clock menu more-arrow no yes mobile
NFL: Minnesota Vikings at Buffalo Bills Mark Konezny-USA TODAY Sports

Filed under:

Yes, the Minnesota Vikings are frauds, but what exactly does that mean?

A look at regression to the mean, outperforming your underlying numbers, and good old-fashioned lucky breaks.

One thing you will hear repeatedly leading up to the Green Bay Packers’ game this weekend against the Minnesota Vikings is that the Vikings are frauds. But what does it mean to be a fraud in the first place? Is this some analytical gobbledygook that can be ignored? Is it even a bad thing for the Vikings? And does it mean the Packers are going to win this weekend?

Let’s take a closer look at football frauds and see if we can learn something.

1. What is a fraud, and is it bad to be one?

First of all, “fraud” is a bit of a loaded term. Fraud, in the common vernacular, implies a person or party actively trying to deceive, and the Vikings haven’t done anything of the sort. They’ve simply played their games, and in several instances, they’ve benefitted from incredible luck. We use “fraud” as part of rivalry speak. In some limited cases, when an analyst is discussing Minnesota as one of the NFC’s elite teams like Philly, San Francisco, or Dallas, that analyst may be engaging in something like fraud, though it’s more likely they just don’t understand or trust this kind of analysis, which again, isn’t intent to deceive.

The better term is probably just “lucky.” The Vikings have been unquestionably lucky, in both obvious ways (the fumble in Buffalo and the greatest comeback in NFL history empowered by the Jeff Saturday-led Colts), and non-obvious ways (success in one-score games). But “lucky” isn’t as fun as “fraud” and doesn’t garner the same troll-reaction on Twitter, so fraud will likely retain its popularity.

And no, it’s not bad at all to be a fraud! The 2019 Packers were frauds, going 13-3 with a +63 point differential, which ranked 5th in the NFC and 9th in the NFL. They were 9th overall in DVOA, which is good, but hardly reflective of their elite 13-3 record; when they ran into the much better San Francisco 49ers team, who were 5th in overall DVOA and 2nd in defensive DVOA, it showed. Every team will, at some point, find themselves outperforming their underlying numbers, and if you can ride that hot streak through the end of the season, you can occasionally even win a title as a fraud.

It’s difficult, for reasons we will get to in a moment, but the 2011 Giants did it with a negative point differential and the 2007 Giants also did it, going 7-1 in one-score games. Damned Giants. Getting lucky can present you with opportunity, and if you can capitalize on the opportunity, no one cares if you weren’t supposed to be there.

2. How do we identify frauds?

The Vikings are frauds because their record is much better than their underlying numbers, but what exactly are “underlying numbers” and why do we care? While old school analysts are keen to quote Bill Parcells’ famous “you are what your record says you are,” that’s not really true, especially in a small sample size season like the NFL’s. This concept, as it relates to sports, goes back to baseball, Voros McCracken, and Defense-Independent Pitching (DIPs) theory.

After Bill James and Billy Beane, McCracken is perhaps the most famous old timey sabermetrician. McCracken’s key insight was that there were some aspects of pitching that pitchers control (walks, strikeouts, and home runs) and some they do not (balls in play, which were impacted by defenders and park-specific issues). Based on this, he developed the statistic Defense Independent ERA (dERA), which, over time, morphed into the more familiar FIP (Fielding Independent Pitching). FIP is scaled to resemble the more popular Earned Run Average (ERA), where 3.00 is pretty good. The crazy thing about FIP, and the reason it’s interesting and still a part of every Baseball Reference stat box, is that by focusing on what pitchers control, and excluding what they don’t, FIP predicts future ERA better than ERA does. And that is, fundamentally, what we mean by “underlying numbers.”

While FIP and its ilk kicked off a monumental shift in baseball statistics, the more important impacts were focusing an entire generation of nerds to start quantifying which statistics were prone to regression to the mean and to see if we might be able to do better or isolate luck from skill.

Getting back to football — as it turns out, a team’s record is a terrible predictor of future performance. There is just so much luck inherent in football that teams playing above or below expectation is the norm, not the exception. Aside from the fact that we have a mere 17 regular season games in a season, football also incorporates a ton of high-leverage plays that are subject to randomness. As an example, way back when Football Outsiders was just getting started, they would often focus on how teams performed on third down versus first and second down. Third down is, of course, much more important than the other two downs, and failing (or succeeding) on third down has an outsized impact on winning. FO found that teams with huge splits in their early-to-late down performance would often have huge shifts the following year as their third down performance fell back in line with their overall offensive performance. (This has declined slightly over the last two decades as teams have realized the increased importance of running plays past the sticks, and specifically targeting high-percentage conversions on third down. Yes, this was not always a thing.)

Relative success on late downs is just one of many hugely impactful events prone to random variation. Fumble luck is another huge contributor to team success or failure from year to year. Excluding botched snaps (which are very likely to be recovered by the offense), fumble recoveries are random events, and overall, we can expect half to be recovered by the offense and half by the defense. When a team benefits from more than their expected share of fumble recoveries, they will often find themselves crashing back to 50/50 (or worse) the following season.

Opposing field goal percentage is another one. Field goals, especially near the end of games, are frequently high leverage events, and if a given team’s opponents happen to miss more kicks than we would expect, it can have a huge impact on win-loss record. Opposing dropped passes can have a similar impact.

On a larger scale, we know, from decades of data, that one-score games in the NFL are basically 50/50 toss ups. Our own Justis Mosqueda made this point annually back in his Setting the Edge days, and it remains the case that teams with well over (or well below) .500 records in one-score games will almost always bounce back to closer to .500 the following season. There are few more reliable trends in football.

Finally, to bring this back to DIPs theory, we now have more sophisticated models like DVOA and EPA to tell us just how efficient (or inefficient) and offense is on a play-to-play basis, and a good idea about how many points any given team should have scored or allowed based on these models. The old Pythagorean Record still works surprisingly well, but we can do even better.

3. Why specifically are the Vikings Frauds?

The Vikings have all of the hallmarks of an incredibly lucky team. Their point differential is just +5, indicating right away that they are more like a .500 team than their actual 12-3 (.800) record. The top three other NFC teams have differentials of +145 (San Francisco), +137 (Philadelphia), and +131 (Dallas). Minnesota sticks out like a sore thumb.

Minnesota is also extremely lucky in one-score games, where they are currently 9-0, which is, statistically speaking, bananas. In this case we will define one-score games as seven points or less, but the Vikings also have two 8-point wins. Indeed, their only truly impressive win of the season came in week one against Green Bay, 23-7.

The Vikings haven’t lost much, but when they have, they’ve lost convincingly: 24-7 to the Eagles, 40-3 to the Cowboys, and 34-23 to the Lions. Some will claim that this team just happens to be clutch, and hey, Justin Jefferson has made some truly incredible plays to get his team back in games. The thing is, all similarly lucky teams in the past have made the exact same claim, and they are just never correct. The thing about incredible plays is that they are rare and not to be relied on for consistent production, even from a talent like Jefferson.

On a play-to-play basis, Minnesota simply isn’t very good. They rank 25th overall with a -9.4% DVOA, where their average-y offense and defense are dragged down by atrocious special teams, which rank 28th. By EPA/Play they are almost perfectly average on offense, where they rank 17th with -0.0002, and on defense, where they rank 14th with +0.006. Everything in their underlying numbers screams that they’re a .500 team.

So where is the luck coming from?

Penalties

First, it’s in penalties. The Vikings have unquestionably benefitted from the officials this season. I want to be clear that this isn’t some conspiracy to make the Vikings successful, as accusations about officiating often come bundled with such things. This is simply a result of the natural variance in officiating. Sometimes you get calls, and sometimes you do not.

The Vikings, Penalties, and Fumbles

Minn Penalty Yard Fumbles Fumbles Lost Opponent Penalty Yards Fumbles Fumbles Lost
Minn Penalty Yard Fumbles Fumbles Lost Opponent Penalty Yards Fumbles Fumbles Lost
Minn 3 15 2 1 DET 7 59 0 0
Minn 5 35 0 0 NO 10 102 2 2
Minn 7 70 1 0 CHI 6 40 3 1
Minn 2 20 1 0 MIA 10 97 2 1
Minn 10 86 1 1 ARI 10 67 2 1
Minn 5 42 0 0 WAS 3 23 0 0
Minn 7 64 2 0 BUF 7 64 2 2
Minn 3 20 0 0 NE 6 55 1 0
Minn 3 44 0 0 NYJ 6 40 0 0
Minn 9 74 1 1 IND 11 103 3 1
Minn 2 14 0 0 NYG 7 63 2 1
Minn 5 25 1 0 PHI 8 60 1 0
Minn 7 59 2 1 DAL 6 39 1 0
Minn 3 25 2 2 DET 7 51 0 0
Minn 3 20 0 0 GB 2 10 1 1
Totals 74 613 13 6 106 873 20 10

The Vikings have been called for 74 penalties on the year, totaling 613 yards. Their opponents have been called for 106 penalties, totaling 873 yards. In their wins of 8 points or less, they have been flagged 56 times for 484 yards, while their opponents have been flagged 83 times for 713 yards. The Vikings suffered double digit flags on just one occasion, against Arizona. Viking opponents suffered double digit flags four times. On the other end of the spectrum, Viking opponents were flagged for 5 or fewer penalties in a game twice. The Vikings were flagged 5 times or fewer in a game 10 times. Refs matter.

Is Kirk Cousins clutch?

What about Kirk Cousins? Has he been clutch in the 4th quarter? Over the course of his career, Cousins has his highest passer rating (108.2) and completion percentage (70.55%) in the first quarter, after which he settles in at around a 94-97 passer rating and 65% completion percentage.

If Cousins seems to be extra clutch this season, it has much more to do with a slight decline in earlier quarters than some newfound fourth quarter magic. In 2022 his first quarter has dipped to a 93 rating and 69% completion percentage, and his third quarter has really taken it on the chin, dropping to a 75.4 rating, and 62.5% completion percentage. In the 4th he soars back to a 65.48 completion percentage, almost identical to his career numbers. His passer rating is better at 104.8, due to one fewer interception than he usually throws. Cousins has been “good” in the fourth quarter mostly because he’s been worse in the 3rd quarter. It’s really more about distribution of his good and bad play than about his overall level of play, which shows up in his DVOA. Since joining the Vikings in 2018, Cousins has ranked 20th, 10th, 10th, and 8th in DVOA. This season he ranks 12th, right in line with where he always is.

I would urge anyone who is still skeptical to just take a look at the play-by-play on Pro Football Reference for the last 5 minutes of every Viking win, just to see all the weirdness for yourself. You probably remember the insane Colts game and the equally unlikely finish to the Bills game, but almost every Vikings’ win has some sort of asinine ending. From the Lions trying and missing a game-sealing 54-yard field goal with 1:14 left, followed by consecutive 28-yard passes from Cousins to KJ Osborn, to a late Jaylen Waddle fumble in a game that Tua didn’t even play, to the Bears losing a late fumble by Ihmir Smith-Marsette, to Washington losing a chance at a comeback due to an unnecessary roughness penalty against John Ridgeway on a field goal attempt. The Vikings are lousy with statistical mediocrity, to be sure, but their “obvious luck” is also right there for anyone who cares to look.

4. Does this mean the Packers rout the Vikings on their way towards making the playoffs?

Absolutely not! For one thing, the Vikings already own a win over Green Bay, and have a better point differential than Green Bay. The Vikings may be frauds, but their “true talent” is more like what the Packers have been all season. Minnesota isn’t a bad team, they’re just not a 12-3 team. The Packers may very well be without Christian Watson and/or Keisean Nixon, their offensive line is banged up, and Rashan Gary is out. There is a good case to be made that Minnesota is the better team, and I expect this will be a tough game for both sides. While DVOA prefers the Packers, by EPA they’re pretty much dead even.

So yes, the Vikings are frauds. They have all of the telltale markings of a lucky overachieving team, and then some. That doesn’t mean they aren’t still dangerous, and it doesn’t mean they can’t parlay this bit of luck into something special if their luck just holds out long enough. Fortunately for the Packers, she has a very unladylike way of running out.

NFL

Here’s what you missed in Aaron Rodgers’ astrology webinar

NFL Draft 2023

NFL executives name standout draft prospects from Senior Bowl practices

Former Packers QB Brett Hundley set to be XFL’s highest-paid player