Pete Rozelle once envisioned a league that would have every team finish 8-8.
We know that kind of level playing ground is a pipe dream. But how is competitive balance done in the NFL? How close have we come to that parity?
There are numerical ways that we can go about measuring this. Rozelle was concerned with wins and losses, so why don't we concern ourselves with the same?
The first thing we have to ask ourselves is: What is parity, and how do we achieve it?
Rozelle said that ultimate parity comes when everyone finishes 8-8—meaning no separation between teams in the standings.
Statistics give us a perfect, easy-to-find measurement to gauge the separation in teams' standings. We simply use the standard deviation (STDEV) between the teams' win totals.
Standard deviation measures the average distance between a set of numbers. For 2009, the standard deviation between NFL teams was 3.22 wins.
When it comes to this figure, lower is better. The lower the difference between the teams, the more competitive the league.
Another measurement we can use is the separation between the best and worst teams.
This is easy. Take the team with the most wins, and the team with the fewest, and see how many wins separate them.
This year, the Colts led the pack with 14 wins, and the Rams brought up the rear with only one. This makes a range of 13 wins. Again, the lower the better. The smaller gap between the best and worst teams, the more competitive the league.
Finally, we can use the median of the data set to measure how well the middle-of-the-road teams did. The median is the number in the center of the data set. If the median is eight, for example, there will be the same number of teams with more than eight wins as there are with fewer than eight. A higher median means the middle of the league did better.
Now, how can we take these three numbers and produce something that will tell us how competitive the league was?
I am not a mathematician, but I made a simple formula to produce an index of the league's competitive balance—a number I called the "Quality Index." My formula was MEDIAN/(STDEV+DIFFERENCE). The higher the result, the more competitive the season was.
For example, Pete Rozelle's perfect season would produce an infinite Quality Index, since every team going 8-8 would produce a difference and standard deviation of zero, meaning we would be dividing by zero.
If every team went 9-7 or 7-9, we would have a Quality Index of 2.649. On the flip side, if every team went either 15-1 or 1-15, the Quality Index would be .379.
When the 2009 totals were plugged in, it returned a value of .493.
Now, on its own, this number means nothing. So, in order to give it meaning, we have to compare it to past seasons. I excluded the 1987 and 1982 strike-shortened seasons, since the teams did not play 16 games those years.
I ran every season since the expansion to a 16-game schedule through the formula and compared them next to each other.
The results were not pretty for 2009.
Overall, the season finished on the wrong side of average in every category. The average deviation of wins since 1978 is 3.04, compared to 2009's 3.22 (remember: lower is better). The average Quality Index over this time is .557 (this time, higher is better), and 2009 clocked in at .493.
Overall, 2009 ranked as the sixth least competitive season since 1978.
Here is a brief look at the most competitive seasons using my method:
1988- QI .726
1979- QI .704
1993- QI .691
2002- QI .674
1995- QI .643
And here are the five worst:
1984- QI .488
1991- QI .485
1990- QI .462
2007- QI .437
2001- QI .431
There are a couple things to note from this study.
The first is that competitive balance has shown no trend from season to season. The 2001 season was the least competitive of the 16-game era, and the very next year was the fourth best. Looking at these numbers in a line graph will show that the numbers bounce around like the sea in a hurricane over the 30 seasons.
Second, this is only one method of doing things. We are not taking into account playoff races. For instance, 2009 featured a crazy AFC playoff race that had five teams fighting for two playoff spots in the last week of the season. I am simply trying to create a numerical value for how competitive the teams were.
Third, this formula can be transferred over to any sport. As long as all the seasons you compare had the same number of games, you can compare the level of parity from any league.
We learned that competitive balance is nowhere near what Pete Rozelle envisioned, but I doubt that is even possible.
We also showed that there is no season-to-season trend for competitive balance. Just because one season was not as strong doesn't mean the next will also have a poor showing.
This is only one step to measuring the NFL's level of parity. Many more methods can and should be tried in the future. Let me know what you think of the study and how it can be improved.