In search of truth about the Vikings' point differential
By now you have probably heard the statistics about the Minnesota Vikings’ point differential but it bears repeating because it’s pretty wild.
No team that’s ever started 10-2 has had a lower point differential. Last year the Vikings were plus-3 through 12 games and had a 5-7 record. This season they have a plus-10 differential and have thrived on the back of nine one-score wins.
In the 20 years since Moneyball came out, we have all come to understand the concept of regression. If a batter hits .400 in May, he’s unlikely to keep it up through June, for example. If a football team isn’t running away from opponents on a regular basis and they are still winning, odds are that they will come back to earth at some point. Only one team has ever reached the Super Bowl with a lower point differential than the Vikings have right now and last year’s representatives were both over plus-80.
But there’s a few things worth noting about regression and football. It isn’t a big sample size sport like baseball. In a month, baseball clubs will play nearly twice the games as an NFL season. Can you keep hitting .400 through 21 games? Sure. Will the Vikings stop winning this way over the next 100 games? Of course. But they don’t have to play 100 games.
Point differential is certainly a good indicator but it’s not the only thing to factor when determining team strength. We have access to many different stats these days that tell us more specifically which teams have played the best ball, so PD can’t be the only measure of a squad.
How are we supposed to sort it all out? Let’s see if we can pick apart the point differential conundrum…
We should start with their schedule, which appears to be a factor in all the close games. Pro-Football Reference’s Strength of Schedule metric lists 2022 as the eighth most difficult in team history. The Vikings have matched up with the No. 4, 5, 6, 7 and 8th ranked defenses in the NFL and come out 3-2 in those games.
On the matter of the one-score games, head coach Kevin O’Connell gave his take on Monday:
“I don’t think it’s an accident that our team continues finding ways to win and doing enough to win,” he said. “The important thing is in those moments whether it is situational football, understanding how clock management and how our team plays a role in that in all three phases. How we can play smart aggressive football, but also understand that we’re setting ourselves up to do those things to win those games in the end… I do know our team is confident that when we can get into the fourth quarter of these games with winnable situations.”
The eye test would align with O’Connell. We have repeatedly seen the Vikings make plays at the end. Sometimes they have been aided by a miscue but it doesn’t seem completely random that players like Harrison Smith, Patrick Peterson, Danielle Hunter and Za’Darius Smith can stop the other team when it matters most.
And not all one-score games are built equally.
The analytics website FiveThirtyEight attempted to create a more accurate measure for coin-flip games than just using the final score. In 2020, they looked at ESPN’s Win Probability model and ruled the coin-flip games as ones where the win probability was between 40%-60% with under five minutes to go at any point. Those games, they decided, truly could have gone either way.
Let’s do the same exercise for the Vikings’ one-score games.
— The Lions had a 90% chance to win with 2:36 left but that flipped to a Vikings win in the final moments on a KJ Osborn touchdown. Check.
— With 5:07 left in the game against the Saints, New Orleans had a 66.3% chance but a pass interference call swung that to 66.6%% in Minnesota’s favor. Check.
— The Vikings never sunk below 68% against the Dolphins. This one doesn’t count as a coin flip.
— Arizona only had an 11% chance to win with five minutes remaining. No coin flip here.
— Strangely even with the score at 17-17, the Win Probability model gave the Vikings a 68.8% shot to win versus the Commanders with five minutes remaining. No.
— Vikings-Bills basically broke the WP calculator. Both teams were 90% or higher odds within the final five minutes. Check.
— At 4:32 left against the Patriots, the Vikings sat with a touchdown lead at 87.8%. No.
— With three minutes left versus New York, the Jets were over 50% to win. Check.
So, by FiveThirtyEight’s method, only four out of the nine games were truly coin flips. You can certainly talk about the breaks the Vikings received like a Taylor Heinicke overthrow in the fourth quarter or a kick return for touchdown but the Vikings earned those wins more than they were the beneficiaries of randomness. Of course, if they had bad luck and went 0-4 in coin-flip games, a 6-6 record would certainly feel wildly different than 10-2.
Going 4-0 in mostly random endings doesn’t have to spell impending doom. In 2019, the 13-3 San Francisco 49ers went 7-3 in games that fit the FiveThirtyEight coin-flip qualifications. They were a first down away from winning the Super Bowl. Though the Seahawks also went 5-1 in coin-flips and lost in the divisional round in 2019 as well.
Not all random-influenced teams are the same. The 49ers were a complete team in 2019 with a top-five Expected Points Added passing game and that stat has most correlated with reaching The Big Game. Every Super Bowl winner since 2015 has ranked in the top five in passing EPA. The 2019 Seahawks ranked sixth in passing EPA but were more of a paper tiger, ranking ninth overall scoring and 22nd in points against.
The Vikings are presently 13th in passing EPA, by the way. And on the defensive side only one team has allowed more than 6.0 yards per play against on defense and reached the Super Bowl since 2010 and the Vikings are currently giving up — you guessed it — 6.0 yards per play.
Pro-Football Reference’s metric to weigh overall performance versus schedule only puts the Vikings in the ballpark of the 1994 and 2012 versions.
Not quite conclusive, not quite without concern.
Justin Jefferson put it this way to NFL Network’s Tom Pelissero on Sunday:
“We’re going to celebrate this win, look at the things that we need to fix offense — because there’s a lot — and go into next week ready to roll.”
It’s important to again remember that there is still a ways to go before the end of the season. Five games represents 30% of the year yet to be played. The Vikings, mostly healthy, can fix problems. It’s the best (and possibly only) way to fight regression and randomness.
The Vikings have an opportunity to steamroll some teams like the Colts and Bears and, while it’s not likely that they end up with the standard for Super Bowl teams which is around a plus-100 differential, they could still bring themselves out of the outlier range and into the lower-end Super Bowl contender ballpark by Week 18 by point differential.
We could end up saying, “Well, if they had schemed X/Y/Z earlier in the year they wouldn’t have had those close games,” or “Well, if they played Indy and Chicago in Week 10 and 11, the point differential discussion wouldn’t have been a thing.”
That’s the best-case scenario. The evidence regarding the Vikings’ point differential heading into Week 14 is that it tells a story but not the whole story. We’ll have a better idea if it matters based on what they look like when the playoffs start.
