Last season the Vikings were thought to be one of the luckiest teams in the league. Most pointed to their 11-0 record in one-score games- 11-1 including the playoffs- but there are a particular set of luck factors that happen on the field and influence the outcome of football games.
NFL Football Operations have analyzed several events that have a significant impact on the outcome of an NFL game and which the outcome is significantly determined by luck. These include dropped interceptions by an opponent, dropped passes by an opponent, field goal and extra point attempts by an opponent, and fumble recoveries by either team. They went on to quantify the net luck factor for each event in terms of wins added or subtracted over the course of the 2022 season. Each event affects win probability based on game situation- time remaining, down, distance, yard line, timeouts remaining. Those win probabilities are added up for each event in each category to arrive at the total.
Based on those factors, the Vikings weren’t the luckiest team in the league last year- they were fifth luckiest with an additional 0.84 wins attributed to luck.
What a Difference a Season Makes
So far this season, not only has the Vikings lucky streak ended, it has turned sharply negative. Neil Greenberg of the Washington Post did a story on the luckiest teams of Week One based on the same factors mentioned above. He found that the Buccaneers were the luckiest team in the league Week One- beating the Vikings despite just a 36% win probability based on the events of the game.
In fact, the Bucs were the 8th luckiest winning team in Week One going back to 2002:
But the Buccaneers luck in Week One against the Vikings pales in comparison to the luck the Eagles enjoyed on Thursday night. Consider again the luck categories and what happened in the Eagles game:
Field Goals and Extra Points
Vikings: 4-4 XP
Eagles: 4-4 XP, 2-3 FG (24 yards, 55 yards (missed), 61 yards)
There were a total of five fumbles in the game- 4 by the Vikings and 1 by the Eagles. The Eagles recovered four, and the last by Justin Jefferson resulted in a touchback- effectively another recovery for the Eagles.
Of these four categories, the results of three of them favor the Eagles, one significantly and another overwhelmingly.
My guess is that the dropped pass advantage for the Eagles, while still an advantage in luck, wasn’t that significant in terms of win probability. But the likelihood of making a 61-yard field goal is only 16% according to NextGenStats, This is off-set some from the Eagles missing a 55-yard field goal, which has around a 60% probability of success, but overall still a luck factor that favors the Eagles.
The fumble recoveries were an off-the-charts luck advantage for the Eagles. If you consider that a team has a 50% chance of recovering a fumble, the probability of a team recovering all five fumbles in the game is just 3.125% - same as getting heads five times in a row when flipping a coin.
And if you consider that three of those fumble recoveries happened deep in negative territory- Jefferson’s fumble for a touchback, Cousin’s strip sack, and Covey’s punt return fumble- basically resulted in close to a 21-point swing in favor of the Eagles, while the Powell and Mattison fumbles added at least 6 more points in favor of the Eagles, that’s a total of around a 27-point swing in favor of the Eagles, probably a bit less in EPA terms, but still an overwhelming swing in favor of the Eagles. Given that the Eagles won by only 6 points, if even one of those turnovers is recovered by the Vikings, that would’ve changed the outcome of the game.
Be that as it may, the probability of the Eagles winning a game in which five fumbles happened- as in this game- along with everything else is quite low. If the Bucs had only a 36% chance of winning in week one, my guess is that the Eagles had only around a 10% chance of winning the game on Thursday night.
Regressing Way Past the Mean
When it comes to the luck factors, the Vikings have not only regressed to the mean, but their luck has also swung to at least twice as negative as it was positive last season.
For example, last season the Vikings fumble recovery rate was 52.63% (20/38). They recovered 55.56% of their own fumbles (10/18) and 50% (10/20) of their opponent fumbles. The circumstances of those fumbles and recoveries added just over one win to the Vikings’ total last season. In fact, at +1.07 wins added, the Vikings were the luckiest team when it came to fumble recoveries, based on the circumstances of each fumble, even though they recovered only one more than 50% of them.
Looking at specifics, you could say the Vikings won the Bills game based on the fumble recovery in the end zone, giving them +1 win, while all the other fumbles and recoveries last season came out about even. Had the Vikings not recovered that fumble in the Bills game, the fumble recovery percentage would have been an even 50% for the Vikings and their opponents and win probability gained/lost from them would’ve been zero.
Now consider this season.
Already there have been seven fumbles in the Vikings’ two games (six by the Vikings, one by the Eagles). And the Vikings’ fumble recovery percentage is 0%. Compared to last season, when the Vikings were one fumble recovery above average, they’re already down 3.5 fumble recoveries with basically two games subtracted from their win total. This isn’t regression toward the mean. This is an extreme swing at least twice as much to the negative as the Vikings were to the positive in luck factors last season. And this is just considering the fumble recovery category. The Vikings haven’t, on balance, been lucky in the other three luck categories this season either.
What It Means for the Rest of the Season
Given the Vikings’ fumble recovery rate so far this season and the number of fumbles, if there were a total of 38 fumbles again this season (same as last season), the Vikings could recover 18 of the next 31 fumbles (a 58% rate) and still have regressed fully to the mean over two seasons. If you assume 20 fumbles for the Vikings and 18 for their opponents this season (splitting fumbles evenly between Vikings and opponents over the two seasons), that would give the Vikings a +6 turnover margin on fumbles over the rest of the 15-game season - and this would still be a full reversion to the mean (50%) in terms of fumble recovery percentage over two seasons.
Of course anything can happen, but even the unluckiest of outlier teams recovered at least 33% of fumbles over the course of a season, with roughly 66% of teams recovering between 55%-45% of fumbles and 95% recovering between 62%-38% of fumbles. The Vikings have currently recovered 0% of fumbles.
If you also consider that the Vikings have averaged about 21 turnovers a season (adjusted for a 17-game season) over the past ten seasons, and they’ve already had six, that means an average of one per game the rest of the season or 15 more. If you couple that with an average of 24 takeaways per 17-game season over the last ten seasons, with a low of 20 takeaways, that suggests somewhere between a +4-8 turnover margin the rest of the season.
There is also the Detroit Lions example last season. They started out with a -6 turnover differential (which is where the Vikings are now) going 1-5 in the process- but then finished +13 in turnover differential and went 8-3 over those 11 games. If the Vikings have concentrated most of their bad luck in the first two games, as the Lions did in their first six last season in terms of turnovers, and the Vikings had the same winning percentage as the Lions following the change in turnover fortunes over their remaining 15 games (.727), they’d go 11-6 this season. Keep in mind teams that win the turnover battle win 70% of the time.
There is also the example of the 49ers last season. They recovered just 45.24% of fumbles last season (24th in the league) but still led the league in turnover margin at +13.
Overall, given how the season has started for the Vikings, regression toward the mean should now result in the Vikings having significantly better luck the rest of the season.
Follow me on X/Twitter @wludford
The Vikings currently have a 0% fumble recovery rate. 50% is average, most teams have a 45% or better rate, and roughly 95% of teams recover at least 38% of fumbles. What will the Vikings fumble recovery rate be closest to at the end of the season?