1.   AFC Power Ratings

            Cleveland has replaced Oakland as the worst team in the conference; New England continues to be the best, despite a few losses.    

            Most people will tell you that Indianapolis is the best team in the conference, of course, because they are 10-0.   The heart of the dispute is the issue of whether there is an ability to win, as distinguished from an ability to score points and an ability to prevent the other team from scoring points.   In baseball we have had this debate for many years, and it seems fairly clear that the ability to win distinct from offense and defense, if such a thing exists at all, is a low-level element.  In football this is less clear.  In football you can substitute to fit the needs of the game.   Indianapolis has won their last four games by 4 points, 3 points, 1 point, and 2 points.   No one questions but that they are a good team, but are they are a good team that has been lucky, or are they are good team that has an ability to win close games—therefore, a great team?   All good sportswriters know the answer to this question.  I, unfortunately, do not.

2.   NFC Power Ratings

            Ya know, actually, the NFC power ratings battle would be a lot more interesting if you turned them upside down.    St. Louis and Detroit are waging a great battle for the position of the conference’s worst team.   New Orleans, on the other hand, is making a mockery of the contest for the top spot:

3.   Best Offenses

            New Orleans and New England, obviously; the Monday night matchup should see somebody in the 40s.  The very definition of a marquee matchup:  the two best teams in the league, the two best offenses in the league, the two hottest teams in the league, Monday night football.   Oakland’s “offensive contribution score” is up to zero:

4.  Best Defenses

            Much of sports writing is an echo chamber.   People are convinced that Indianapolis is the best team in the NFL not only because the Colts are 10-0, but also because they have been good for about ten years now, and the repetition of excellence has gradually ground the dissenters into meal.  

            A few years ago, however, the team was in the Super Bowl with a historically awful defense (2006).   This year, at least by our analysis, the biggest key to their success is their defense:

            But the Colts, as you probably know, are in the middle of the pack when it comes to surrendering yards of offense, but in first when it comes to surrendering points.   If I understood that better I could figure it out.

            As Yogi might have said. . ..

5.  Strength of Schedule

            Miami has still played the toughest schedule in the league, but it’s evening out a little.   Minnesota has passed Washington as the darling of the schedule makers:

6.  Current Temperatures

            Have yet to establish that it means anything—or that it doesn’t.  

7.  Consistency

            The NFL’s most consistent teams have been:

            Baltimore’s “consistency” is a sort of odd concept.   Baltimore’s two worst games this year, by the game output score, are the two losses to Cincinnati, by scores of 17-14 and 17-7—but in that Cincinnati is a good team, losing to them by 3 points at home or 10 points on the road is not a horrific outcome.   Their best game of the year is a 30-7 thrashing of Denver.

            Baltimore will probably be the best team this year that doesn’t make the playoffs.   We’ve kind of gotten into watching Miami, struggling to get back in the playoff hunt after starting out 0-3, 2-4, 3-5, facing a brutal schedule, and Tennessee, trying to become the first team ever to rally from 0-6 to a playoff spot.  But Baltimore is a significantly better team than Miami, and has the same record and the same problem.   Baltimore has outscored their opponents this year by 66 points, but has not been able to win the close games—consistently.  When they have faced good opposition, they have played consistently just well enough to lose.  

            The most inconsistent team has been the Titans:

8.  Performance Following Bye Weeks

            Every NFL team has now had its Bye Week and played a game following the Bye Week, so we can now calculate the apparent effects of the Bye Week.   This chart compares each team’s season average performance—their rank—with their game output score for the game following the Bye Week:

            18 teams had better-than-average performances following their Bye Week, 13 had worse, while Washington exactly hit their usual mark.  Six teams did play their best game of the week following the Bye Week (Atlanta, Baltimore, Dallas, Pittsburgh, St. Louis and Tampa Bay), whereas at random we would expect only three teams to do that.  

            However, the “average” or expected effect does not appear to be large—a gain of a little less than a point.   That’s misleading, because there were several games that matched two teams both coming off a bye week.   If we threw those out of the calculation, as we should, that would make the average a little more than a point, rather than a little less than a point.  But it still doesn’t appear to be as big an edge in fact as it is in the popular imagination—at least in this season’s data.

9.  Last Week’s Predictions

            In last week’s predictions we were 11-5, our first “good” week in four, making us 55-41 on the season.   We got quite a few of the spreads and scores about right.   We had Jacksonville over Buffalo 20-17; it was actually 18-15.   We had Indianapolis/Baltimore as a two-point game; it was actually one.  We had Cleveland/Detroit as a one-point game; it was actually two.  We had New England over the Jets 24-14; it was actually 31-14.   Even the Atlanta/Giants game, which we were wrong about, we were actually wrong about by only a few points.    Of course, we didn’t foresee the Kansas City upset of Pittsburgh or Oakland over Cincinnati, but. . .who did?

            But wait a minute; we were 11-5.  11-5 is not that good.   Last year we were going 11-5, 10-6, 12-4 just about every week.   11 and 5 only looks good because we’re been doing so badly week in and week out.

            What I was trying to get to was the question of why?    What is different about this season, that is causing so many games to go against us?  There are four obvious things that it could be.  

It could be that teams are up and down more this year than they usually are, more inconsistent week to week.    It could be, but I don’t think it is.

It could be that we’re just missing the close games.   It could be that when we say Team A should win by 2, they lose by 3.   It could be, but I don’t think it is.

It could be that the league is “tighter” this year than it was last year, more compact.   A lower standard deviation of the quality of teams would cause more games to be within range of an upset.  It could be, but I don’t think it is—in fact, I’m fairly sure the standard deviation of the quality of teams is higher this year than it normally is.   We’ve got two 10-0 teams, one of whom (New Orleans) has only played three competitive contests.   We’ve got several teams that are truly, epically terrible.  

What I think it is is, the quality of several teams has shifted in mid-season—causing us to mis-read where teams are by looking at early-season data.   Last year, the teams that were good early in the year, with maybe one or two exceptions, were good late in the year.   This year there are like eight teams that were good early in the year but stink now, or vice versa.  

Last week I “predicted” that Denver would beat San Diego—fully expecting this prediction to be wrong.   That’s what the numbers show; that’s the outcome of the process.   I’m not an expert; I just use the process.  

But you’d have to be an idiot to actually think that Denver, at this point of the season, is on the same level as San Diego.   These are Denver’s game output scores for the season:

They’re just not the same team they were early in the year, not even close.   Early in the season they were playing at the level of Pittsburgh and Arizona and Philadelphia.  Now they’re playing at the level of St. Louis or Detroit.  

I’m not even sure that’s the most dramatic example.   Tennessee turned it around at the bye week, in the opposite direction:

There are numerous other examples.   Chicago started out playing well.  Houston was absolutely awful two of the first three weeks.    The Giants played well the first five weeks, stunk the next four.   The Giants are playing Denver this weekend; it’s hard to guess which team has disintegrated more completely.  San Francisco is not playing at the same level now they were early in the season.   Kansas City and Washington are playing much better than they were early in the year.   That’s why we’ve missed so many predictions—the normal misses that we will always feature have been augmented this season by the system using game data that just really doesn’t describe where teams are anymore.  

10.  Predictions for This Week (Week 12)

But. . .this is my story and I’m sticking with it.   Systematic analysis, I suspect, almost always starts out at a level below the accuracy of conventional wisdom.   It should.   The difference is, systematic analysis can be systematically improved.   You reach a point, eventually, where you’re able to contribute to conventional wisdom, and then perhaps you reach a point at which conventional wisdom races to stay astride.   Yes, we’re idiots.   That is the genius of what we do.