Let’s start with the SRS ratings for every team in the NFL. The SRS ratings are generated based off of the points scored, points allowed, home field, and opponent for each game. In its simplest form, the SRS is just an SOS-adjusted version of points differential, although the devil is in the details. After running hundreds of iterations to get the ratings to converge (and awarding 3 points to the home team), below are the ratings through week 8:

You might be surprised to see the Chiefs at #2 — I know I was — but remember Kansas City beat New England by 27 points. The Chiefs also beat the Dolphins by 19 points and the Rams by 27 points, and split two close games against the Chargers and Broncos.

But that’s not the point of this post. The point here is to look at non-conforming games, and no game was weirder than the Chiefs first game of the season. Take a look at the table above: Kansas City has an SRS rating of 9.8, while the Titans have a rating of just -6.7. Based on those ratings, we would expect the Chiefs to win a head-to-head matchup by 16.5 points; put the game in Arrowhead, and our projected line jumps to 19.5.

Yet in week one, the Jake Locker-led Titans beat Kansas City by 16 points! That means the game went a whopping 35.5 points in the other direction, which makes it the Single Weirdest Game of the Year. The table below lists every game this season from the perspective of both teams, although I’ve limited the search function to only search for the team in the first column. Here’s how to read New York/Washington row, which represented the second weirdest game of the year. In week 4, the Giants traveled to Washington; New York has an SRS rating of -0.2, while Washington is slightly worse at -3.0. Therefore, giving 3 points for home field, we would expect New York to lose by 0.2 points. In reality, the Giants won by 31 points,which means New York’s margin of victory was 31.2 points greater than expected, making it the second weirdest game of the year. The table below shows all 242 games, although the table by default only lists the top ten.

We can also break things down by units. For example, the Broncos offense ^[1]Using points scored as a proxy for offense. has been 8.0 points better than average, while the defense ^[2]Using points allowed as a proxy for defense. has been 2.4 points better (i.e., stingier) than average. Meanwhile, Denver’s offensive strength of schedule (i.e., measuring the strength of the defenses the Broncos faced) was 2.8 points tougher than average, while the defense (i.e., the quality of the opposing offenses) has been 1.9 points harder than average. Add it all up, and the Broncos have an SRS of +15.2 points.

In what should not be surprising, the Steelers/Colts game comes in as the least conforming game of the year with respect to points scored by one team. Even including that game, the Colts defense has been 0.5 PPG above average, while Pittsburgh’s offense has been just 1.7 points above average. On a neutral field, that would mean we would expect the Steelers to score 1.2 more points than average against Indianapolis; put the game in Heinz Field, and that number jumps to 2.7 points (if we split the 3 point advantage evenly among both offense and defense).

Since the average team has scored 23.3 points this year, we would project Pittsburgh at home against Indianapolis to score 26.0 points; instead, the Steelers scored 51! By finishing 25 points above average, it was the biggest outlier performance by an offensive team this year. The other biggest outliers: Atlanta against Tampa Bay (20.9 points higher than expected), New York/Washington (20.7), Carolina/Cincinnati (18.9), and Minnesota/Atlanta (18.5).

The least conforming offensive performances in a negative way? That belongs to the Giants, who scored 21.3 points below expectation against the Eagles. New York was shut out on the road, in a game featuring roughly average offenses and defenses. Also on the bad side: Cincinnati’s shut out against Indianapolis (-21.1), Kansas City/Tennessee (-18.7), Patriots/Raiders (-17.5), and Browns/Jaguars (-16.5). Of course, by definition, the least-conforming offenses in a negative way list is the same as the most-conforming defenses in a positive way list.

Leave your thoughts in the comments.