NFL Parity

(Originally posted 10/21/09)

It seems like every year - especially early in the season - you hear lots of people talking about parity in the NFL.  Some years the talk is about how even all the teams are (high parity) and some years the talk is about how many bad teams there are (low parity).  This year seems to be a year where there is quite a bit of talk about how parity has disappeared.

After just 6 weeks (some teams have only played 5 games), it is really not reasonable to make a determination about parity.  The apparent parity (or lack thereof) this early in the season is going to be strongly related to the NFL schedule.  For instance, the New York Giants have played 2/3 of their games against some of the worst teams in the league (Was, TB, KC and Oak) - so their 5-1 record might make the Giants look like a better team than they really are.

So I thought I would take a quick look and see how parity this year compares to other years.  I only looked at NFL seasons since the AFL-NFL merger in 1970.

For simplicity's sake, I calculated parity by determining the average games over .500 that each NFL team finished in a given year.  For instance, if one team finished 12-4, they were 8 games away from .500.  If another team finished 7-9, they were 2 games away from .500.  I did this for every team, added them up and then divided by the number of teams.  I also normalized the values to a 16-game season for all seasons in which fewer than 16 games were played.

Here are the all-time parity rankings. 1975 comes in as the year with the least amount of parity. In 1975 there were 26 teams in the league and they each played a 14 game schedule. 19 of the 26 teams either won or lost 10+ games that year. There were 3 other teams that either won or lost 9 games (meaning if they played a 16-game schedule that year, there was a decent chance that 22 of the 26 teams would have won or lost 10+ games!).

Year	Average
1975	7.2
1976	7
1973	6.4
2005	6
1972	5.9
1977	5.7
1991	5.7
1992	5.7
1970	5.6
1986	5.6
1982	5.5
1998	5.5
1990	5.4
2003	5.4
1984	5.3
2000	5.3
2001	5.3
2008	5.2
1971	5.1
2007	5.1
1980	5
1985	5
1974	4.9
2004	4.9
1997	4.8
1979	4.6
1988	4.6
1987	4.5
1996	4.5
1989	4.4
1994	4.4
1999	4.4
1978	4.3
1981	4.3
2002	4.3
2006	4.3
1993	4.1
1983	3.9
1995	3.9

Here is a graph of the parity trend over time:

Click the image for a larger version.

Now the question is: How does 2009 (so far) compare?  If I use the same method to calculate parity for 2009, I will get a parity rating of 7.2 through Week 6, which would tie the 1975 rating.  But that's not a fair rating.  Right now the NFL sports 9 teams that are either undefeated or winless.  A straight extrapolation of records to 16 games means we'd end the season with 9 teams undefeated or winless - we know that won't happen.  Instead, I calculated each team's pythagorean record through 16 games to come up with an expected W-L record for each team.  A perfect-parity season would be if every team finished 8-8, and the parity rating would be 0.  A perfect un-parity season would be if half the teams finished 16-0 and half finished 0-16 (which could probably never happen due to the schedule) and that provide a parity rating of 16.

At the top of the standings, Pythagorus says that the Saints will win 14 games, while the Colts and Broncos will win 13.  It also says that the Rams will go 1-15 and the Titans, Browns and Raiders will go 2-14.  These seem like reasonable approximations of what could happen - though I'd guess that there will be fewer than 4 teams that win fewer than 3 games.

Using these figures, I come up with a parity rating of 6.3 for 2009 through Week 6, which would be the 4th most un-parity season of all time.  It wouldn't surprise me if the actual parity rating for 2009 is fairly high (remember, high rating means low parity), but I suspect it won't be as high as 6.3.

Updates:

Through	W-L	Pythag
Week 7	7.3	6.9

Update:
With a full season of data, 2009 finishes with an actual rating of 5.0 (using actual W-L instead of Pythag. The Pythag rating was 5.1) That puts 2009 approximately in the center of the all-time rankings. Also, the all-time average is a parity rating of 5.1 which means that 2009 ends up finishing with a less varied league than average.

Even if the Colts had gone on to a 16-0 season, that would have only bumped 2009 up to 5.1 - right at the all-time average.

So in conclusion, the early-season perception that the league had lots of really bad teams and lots of really good teams was not accurate. It probably had more to do with the way the schedule worked out early on, and maybe a higher than usual number of hot and/or cold teams early.