Let’s start with the SRS ratings through nine weeks, excluding the Green Bay/Chicago Monday Night Football game:

Today, I want to use the SRS on the single-game level to analyze two things: the best/worst performances of the year, and the most surprising results of the year (regular readers will remember that I did this last year, too).

The two best single-game SRS scores come courtesy of victories over the 49ers. Seattle won by 26 in Seattle (after adjusting for home field, a Margin of Victory of 23), giving them the same score as the Colts, who won by 20 points in San Francisco. Since the 49ers are 8.4 points better than average, each of those wins gets recorded as +31.4.

What about the biggest surprise of the year? That came when the Rams upset the Texans. St. Louis has a rating of -4.4; when the Rams traveled to Houston (SRS of -6.0), we would “expect” — when retrofitting the data — that the Texans would win by 1.4 points. Instead, St. Louis won by 25 points, producing a result that is 26.4 points off of our “projection.” That makes that game the biggest surprise of the year. The second biggest surprise? The Eagles (-1.3) winning in Oakland (-8.1) by 29 points, when they were expected to win by just 3.8 points.

As always, the table below is fully sortable and searchable (it’s showing 25 entries right now, but you can change that, too). Here’s how to read the first row. In week 3, Indianapolis traveled to San Francisco, and won, 27-7. That gave them a Margin of Victory of 23 (since the Colts were on the road), against a team that has a rating of +8.4. Therefore, the Colts get an SRS score of 31.4 in that game. The “Exp” column shows what you would expect to happen based on the location of the game and each team’s SRS rating: here, the Colts (6.9) in San Francisco (8.4) would be expected to lose by 4.5 points. Since Indianapolis actually won by 20 points, the Colts get a “Diff” of 24.5.

There’s a lot to digest, so I’ll leave the analysis to you. But there’s one more thing I want to do with that data: figure out which teams are most/least consistent. If you take the absolute value ^[1]By definition, the sum of all the “Diff” results for each game will be zero for each team on the season level. of the difference column for each team in each game, you get a sense of the most/least consistent teams. If you ask me, the results here match my intuition: the Texans, Jets, Colts, 49ers, and Eagles are the least consistent teams, while teams like Green Bay (again, this is prior to Monday Night) and Detroit have been successful but flying under the radar. The “variability” of each team — defined as the average of the absolute value of the “Diff” column in the table above — is in the “Var” column in the table below.