Predictably Unpredictable: Welcome To A Week In The NFL
OAKLAND, CA – DECEMBER 19: Quentin Groves #52 of the Oakland Raiders celebrates with Kamerion Wimbley #96 and Matt Shaughnessy #77 after he stopped Correll Buckhalter #28 of the Denver Broncos for a safety at Oakland-Alameda County Coliseum on December 19, 2010 in Oakland, California. (Photo by Ezra Shaw/Getty Images)
On October 24, 2010, the Oakland Raiders traveled to Invesco Field to take on their division rival Denver Broncos. The Raiders had entered the season riding a streak of seven-straight losing seasons, and the new decade looked to be picking up where the previous one had left off. After dropping four of their first six games—including an anemic 179-yard performance the week prior against the then-mediocre 49ers—they entered Denver as touchdown underdogs.
And then the game began. Six minutes in, the Raiders were up 21-0. By halftime, it was 38-7. At the end of three quarters, the carnage had grown to 59-14, after which the sudden juggernauts called off the dogs and cruised through a scoreless fourth. One week after struggling to get past the line of scrimmage, the Silver and Black pounded their way to 508 total yards (328 on the ground, most in the NFL that season) despite sleeping through the final quarter. When the dust finally settled, the Raiders had outperformed the pundits in Vegas by a staggering 52 points (45-point margin + 7-point spread).
Plenty of writers spill plenty of virtual ink every fall day telling you what is totally, definitely, without-a-shadow-of-a-doubt going to happen that coming Sunday. Rather than try to suppress the chaos that is the National Football League, I’m going to embrace it. This article is about the most surprising games, teams, and seasons of the NFL’s recent past.
To keep things simple, I’m going to define “surprise” by how much a game fooled the folks who have the most on the line with their predictions: the Vegas oddsmakers. The greater the gap between the final margin and the pregame point spread, the greater the surprise. Whichever team covers the spread is considered the “over-performer”.
As you may have guessed, the Raiders’ 52-point over-performance ranks as the most surprising game of the time period that I studied (the past 8 regular seasons). The New England Patriots’ 59-0 shellacking of the Tennessee Titans in 2009 (50-point over-performance when taking into account the 9-point spread in the Pats’ favor) takes second place.
Already, we see the beauty of this study. On the top of the leaderboard are two teams that couldn’t be more different: a perennial doormat as a road dog versus a perennial powerhouse as a home favorite. They even dominated their respective games in completely different fashions: Oakland with 300+ yards on the ground, New England with 400+ through the air.
The bronze medalist 2012 Seahawks show us yet another way to shock the football world. They took advantage of a +7 turnover margin (yes, you read the correctly) to steamroll the 10-point underdog Cardinals 58-0 (over-performance of 48 points).
In case you’re wondering, the biggest shocker so far in 2014 was a 35.5 over-performance by the Falcons in a 56-14 pasting of the Bucs (a margin that could’ve been worse without two meaningless Tampa fourth quarter touchdowns).
Here’s the rest of the top 10 from 2006-2013:
Believe it or not, the Patriots’ absurd 433 net total yard advantage over the Titans was the not the highest in the time period studied. That honor goes to last year’s Lions, who out-gained the Packers by 435 yards in late November (thank you, Matt Flynn!). Detroit “only” over-performed by 24 points in that game. A month earlier, those same Lions also accomplished the bizarre feat of outgaining the Cowboys by 355 yards but under-performing the line by two points (they still won the game, 31-30). If that wasn’t enough Motor City madness for you, take a look at the Lions’ December 11, 2011 game against the Vikings. Detroit managed to under-perform by four points despite benefiting from a +6 turnover margin.
While we’re on the subject of hard-to-read teams, which club holds the distinction as the least predictable? In order to measure each team’s 8-year “Predictability Index,” I calculated the standard deviation of its line performance. In plain English: the more spread out a team’s line performances, the higher its unpredictability score.
This should make intuitive sense. A team that usually is within a few points of its line each week is, relatively speaking, predictable. The real head-scratchers are teams that can go from a +20 to a -20 in consecutive weeks, a behavior that’s reflected in the standard deviation metric.
Since the standard deviation figures don’t make much sense in the abstract, what you see below is the percent above or below average that each team’s predictability score lies. The higher the number, the more unpredictable.
Is it fate that the teams who gave us two of the most surprising Super Bowls ever sit at the top of this list? The Patriots owe their ranking to being so good that they are capable of massive over-performances, and the Giants to the mercurial nature of Eli Manning. This finding is probably old news to New Yorkers, who are used to the local papers alternating between sainting and shaming Manning on a weekly basis.
On the other end of the spectrum lie the Browns, who have stumbled along in their losing ways in a highly predictable fashion. At least the Raiders, who have been equally inept over the past eight seasons, keep things interesting with big wins and even bigger losses.
Speaking of the Silver and Black, the Raiders also hold the distinction of the most unpredictable single season: 2010, the one that included the 52-point shocker over the Broncos. In fact, that year the Raiders went 3-0 against the Broncos and Seahawks, winning by an average margin of 44-13. Good luck trying to repeat that showing this year, boys.
The most predictable season belonged to the 2010 Ravens, who had just a 20.5-point gap between its largest over-performance and under-performance. Twelve of their games were within six points of the line.
The award for biggest one-week Jekyll and Hyde turnaround goes to the 2012 Cardinals, who went from the aforementioned woeful 48-point under-performance against the Seahawks one Sunday to a 34.5-point over-performance against the Lions (again with those Lions!) the next.
Despite the lack of information and game tape, September is actually the most predictable month. 2009 was the most unpredictable overall season, while last year was the most predictable, but there was no general pattern over time. And in case you’re wondering whether a team’s over/under-performance one week predicts its performance the following week, try again. The correlation between the two is -0.02 (i.e. no relationship whatsoever). Which, given what we know about the NFL, should have been completely predictable.
***
Martin Kleinbard is the co-founder of The Bandwagn, a newsletter/website that allows non-sports diehards the chance to join the sports conversation. He can be reached at martin.kleinbard@gmail.com.
You May Also Be Interested In These Stories
