild World of the 2022 Minnesota Vikings

Vikings WR Justin Jefferson and Adam Thielen
Vikings WR Justin Jefferson and Adam Thielen
Photo: USA Today Sports Images

NFL Week 13 - It's Trading Places Week at the top of the Football Outsiders DVOA ratings, with a number of teams swapping spots while the team(s) in between them stay the same.

Let's start near the top with Baltimore and Dallas. The Cowboys crushed the Colts on Sunday night and move up to No. 2 in DVOA. The Cowboys are also now No. 1 in weighted DVOA, which lowers the strength of the earliest games to try to get a better idea of how good teams will be going forward. The Cowboys are now the No. 1 defense by five percentage points over the rest of the league, with San Francisco climbing to No. 2. We'll get to the 49ers in a moment.

Baltimore slid past Denver 10-9 but the Broncos significantly outplayed the Ravens in that game. Baltimore ended up with just 4.7% Post-Game Win Expectancy. The Broncos outgained the Ravens 5.2 to 4.1 yards per play and had a +2 turnover margin. So the Ravens drop in DVOA despite the win, two spots down to No. 4 where the Cowboys used to be.

The Philadelphia Eagles remain in between at No. 3 although they climbed up from 26.9% to 30.0% with a big win over the Titans.

Next swap: San Francisco and Kansas City. If you remember that Cincinnati was just one spot behind the Chiefs last week, you probably aren't surprised that the Bengals' win moved them past Kansas City. However, San Francisco leaped both of them with a big win over Miami, especially by shutting down the Dolphins offense. So we now have San Francisco fifth and Kansas City seventh, with Cincinnati remaining in between at No. 6.

Final swap: How about those red hot Detroit Lions? The Lions smashed Jacksonville and climb three spots to No. 13. Over the past six weeks, Detroit ranks seventh in overall DVOA including ninth on both offense and defense plus sixth on special teams. The Lions swap spots with the Tennessee Titans, who lost big to Philadelphia as mentioned above. They drop down to 16th but are still clear favorites to win the AFC South with the rest of the division also losing this week. Jacksonville plummets six spots from 18th to 24th in DVOA after the Lions' win.

Cleveland (No. 14) and Green Bay (No. 15) drop slightly but stay in the same places between the Lions and the Titans.

Speaking of the Lions, you may have seen that the 5-7 Lions are favored in this week's game against the 10-2 Vikings. Clearly, Vegas oddsmakers are smoking whatever DVOA is smoking when it comes to the crazy year the Vikings are having. Honestly, it doesn't matter what stat you look at, there's never been a 10-2 team anything like this year's Vikings. I've written about them so much but things just get crazier and crazier as they reel off close win after close win.

Check out the list of the lowest 10-2 teams in DVOA history and you can see just how far away the Vikings are from every other 10-2 team of the last 42 years except themselves in 2000.

Worst 10-2 Teams by DVOA, 1981-2022
Year Team DVOA Rk WEI
DVOA
Rk
2022 MIN -5.5% 20 -6.8% 21
2000 MIN 2.4% 19 3.4% 17
1986 NYJ 9.6% 9 9.6% 10
2003 NE 10.5% 8 14.3% 7
1998 ATL 11.4% 10 16.7% 7
1999 IND 11.5% 10 13.3% 10
2001 PIT 14.1% 8 18.3% 4
1991 BUF 14.2% 7 14.3% 8
2000 OAK 14.5% 9 16.7% 6
2016 OAK 15.5% 5 17.7% 4
1986 NYG 15.6% 8 16.1% 8
1989 DEN 15.9% 7 15.4% 7

I added weighted DVOA to the table so you could see just how little these teams are similar to this year's Vikings. Many of the worst 10-2 teams were on the upswing, getting hot in the second half of the season (1998 Falcons) or rebounding from particularly bad performances very early in the season (2003 Patriots). But the 2022 Vikings are actually... well, not quite on the downswing, but three of their four best games of the season came back in Weeks 1-6 before their bye. (The Jets game this week is the fourth of those four games.)

2022 Vikings DVOA

I thought we might end up in a situation where the Vikings were not just the worst 10-2 team in DVOA history but also worse than all the 9-3 teams. That doesn't turn out to be the case because the Vikings climbed a bit with 18.9% DVOA in the game against the Jets. So the Vikings do have one 9-3 team that comes in below them, the 2020 Cleveland Browns. Check out this list of the worst 9-3 or 10-2 teams, and how they finished the season:

Worst 9-3 or 10-2 Teams by DVOA, 1981-2022
Year Team W-L DVOA Rk Final
W-L
Playoffs
2020 CLE 9-3 -6.8% 23 11-5 Lost Div
2022 MIN 10-2 -5.5% 20 -- --
2004 ATL 9-3 -1.4% 14 11-5 Lost CCG
2010 CHI 9-3 -1.0% 17 11-5 Lost CCG
2014 ARI 9-3 1.8% 16 11-5 Lost WC
2000 MIN 10-2 2.4% 19 11-5 Lost CCG
1999 TEN 9-3 3.5% 18 13-3 Lost SB
2019 BUF 9-3 3.7% 11 10-6 Lost WC
2009 CIN 9-3 4.0% 15 10-6 Lost WC
2001 CHI 9-3 4.0% 14 13-3 Lost Div
2019 GB 9-3 4.7% 9 13-3 Lost CCG
2003 STL 9-3 5.4% 13 12-4 Lost Div

It's interesting that none of the teams on this list come from any earlier than 1999. I'm not sure why that is. There certainly were plenty of overperforming teams in the 80s and 90s, but the most extreme ones just happened to be 8-4 or 7-5 instead of 9-3.

The Vikings also stand out among history's 10-2 teams if we use conventional stats. For example, only six teams in NFL history have had negative net yardage with a 10-2 record. The Vikings, at -754 yards, are the most negative of those teams. The 2010 Patriots were at -445 yards; the other 10-2 teams that had net negative yardage were the 1984 Seahawks, 1965 Browns, 1986 Jets, and 1960 Eagles. The Vikings also have the lowest point differential of any 10-2 team in history at +10 points. The 2019 Seahawks, at +36, were the only other 10-2 team below +40.

And here's what's even more remarkable... the Vikings are probably going to keep winning. The Vikings now have the No. 30 remaining schedule in the NFL by average DVOA of opponent, with no games left against the top dozen teams in the league. DVOA will favor their opponents when they travel to Detroit this Sunday and Green Bay in Week 17, but those are definitely winnable games for an average team. Even if Vegas does have the Lions favored this week!

Introducing Game Variance

What's the best way to measure a team's consistency?

Right now, we measure consistency at Football Outsiders with variance. This is a basic statistical formula that's similar to standard deviation. It measures the average of squared deviations from the mean of a sample. But does this match the way we tend to think about a team's consistency as fans?

Let me throw things back to last season for a moment. Here's our week-to-week DVOA graph for last year's New York Giants. The Giants ranked 23rd in variance last season.

2021 Giants DVOA

This does look like a high variance team at first. Look at the big changes between Week 5 and Week 7, for example! But look a little closer and things appear more consistent. For the first part of the season, the Giants were mostly an average team overall. Then Daniel Jones got hurt in the Week 12 game against the Eagles and the team collapsed. The first few games are mostly grouped together. The last few games are mostly grouped together.

A team that the public tends to think of as inconsistent would have a graph that pinged up and down and up and down all season long. When we think of inconsistent, we tend to think of teams that play very differently from week to week. We think of something more like what the Giants did in Weeks 6 and 7 of last year, where they lost 38-11 to the Rams and then beat the Panthers 25-3. That's inconsistent. The Giants come out with a high variance because they had a lot of average games and a lot of very, very bad games and most of those games were a significant distance from their mean performance which was around -30%. But for the most part, they weren't constantly alternating between good games and bad games.

Compare that to, say, this year's Atlanta Falcons:

2022 Falcons DVOA

The Falcons have been up and down all year long. They have seven games below 0% DVOA but only twice have they had two of them in a row. By variance the way we do it now, the 2021 Giants (15.2%) were a much more inconsistent team than this year's Falcons (7.9%). But when you look at the graphics... the Falcons kind of look the less consistent team, right?

So I've been playing around with an idea I call "game variance." It's a different way to measure consistency that concentrates on how the team's performance changes from week to week without looking at the team's average for the entire season. It's very simple. Game variance takes the changes in single-game DVOA from each week to the next week and then averages the absolute values of those changes.

The 2021 Giants had a game variance of 35.5%, which ranked fifth in the league last year. By this measure, they were one of the most consistent teams from week to week. The 2022 Falcons have a game variance of 45.3%, which currently ranks 24th. So by game variance, the Falcons are the more inconsistent of the two teams.

On the extremes, both methods tend to pick out the same teams as the most or least consistent. The two most consistent teams in the league this year, by both measures, are Washington and the New York Giants. The most inconsistent team in the league, by both measures, is Carolina. But in between, these are some differences. Tampa Bay, Tennessee, and Denver have all been more consistent by game variance. Green Bay and Arizona have been more consistent by standard variance. Here's a look at all 32 teams, including offense and defense split out.

TEAM TOT
GVAR
RK TOT
VAR
RK OFF
GVAR
RK OFF
VAR
RK DEF
GVAR
RK DEF
VAR
RK
NYG 15.5% 1 4.9% 2 22.5% 8 4.3% 8 24.0% 18 2.8% 7
WAS 15.5% 2 4.1% 1 16.8% 3 1.6% 2 16.6% 5 2.2% 2
SEA 20.4% 3 5.0% 3 22.8% 9 3.7% 7 20.0% 7 4.9% 22
BAL 22.9% 4 7.5% 7 22.1% 6 6.6% 22 21.3% 10 4.0% 15
TB 27.8% 5 9.9% 14 31.3% 24 5.0% 15 20.8% 9 7.6% 29
DEN 29.7% 6 10.9% 17 15.4% 2 1.4% 1 27.2% 24 4.4% 19
LAC 30.0% 7 8.4% 10 22.3% 7 4.3% 9 16.8% 6 3.9% 14
TEN 30.0% 8 10.8% 16 22.9% 10 4.6% 10 11.5% 1 2.4% 4
HOU 30.1% 9 6.1% 4 28.5% 18 7.7% 24 23.8% 16 4.7% 20
LAR 30.3% 10 6.5% 5 28.8% 19 4.9% 14 23.0% 14 3.6% 10
MIA 33.9% 11 8.2% 9 22.9% 11 7.5% 23 28.5% 27 5.3% 24
MIN 34.1% 12 7.1% 6 25.3% 13 3.6% 6 22.4% 13 2.3% 3
LV 34.6% 13 13.7% 21 33.6% 25 10.4% 29 12.7% 2 2.2% 1
DET 35.3% 14 14.8% 28 24.8% 12 4.7% 11 29.3% 28 7.6% 30
BUF 35.6% 15 9.7% 12 19.7% 5 3.0% 5 25.2% 19 4.2% 18
NO 39.2% 16 9.2% 11 30.4% 21 6.5% 21 20.2% 8 3.8% 13
PHI 41.6% 17 10.8% 15 27.3% 16 4.9% 13 25.3% 20 4.7% 21
KC 41.9% 18 15.1% 29 39.7% 29 11.2% 31 23.9% 17 3.7% 11
NE 42.9% 19 11.3% 18 29.7% 20 5.6% 16 26.8% 23 6.7% 25
CLE 43.1% 20 14.6% 25 39.6% 28 7.8% 25 30.2% 29 7.2% 26
JAX 44.0% 21 14.3% 24 28.4% 17 5.9% 17 31.0% 30 7.3% 28
CIN 44.4% 22 16.6% 30 30.7% 22 9.9% 28 23.3% 15 2.8% 6
IND 45.1% 23 14.6% 26 43.2% 31 10.7% 30 28.0% 26 4.1% 17
ATL 45.3% 24 7.9% 8 18.1% 4 2.0% 4 26.6% 22 3.1% 9
GB 46.9% 25 9.8% 13 36.6% 27 6.4% 20 16.4% 3 2.9% 8
CHI 46.9% 26 12.0% 20 25.6% 14 6.2% 19 31.6% 31 7.3% 27
SF 47.8% 27 19.0% 31 35.7% 26 9.3% 27 21.4% 11 8.9% 31
DAL 48.2% 28 14.8% 27 25.9% 15 6.1% 18 25.8% 21 3.7% 12
PIT 48.6% 29 13.8% 22 13.6% 1 1.7% 3 40.5% 32 9.1% 32
ARI 48.7% 30 11.5% 19 31.0% 23 4.8% 12 27.7% 25 5.1% 23
NYJ 48.7% 31 14.0% 23 41.4% 30 8.2% 26 16.5% 4 2.8% 5
CAR 62.0% 32 19.7% 32 50.8% 32 12.8% 32 21.4% 12 4.1% 16

My first thought looking at this table is that I talk a lot about Josh Allen being inconsistent but based on these numbers he's actually been quite consistent this season. And how about the difference between the Pittsburgh offense and the Pittsburgh defense?

I'm curious to hear in the comments what the readers think about the two forms of variance. Which one do you think is a better measure of a team's consistency? Would you like to see us switch from VAR to GVAR in the future?

* * * * *

Football Outsiders playoff odds, snap counts, and the FO+ database are now all updated through Week 13.

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, Derrik Klassen's All-32 game preview column, and picks against the spread (now including projected totals for over/unders).

*New: It's not a paywall, except on Mondays! We now post DVOA data on Monday instead of waiting until Tuesday as in the past. If you want to see data from the current season on Monday, including all of Sunday's games, you will need to be an FO+ subscriber. On Tuesday morning when we post the update with Monday Night Football added in, all of the free stats pages become free again. The exception is snap counts, which are available to everyone Monday.

* * * * *

This is the Football Outsiders Top 16 through 13 weeks of 2022, 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). Click here for the full table.

OFFENSE and DEFENSE DVOA are adjusted for opponent and performance indoors and consider all fumbles, kept or lost, as equal value. 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.

WEIGHTED DVOA represents an attempt to figure out how a team is playing right now, as opposed to over the season as a whole, by making recent games more important than earlier games.

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>

RK TEAM TOTAL
DVOA
LAST
WEEK
WEI.
DVOA
RK W-L OFF.
DVOA
OFF.
RANK
DEF.
DVOA
DEF.
RANK
S.T.
DVOA
S.T.
RANK
1 BUF 32.8% 1 29.6% 3 9-3 16.1% 4 -13.0% 4 3.6% 3
2 DAL 30.7% 4 33.5% 1 9-3 7.4% 11 -20.1% 1 3.2% 5
3 PHI 30.0% 3 30.0% 2 11-1 19.4% 3 -11.7% 6 -1.2% 21
4 BAL 23.8% 2 21.7% 6 8-4 11.6% 6 -7.0% 9 5.3% 1
5 SF 19.7% 7 24.0% 4 8-4 4.8% 14 -15.1% 2 -0.2% 17
6 CIN 18.1% 6 22.0% 5 8-4 15.0% 5 -4.5% 11 -1.5% 24
7 KC 17.7% 5 20.8% 7 9-3 24.9% 1 7.0% 25 -0.2% 16
8 MIA 14.1% 8 16.2% 8 8-4 20.1% 2 1.1% 17 -4.9% 32
9 NYJ 9.5% 9 11.2% 9 7-5 -2.9% 19 -12.5% 5 -0.1% 14
10 SEA 8.9% 12 10.2% 10 7-5 9.6% 8 4.4% 20 3.7% 2
11 TB 7.7% 10 4.0% 13 6-6 1.5% 16 -8.4% 8 -2.3% 29
12 NE 7.1% 11 6.3% 12 6-6 -8.8% 24 -14.8% 3 1.1% 10
13 DET 6.2% 16 8.7% 11 5-7 8.8% 9 3.9% 19 1.2% 9
14 CLE 1.6% 14 3.0% 14 5-7 10.6% 7 8.9% 27 -0.2% 15
15 GB 0.9% 15 1.9% 15 5-8 8.7% 10 4.7% 22 -3.1% 30
16 TEN -1.5% 13 0.4% 16 7-5 -3.6% 20 -2.6% 12 -0.5% 18

Comments

87 comments, Last at 08 Dec 2022, 12:27pm

#1 by Jetspete // Dec 06, 2022 - 2:13pm

The 1986 jets were my introduction to Football. It’s worth noting they started 10-1 and game 12 was a Monday night destruction at the hands of Marino, 45-3. They finished 10-6 and lost what every jet fan refers to as the Gastineau game in Cleveland in the divisional round. 

Points: 1

#2 by BlueStarDude // Dec 06, 2022 - 2:38pm

RE: "They have seven games below 0% DVOA but until the last two weeks, they had not had two of them in a row" — The graph shows weeks 7 and 8 were both below 0%.

Points: 2

#3 by Aaron Brooks G… // Dec 06, 2022 - 3:07pm

Looks like you don't want to be at either end of the variance extreme -- lot of bad teams clustered there.

Points: 0

#71 by Dales // Dec 07, 2022 - 4:22pm

Feels logical that it is easier to be consistently bad than consistently good, and also that it would be hard to be a very good team that is throwing out clunkers *often*; if they were, they wouldn't be recognized as a good team and unlikely to be a statistically good team.

Points: 0

#4 by zenbitz // Dec 06, 2022 - 3:29pm

It's interesting that none of the teams on this list come from any earlier than 1999. I'm not sure why that is.

 

I would guess that it's going to 4x4 divisions in 2002, smaller division higher variance in opponents, more opportunity to have outlier results.

Points: 4

#5 by Aaron Brooks G… // Dec 06, 2022 - 3:41pm

Now I'm curious what the easiest schedule in the DVOA era was.

Points: 0

#43 by Aaron Schatz // Dec 07, 2022 - 10:28am

Easiest schedules since 1981

1999 STL -15.5%

1991 BUF -15.5%

1982 CIN -13.6%

1991 MIA -11.5%

1998 ARI -11.1%

2010 STL -10.8%

1999 JAX -10.7%

1991 NYJ -10.5%

Points: 2

#63 by tjb // Dec 07, 2022 - 1:57pm

Indeed, it sucked.  The rest of the division, by DVOA:

MIA: -2.4% (19th)

NYJ: -6.2% (20th)

NE: -34.3% (26th)

IND: -49.2 (28th)

(There were only 28 teams)

Points: 1

#6 by Mostly Anonymous // Dec 06, 2022 - 3:46pm

I'm curious to hear in the comments what the readers think about the two forms of variance. Which one do you think is a better measure of a team's consistency? 

My 2 cents is that you should compare GVAR to GVAR for each of the three phases of DVOA (offense, defense, and special teams).  That way, being consistent (or inconsistent) in any of the three phases of the game will stand out in addition to seeing the overall impact on a combined metric.  It may even lead to weighting the three individual GVAR components into a better aggregate GVAR, similar to the way DVOA weights them differently.  Thinking back to stats in college, VAR(aX + bY) was something like a^2 * VAR(X) + b^2 * VAR(Y) + 2ab * COV(X, Y).  I forgot what it would be for three random variables.  Anyway, it would make for a fun off season project.  Thanks for listening and having a great website.

Points: 0

#7 by owleabf // Dec 06, 2022 - 3:50pm

Is there any thought to DVOA possibly missing something in the Vikings result? Or is the thought really that the Vikings have just hit an incredible lucky streak?

They have a 10-2 record against DVOA's 3rd hardest schedule and DVOA keeps insisting they're in the bottom third of the league. 

The model is essentially an EPA/play model as I understand it, I'd wonder if there's any room for things like adjusting it's measures for time left and/or current score, win probability in other words. The Vikings have been winning in high leverage situations and I'm pretty sure that doesn't get valued in the system.

Points: 0

#8 by BlueStarDude // Dec 06, 2022 - 4:06pm

I think they are 9-0 in one score games. There have been multiple games, like this past week's Jets game, where the Vikings really should have lost. Fluke things happen—an inch short here, an inexplicable fumble there—and that really seems to be the simple reason for the discrepancy between DVOA ranking and wins.

Having watched them 3 weeks in a row, they seem average (except in that Dallas game)—which is not so far off from what their actual DVOA says. They sure as heck don't look like a 10-2 team. They do seem to have the pieces in place that they should be better, and I don't know but I guess you lay that at Cousins's feet—Will or someone could say better.

 

Points: 1

#15 by serutan // Dec 06, 2022 - 4:50pm

There will be regression in those one score games next year if not this.

Points: 0

#19 by KnotMe // Dec 06, 2022 - 5:25pm

Yeah, although it really depends on how you get to a one score game.  If you get to 1 score bc your opponent scored a last second TD to get within one score...it actually wasn't that close bc all you had to do was run the clock competently. If you get something like the Jets game where the Jet's had chances but couldn't punch it in..those are the types of things that tend to regress as even the best defense won't get 100% stops in the red zone. Another example would have been if the Saints had been competent to sustain pretty much anything after the Buc's first TD.  It would have been a 1 score game, but not actually all that close. (I've always thought we should differentiate between 1 (FG) score and 1 (TD) score games as they are both a single score but have a pretty big difference in difficulty). 

Points: 0

#25 by bruddog // Dec 06, 2022 - 8:23pm

The one area the Vikings are doing extremely well in is turnover differential at +8. +3 on interceptions and +5 on lost fumbles. They are also +15 on first downs by penalty. That probably explains a few wins at least. 

Points: 0

#74 by laurin // Dec 07, 2022 - 4:48pm

and even their turnover differential can be traced back to luck. they have recovered an insane 16 of 25 fumbles in their games, the general recovery rate is 50% or in this case about 12 to 13 fumbles. and still, it gets even worse once you adjust for the type of fumble, which brings the Vikings' "expected recovery rate" down to 45.6% or about 11.4 fumble recoveries

Points: 0

#88 by Xexyz // Dec 08, 2022 - 12:15pm

Is that really "insane", though?  Yes the general recovery rate is 50% but you'd expect being within 2-3 fumbles of that number to be normal with a small sample size of 25.  Yes they've been a little lucky with fumble recoveries but not so much I'd call it a huge outlier.  Even if you regress them back three fumbles they're still +5 on the year which would drop them from 3rd down to 5th in turnover differential.  

Points: 0

#33 by Mike B. In Va // Dec 07, 2022 - 7:06am

They're the inverse of a team like the 2021 Bills, who lost *all* of their one-score games despite finishing second in DVOA. Minnesota's luck this year has been insane.

Points: 1

#59 by Jblades13 // Dec 07, 2022 - 1:21pm

It's kind of lazy to say that the VIkings should have lost a game like the jets game, considering their win probabiliy was greater than 70% for 56 minutes of game time and was only below 50% for one minute, that doesn't come across as a game a team "should have lost"

The Vikings have played 4 games in a row against top 5 defenses by DVOA and outside of one terrible game against the Cowboys their offense has done really well. The only game where the offense was poor was the worst pass blocking performance by any team this season, so I'm not sure what you're saying should be laid at Cousins' feet, but I don't think any QB could do well when pressured on over 60% of snaps. 

Points: 0

#9 by Will Allen // Dec 06, 2022 - 4:17pm

My eyeballs tell me they stink on defense,  have a middling qb, a great receiver, some other good ball handlers, an average at best o-line, and lousy special teams. They have reminded me all year of the 2000 Vikings. That team completely collapsed defensively, and lost a conference championship by 40-plus.

Points: 0

#13 by Romodini // Dec 06, 2022 - 4:44pm

The deprecation of this Vikings team by their fans is one of the funnier things on FO this year. I kind of hope the Vikings keep winning so it continues. Assuming the Cowboys do their usual choke job, my hope is for the Vikings to end the Eagles' season in the conference game (vengeance for 2017) and meet the Bills in the Super Bowl so at least one long suffering franchise finally gets to win one. 

 

Points: 2

#21 by Moridin // Dec 06, 2022 - 6:09pm

heh, I had a comment a few weeks ago that I could see the Bills vs Vikings in the SB, and that would lead to the only tie in SB history when the stadium loses power and then has to be evacuated later due to tornadoes (after getting to 3rd overtime).

I've really enjoyed the Vikings when I look backwards, but its a constant state of "they are finally about to screw this up instead of lucky event X saving their ass this time" in the late 3rd and 4th Q and then it doesn't happen. They've def done the full gamut of close game wins (pulling ahead at the last second, back and forth and being the last to score, staving off a strong GWD opportunity, being ahead most of the game and looking fairly good and letting the opponent get close/pull ahead and having to have one last drive, miracle catches to sustain drives, etc). Its hard to expect that to continue every week.

Points: 2

#14 by Aaron Brooks G… // Dec 06, 2022 - 4:45pm

It's amusing to go back to the peak-Bears.

Those 1985-1986 teams only lost four games, but they tended to get their asses kicked when they did lose. 

\then there was the 1940 Redskins...

Points: 0

#16 by jmaron // Dec 06, 2022 - 4:59pm

I don't really see that comparison.  I don't think this team is as soft on defence, nor as talented on offence. This version is 18th in both offence and defence and the 2000 team was 5th in offence and 31st in def. As soft as the defence is, they do get a lot of turnovers late in the game and from listening to the def coordinator and players they are actively trying to take the ball away, particularly in the red zone. Last year it seemed the Vikings allowed teams to score at the end of every 2nd and 4th quarter. This year they have stopped teams quite often. Probably a decent amount of luck, but I doubt it's all luck.

They are also much more disciplined than their opponents. They don't turn it over as often, and they don't take as many pre snap penalties or personal fouls. By my biased eye, they also don't seem nearly as grabby in the secondary as their opponents. I think the coaching staff has done a good job with a team that isn't very talented. They have a net gain of 15 first downs by penalty and +150 yards, vs -9 1st downs and -257 yards last year. 

The one thing I agree with you about, is I don't think this team has any more chance to win the SB than the 2000 team, a massive shellacking would seem a decently high probability against a strong team. 

 

 

 

 

 

Points: 1

#20 by Will Allen // Dec 06, 2022 - 5:50pm

That 2000 team went into the abyss defensively over the last 6 weeks of the regular season, so there's still time. The 2000 team did block better. I think it rather unlikely that their improved red zone defensive performance this year, compared to last, is due to trying harder to get turnovers. 

Points: 0

#26 by vikingflagship // Dec 06, 2022 - 10:30pm

the big difference I see in secondary is emphasis on playing the ball / turnovers. 

they play deeper. 

Zimmer defence prioritised covering the man, and forcing incompletions....and with the right guys (2017), he obtained super results for yardage, points scored, 3rd down conversions....

this incarnation plays softer with more emphasis on the ball, probably as they cannot really cover, hence they have more interceptions.

the timeliness is mostly luck. i say mostly as they are really terrible against play action passes that are less an issue in two minute drills. 

Points: 0

#29 by Will Allen // Dec 07, 2022 - 12:46am

"Cannot really cover" only works with a large amount of luck. Hunter really hasn't returned to form, but if he goes down again, and/or Patrick Peterson gets banged up again, mediocre offenses will be hanging 40 on them in no time. As it is, I'll be surprised if the Lions don't get to thr mid 30s on Sunday

Points: 2

#18 by jmaron // Dec 06, 2022 - 5:21pm

I would put the odds of Vikings being as good as an avg 9-3 to 10-2 team at about 5%

As good as a typical 8-4 team - about 20%

As good as a 5-7 to 7-5 team at about 65%

and worse than typical 4-8 or worse team at about 10%

Points: 0

#23 by AFCNFCBowl // Dec 06, 2022 - 7:17pm

I wouldn't go that far but they are certainly fool's gold.

They have already proven they don't belong in the same league as the elite teams of the league (crushed by DAL and PHI, should have been crushed by BUF based on DVOA splits).

SF is too strong on the D-line for Cousins to handle

They will likely be the #2 seed so they should beat their wild card opponent (likely SEA, WAS or NYG) but beyond that they'll need the same fluke BS they got in almost every win this season except Week 1 vs GB.

Points: 0

#61 by Jblades13 // Dec 07, 2022 - 1:26pm

Saying they're garbage and just getting lucky is about as lazy of a take as you can make. 

If the Vikings had a 50% chance of winning each of these one score games, the odds of them going 9-0 is 0.2%, improbable things certainly happen, but I'm unwilling to bet that something that improbable is actually occurring and when you consider that the Vikings have played one of the harder schedules in the league, they are definitely an above average team. 

Points: 0

#64 by Pat // Dec 07, 2022 - 2:04pm

If the Vikings had a 50% chance of winning each of these one score games, the odds of them going 9-0 is 0.2%,

That's because point probabilities are tiny. The only way they end up going 9-0 even 10% of the time putting all games the same is if they had a 78% chance of winning each game. That corresponds to being favored by a touchdown in all of their games.

Unlikely things happen all the time because there are tons of possible unlikely things.

Points: 0

#32 by David // Dec 07, 2022 - 4:58am

The Vikings have been winning in high leverage situations and I'm pretty sure that doesn't get valued in the system.

It doesn't.  Because it isn't repeatable.  Because if it was, you'd just win in all situations - if you're good, why wait until 3rd down?

Or is the thought really that the Vikings have just hit an incredible lucky streak?

Yes - that's exactly what the model is telling us - so it's not really a 'thought' - the model doesn't think - it's an observation.  The observation, lengthened out, goes something like this:

Football, like almost every sport, is a set of weighted random events.  The events themselves have randomness built in (sometimes, players just drop a pass, or fall over, or whatever) but there is skill inherent to the events, so the events themselves are weighted.  The model is trying to boil out the randomness in order to describe the true skill level.  There are a few things that the DVOA model does to try to accomplish this. 

The first, and most important, is that the model understands that 17 samples (games) isn't enough to boil out the randomness, so it uses events (plays) with a much larger sample size.  Understanding whether a game is successful or not is easy - everybody (apart from the Texans, sorry Rivers) is trying to win the game, so that's simple.  Understanding if a play is successful or not is a little harder.  DVOA doesn't just use EPA - it's more complicated, but you can read the full methodology here - https://www.footballoutsiders.com/info/methods#dvoa.  Teams (again, apart from the Texans) are trying to run successful plays most of the time - though there are a few exceptions*.

Other things that the model does is to literally ignore plays that have nothing to do with skill - recovering fumbles, having your opposition kicker miss a field goal and so on.  These are all worth debating, but I think the model generally does a good job with these.

So, you look at all these plays, for the Vikings, in this case, and see that the model is implying that the underlying skill level, on a per-play basis, is lower than for 19 other teams.  That they have been successful at translating that lower skill into more wins is a disconnect.  The model assumes that the disconnect is due to the randomness - in other words, the Vikings have been lucky that the randomness they have experienced, just like all the teams experience randomness, has contributed more towards their record than the randomness has for other teams.

You don't have to believe in the model.  It's just a tool.  You can ascribe the disconnect between per-play effectiveness and record to a weakness in the model, or the magical thinking that the Vikings are somehow more skilled in high-leverage situations, or anything else you want.  It's only a tool. 

But believing in the magic feather didn't mean that Dumbo didn't have to flap his ears in order to fly.

* For example, running the ball to burn clock rather than gain yardage - the garbage time that commentators often glom onto.  This is probably overstated because teams still want to gain yardage in these situations. If they didn't, they would just kneel, so going forwards is still part of the plan

Points: 1

#34 by Mike B. In Va // Dec 07, 2022 - 7:09am

Understanding whether a game is successful or not is easy - everybody (apart from the Texans, sorry Rivers) is trying to win the game, so that's simple.

Bravo. That's the most FO sentence I've ever read. 

Points: 1

#37 by Aaron Brooks G… // Dec 07, 2022 - 7:59am

if you're good, why wait until 3rd down?

Given modern rules, I wonder about that. It’s so easy to convert 3rd-long compared to historical norms.

Is there a point to having great success on 1st and 2nd down if you fail on 3rd down? I’d rather be mediocre at 1st and 2nd and great on 3rd.
 

Points: 0

#38 by Chuckc // Dec 07, 2022 - 9:27am

It doesn't.  Because it isn't repeatable.  Because if it was, you'd just win in all situations - if you're good, why wait until 3rd down?

 

Because you can't give maximum effort every play. You'd wear out before halftime. So you save that "dig deep" level of effort for the more important plays (3rd down, red zone).

Points: 0

#39 by Will Allen // Dec 07, 2022 - 9:49am

If the median offensive lineman consistently dials down effort in blocking Aaron Donald, his qb is getting carted off before halftime. Football is first and foremost about delivering violence, and scaling back effort in a regular pattern would be just as dangerous as it would be in boxing.

Points: 0

#41 by Chuckc // Dec 07, 2022 - 9:57am

The players aren't robots. They're (sometimes very) emotional human beings. Effort is going to vary from play to play. It's not 80% to 120% but more like 97 to 105. Adrenaline matters and it can amp up when people start to feel pressure. Some will thrive under that pressure and others will underperform.

Points: 0

#47 by Mike B. In Va // Dec 07, 2022 - 10:36am

Right, but that's not *predictable*, at least in any way anyone who's tried to track it has come up with. That same guy who puts in a great performance in go time this week may be thinking about his sick kid next week. So introducing random noise to the predictive model is not going to be helpful.

Points: 0

#49 by Aaron Brooks G… // Dec 07, 2022 - 10:49am

The tendency of statisticians to treat unexplained variance as random noise annoys me.

\most of the unexplored map did not, in fact, contain dragons

Points: 0

#52 by theslothook // Dec 07, 2022 - 11:27am

To me, that's why two things are important before anyone attempts using statistics and models for causal explanations. 

1) A coherent theory about cause and effect. 

2) A theory distilled down to very basic parts. A theory that can explain everything is an incomplete theory. 

A sidenote. If your entire goal is about prediction/forecasting and you don't care about explaining cause and effect, then that's a separate discussion. 

Points: 0

#54 by Aaron Brooks G… // Dec 07, 2022 - 12:08pm

That side note is valid. There's also a concept of good enough.

\now for p-hacking!

Points: 0

#58 by Mike B. In Va // Dec 07, 2022 - 1:03pm

Is unpredictable variance any different from random noise in a system like this at any meaningful level?

 

Edit: This is not snark. I'm not a statistician.

Points: 0

#60 by Aaron Brooks G… // Dec 07, 2022 - 1:24pm

It is tricky to determine whether variance is unpredictable (not able to be predicted) or merely unpredicted (failed to predict).

If you guess "tails" on a two-headed coin, the variance was predictable, but unpredicted. You can also argue the pedantic point that only white noise is random; the other colors are frequency-dependent*. (Think of it as bias, or drift)

A better model from the same data set will leave less unexplained variance.

But you need to be careful, because simply throwing increasing numbers of variables at a dataset will eventually create a perfectly explanatory model that probably can't explain worth a damn, because you overfitted the data. (Using 7 variables, I can perfectly explain a 7-sample dataset, because I just manually copied it)

\ https://en.wikipedia.org/wiki/Colors_of_noise
\\ * -- violent noise is interesting; it's the derivative of random noise, but the process of time-domain differentiation is frequency-dependent

Points: 1

#53 by Will Allen // Dec 07, 2022 - 11:37am

Now all you have to do is collect a large enough sample size of such situations, account for the interactions of the 22 players (the composition of which changes ovet the sample size) the field, the influence of referees, crowd noise, weather, various degree of injury, and then we might  be able to have confidence in the proposition, "The players and coaches for the (insert team name) are reliably more skilled/put forth better effort, relative to their opposition,  in high leverage situations, and then reliably perform markedly less well, relative to their opposition, in situations that are not highly leveraged."

I think a dummy like me is better served in having confidence that the Vikings have a blowout playoff loss in their early 2023 future.

Points: 0

#45 by Aaron Brooks G… // Dec 07, 2022 - 10:33am

If the median offensive lineman consistently dials down effort in blocking Aaron Donald, his qb is getting carted off before halftime. 

Does the median offensive lineman play for the 49ers?

Points: 0

#48 by theslothook // Dec 07, 2022 - 10:40am

DVOA has a heuristic component to it when it comes to defining "success" by downs. That's not a criticism because "success" does not have a true definition. I'd be curious how much DVOA swings if the success criteria were relaxed in both directions.

The other thing to note. AFAIK, DVOA treats every play in isolation. It's what did you do on first down. And what did you do on second down.

But play callers don't call plays purely in isolation. Coordinators and coaches talk at length about using the run to set up play action. It doesn't seem to work in the data, but that doesn't mean the decisions themselves aren't linked. In other words, a team conceivably could be intentionally being weak on first down to setup something incredible on third down.

Points: 1

#76 by Pat // Dec 07, 2022 - 5:07pm

Coordinators and coaches talk at length about using the run to set up play action. It doesn't seem to work in the data

That's just because they're sloppy with language. They're not actually trying to use the run to allow themselves to run play action. It's not like, a cheat code in a video game or something. They're using the run to manipulate the defense to be more susceptible to a play-action pass.

Analytics types try to test the "run to set up play action" by correlating runs with play action, but if you look at what I said, that won't work. You actually need to correlate runs and play action with the defense. You can't just "cheat" and skip over the defense part, because a less well known version of "correlation without causation" is that you can have "causation without correlation." It's perfectly possible to bury all of the correlation in a hidden variable (the defense) connected to the two play types.

The TL;DR version is that if you don't know why play-action works well in the first place, you can't make claims about what helps or hurts it.

Points: 1

#78 by theslothook // Dec 07, 2022 - 6:00pm

they're using the run to manipulate the defense to be more susceptible to a play-action pass.

My original point was that play calling is not done in isolation. 

Points: 0

#81 by Kaepernicus // Dec 07, 2022 - 6:37pm

There is a strategic component tied to play selection and it's sequencing too. Smart play callers know they are not going to score on every single play and have a longer sequence in mind when they call certain plays, jab plays if you will. If a team starts running too much play action without running the defense will stop biting. PA plays that are sniffed out seem to be significantly worse than a traditional running/passing play getting figured out. It would be cool to see the DVOA splits for successful/unsuccessful PA plays and how negative the failures are compared to traditional running and passing plays. 

Points: 0

#89 by Xexyz // Dec 08, 2022 - 12:27pm

It doesn't.  Because it isn't repeatable.  Because if it was, you'd just win in all situations - if you're good, why wait until 3rd down?

Perhaps is easier to play defense on 3rd down?  On 3rd down the defense is going to have a better idea of the kind of plays likely to be called than on 1st down.  This extra knowledge may be more beneficial to a poor defense than a good defense.   

Points: 0

#51 by JonesJon // Dec 07, 2022 - 10:57am

Week 3 Lions miss a FG to go up 6 with 1:14 left and the Vikings out of timeouts but Vikings score in 2 plays to win by 4.

Week 4 Will Lutz double doinks a FG to force OT as time expires, they win by 3

Week 5 they blow a 21-3 lead to the Bears, take the lead back with a 7 minute TD drive that features 5 3rd down conversions, force a fumble inside their 40 to seal the win by 7

Week 7 the Cardinals get into Vikings territory down by 8 on three occasions in the final 8:00 and fail to score each time, Vikings win by 8

Week 8 they rally from down 10 in the 4Q to beat the Commanders on a last second FG and win by 3

Week 9 the miracle Bills win

Week 11 the Patriots drive to the Vikings 30 down 7 late in the 4Q but stall out and the Vikings hold on for a 7 point win

Week 12 they both stop the Jets 3 times at the 1 and pick them off in the red zone in the final 2 minutes to preserve a 5 point win

 

Realistically, they should have lost 2 or 3 of these games and should be 8-4 or 7-5 right now

Points: 1

#10 by jmaron // Dec 06, 2022 - 4:28pm

Worst 9-3 or 10-2 Teams by DVOA, 1981-2022

The record of those teams in their remaining regular season games was 26-18, with an avg DVOA 1.85%, and ranking of 15th.

The teams went 8-10 in playoff games, outscored 415-436. They were good in their 1st playoff games 6-4 outscoring opps 286-235. Only the 1999 Titans got past their 2nd playoff game. 

My gut is DVOA is missing low on the quality of these teams, but not by a huge amount. 

 

 

 

Points: 1

#12 by Romodini // Dec 06, 2022 - 4:31pm

DVOA certainly loves the Cowboys again, but this team is also once again one of the most penalty prone in the league, improving from worst last year to ... second worst this year. I can already picture Brady tossing the bomb to Mike Evans that will draw the PI flag that ends the Cowboys' season in the wild card round.

Dak is also apparently experimenting with a new style of passing for himself which involves blindly tossing the ball to his receivers before they make their breaks. Credit to Dak for trying to improve on trusting his receivers, but with all the uncharacteristic interceptions he's been throwing lately, I don't think that's going to work in the playoffs. Considering how poor the pass block win rate of this Cowboys' line is, it makes me think it has less to do with trusting his receivers and more to do with not trusting his line to hold up. Maybe the coaching staff decided sacks were killing too many of their drives last year, and the risk of more dangerous throws was worth it.

Points: 0

#27 by LionInAZ // Dec 06, 2022 - 10:45pm

Funny, I've always heard that great QBs knew where the receivers were supposed to go and threw to that spot before they made their breaks. Waiting for the receiver to be open was a guarantee that the pass would fail. 

Points: 2

#30 by Romodini // Dec 07, 2022 - 12:56am

I know, it just doesn't work for Dak. He either isn't doing it right or his receivers aren't that great.

Points: 0

#35 by BlueStarDude // Dec 07, 2022 - 7:46am

It has seemed (just my impression from watching on TV, haven't rewatched much this year) like either Dak is slightly late or he just does not have the arm strength of the top tier QBs who routinely make those throws (Rodgers, Mahomes, Allen, Burrow) to consistently drive the ball in there on the right timing ahead of the DB getting there. And sometimes that is combined with the receivers being less than crisp out of their breaks.

Points: 0

#77 by ImNewAroundThe… // Dec 07, 2022 - 5:14pm

I'm sorry, Burrow? Arm strength? That's his biggest weakness and worse than Dak.

Points: 0

#87 by BlueStarDude // Dec 08, 2022 - 12:04pm

You're right (though not sure about worse than Dak's). That was a sloppy inclusion on my part.

Points: 0

#17 by DisplacedPackerFan // Dec 06, 2022 - 5:12pm

I quite enjoy seeing the GVAR split by unit. Though perhaps VAR by unit would provide the same info. As a GB homer yeah the defense hasn't varied much game to game, it's mostly been a turnstile against the run with middling pass coverage. The offense has been all over the place. So GVAR saying they are pretty inconsistent, mostly driven by the offense, feels right. But they have also felt like a pretty consistently average team all season overall never really swinging too high or too low (1st game against MN or maybe the Jets game feeling like the worst lows and Dallas feeling like the biggest high but even those don't feel too wildly off their average). So VAR feels right there too.

Some of the eye test variance that fans feel this season is due to opponents generally being pretty good on defense so that D adjustment kicks in and an eye test game where 3.5 ypc seems bad is actually pretty good, etc.

But there has been what feels like some week to week swings around a relatively small range. So they don't get much above 20% or below -20% for any games but they will go from 20 to -20 and back to 10 then down to -20 then up to 0, then up to 20 from week to week.

I'd have to dig at other teams but they are telling us different things and I think both things have value. Of course I'm always in favor of more measures and more splits. So my vote is just add GVAR. I'd also love to GVAR (and VAR) split by units if possible as the splits felt like the most insightful piece of information presented.

 

Points: 1

#22 by Raiderfan // Dec 06, 2022 - 7:11pm

Something seems wonky.  The link you provided about yardage differential for the 2010 Patriots went to a stat head chart that matches the number you quoted for through 12 games.  But the stat head chart had a link to the team stats  on PFR that showed a difference of less than 100 yards for the season.I have a hard time understanding how the number one team in DVOA that year—the Patriots were at 44,3%—were outgained by over three hundred yards in their last four games when they outscored their opponents 139-44.

help?

 For your question, VAR shows the Raiders rank #1 in defensive variance, while GVAR shows them ranking only#2. Given their uniform horribleness on defense, I vote for VAR.

Points: 0

#24 by AFCNFCBowl // Dec 06, 2022 - 7:22pm

They were outgained by that many yards in their first 12 games, not their last 4 (where they outgained CHI and MIA by huge margins).

Points: 0

#31 by David // Dec 07, 2022 - 4:38am

I'm curious to hear in the comments what the readers think about the two forms of variance. Which one do you think is a better measure of a team's consistency? Would you like to see us switch from VAR to GVAR in the future?

I'm much more interested in VAR than GVAR.  VAR tells us something predictive - how can we expect this team (or unit) to perform in the future.  A team with a low VAR is more likely to replicate its current performance (good or bad) than a team with a high VAR.  GVAR doesn't really help with that, as it is applying (arbitrary) endpoints to the measurement (the end of games).  Games are why we watch, of course, but GVAR isn't telling me anything I can't learn from a hundred other narrative-based analyses.

Points: 0

#73 by Dales // Dec 07, 2022 - 4:40pm

Hmm. Your reply, plus above where it's mentioned about how 17 games is inherently a small sample and DVOA works around it by going to the play level with a much higher n, got me thinking. There is already a "success rate" metric, but not all successes are equal and nor are all failures. Perhaps bucket them into gradients from atrocious to fantastic using some sort of objective measure, and then use *that* to measure variance?

Points: 0

#40 by mehnsrea // Dec 07, 2022 - 9:50am

Is there a way to remove special teams from the total DVOA calculation? I ask because I simply cannot understand how Baltimore is ranked where it is. Baltimore is not a good team. I am forced to watch them often because I live in the D.C. area. I can tell you that they have no chance of beating any sort of KC/BUF like opponent in the playoffs. I am dumbfounded by DVOA’s continued love for Baltimore (and its kicker). 

Points: 0

#46 by Aaron Brooks G… // Dec 07, 2022 - 10:34am

If you think Baltimore is excruciating to watch, is watching Washington like staring longingly at Cthulhu?

Points: 0

#66 by Mike B. In Va // Dec 07, 2022 - 3:17pm

I could definitely see Snyder as the Owner Who Should Not Be.

Points: 0

#67 by Aaron Brooks G… // Dec 07, 2022 - 3:30pm

Snyder is a Joseph Curwen.

I can see Jerry Jones as Henry Akeley, though.

 

Points: 1

#75 by mehnsrea // Dec 07, 2022 - 5:05pm

Hahahaha I don’t watch Washington at all if I can help it.

Points: 0

#50 by Mike B. In Va // Dec 07, 2022 - 10:50am

They came pretty close to beating Buffalo, but have no chance without Lamar. But the tight end, fullback and scrub receiver offense is kinda gut-wrenching to watch, I'll give you that.

Points: 0

#56 by jwkuyken // Dec 07, 2022 - 12:30pm

How about showing both.  Each has its own merits.  As a pro gambler any valid information that provides a certain level of predictability is of value to me.  Variance is a component that I use in determining whether to wager or not.  GVAR, to me, "smooths" out the variance component.  

Points: 0

#62 by JoelBarlow // Dec 07, 2022 - 1:35pm

goff = tua

Points: -1

#68 by andrew // Dec 07, 2022 - 4:00pm

I would like to propose a new statistic to track teams that lose to teams like the 2022 Vikings despite dominating DYAR, Yards, and other metrics:

Games Resulting In Defeat Despite Yards, or GRIDDY.     When can then take the GRIDDY factor into account when trying to explain this season.

Points: 8

#69 by Dales // Dec 07, 2022 - 4:07pm

"But look a little closer and things appear more consistent. For the first part of the season, the Giants were mostly an average team overall. Then Daniel Jones got hurt in the Week 12 game against the Eagles and the team collapsed. The first few games are mostly grouped together. The last few games are mostly grouped together...

"On the extremes, both methods tend to pick out the same teams as the most or least consistent. The two most consistent teams in the league this year, by both measures, are Washington and the New York Giants."

Is... is Daniel Jones exceedingly consistent?

 

I could do my own research but work and life give me little time. But as a Giants fan, it feels like he pretty much is. I can't think of too many games where he has been superb nor too many where he was God-awful.

Points: 0

#70 by Dales // Dec 07, 2022 - 4:19pm

Oh, should have added that I think GVAR is better. 

Points: 0

#79 by Kaepernicus // Dec 07, 2022 - 6:13pm

Has FO looked at defensive adjustments based on unit strengths? Rather than adjusting for overall defense vs. overall offense you could look at splits based on how each team matches up in different aspects of the game. So if the Dolphins torched the 49ers on 3rd and medium they would get less of a defensive adjustment than torching them with passes over the middle. Seems like it could add some additional predictive value if the samples are big enough. If those play counts take longer to get significant samples you could gradually transition to the more intricate defensive adjustments until the samples are large enough.

The MIA vs. SF breakdown FO did this week was incredibly valuable. The best passing defense over the middle vs. the best passing offense over middle was one of the deciding factors in that game. Most pundits have been trying to figure out why Tua was so off in that game and missing targets he had hit all year. The biggest thing I noticed was he started hesitating after the first Bosa sack and the crazy pass deflection by Warner on a key 3rd down in the first half. That split second of hesitation seemed to be leading to his miscues and it seemed to be causing a lot of the inaccuracy in the second half. The other split that was interesting was how bad SF has been on 3rd and medium all year vs. nearly every other defensive scenario. I believe the Dolphins were one of the top teams in the league on offense in those situations as well. Considering Miami went 0-7 on 3rd down I am assuming they did poorly on the 3rd and medium opportunities. SF winning their strength on strength match up convincingly and also winning their worst match up are the reasons they blew them out.

Points: 0

#80 by theslothook // Dec 07, 2022 - 6:29pm

Not to quibble with your overarching methodological point, I think there is a danger in overreacting to one game. 

I can understand why you would see strength versus strength and given the edge went decisively to one side, make conclusions. But remember, even in games featuring rematches, you get wildly different results.

That's why I try not to draw any conclusions of a team based entirely on one or two games. In fact, that's mostly why I try not to let the playoffs color my opinion of what I thought of these teams the minute the regular season ends.

Points: 0

#84 by Mike B. In Va // Dec 08, 2022 - 7:13am

Nathan Peterman's entire career is defined by being a guy who's a split-second slow on passes in real games, as opposed to practice. At least when he was with Buffalo. If you watched him practice, apparently, he made all the right reads and put the ball in the right place, but could seldom do it for real in a game, with real game pressure. While he's an extreme example, I still think there's something to be learned there.

We have a body of work that says Tua can be that same guy who's timing is off - but with enough pressure, every QB is that guy, even TB12. (See two SBs against the Giants if you've forgotten...) The jury is still out on if he's "fixed" or not, much like 2020 was for Allen.

Points: 0

#85 by Aaron Brooks G… // Dec 08, 2022 - 7:59am

Peterman also fumbled the ball like a greased golem.

Points: 0

#86 by Eggolas // Dec 08, 2022 - 11:39am

Giving up a lot of yards is not good. But it also needs to be viewed in context. How do the Vikings compare in the starting position of their opponents’ drives to that of the average team. Is average starting position the correct criteria? Wright has been excellent at pinning the opposition inside their 20 and not punting into the end zone this year.

 

Points: 0

Save 10%
& Support Aaron
Support Football Outsiders' independent media and Aaron Schatz. Use promo code SCHATZ to save 10% on any FO+ membership and give half the cost of your membership to tip Aaron.