SITE NEWS: We are moving all of our site and company news into a single blog for Sports-Reference.com. We'll tag all PFR content, so you can quickly and easily find the content you want.

Also, our existing PFR blog rss feed will be redirected to the new site's feed.

Pro-Football-Reference.com » Sports Reference

For more from Chase and Jason, check out their work at Football Perspective and The Big Lead.

Parity II

Posted by Doug on August 16, 2006

This is a continuation of yesterday's post about parity. In that one, I identified two different kinds of parity:

1. Within-season parity, also known as any-given-Sunday parity. This brand of parity is characterized by having lots and lots of teams in a given season that seem interchangeable. You feel its effects when you look at a Sunday slate of games in midseason and think, "I have no idea who is going to win any of these games."

2. Across-season parity. This is characterized by teams turning from good to bad and vice versa very quickly. The NFL's reputation for this kind of parity was probably born in the late 90s and early 00s when the the Falcons, Rams, and Patriots all went from being terrible to being in the Super Bowl.

I will talk a little bit about across-season parity tomorrow, but for now here is one more look at the history of any-given-Sunday parity.

It seems that a consequence of this kind of parity would be a lot of close games. If a lot of teams are evenly matched, then we ought to see a lot of close games, right? Here is the leauge's percentage of games decided by three or fewer points, and also by seven or fewer points:


Year % within 3 % within 7
================================
1970 18.7 32.4
1971 19.2 41.8
1972 20.9 39.0
1973 15.4 33.0
1974 20.3 50.0
1975 19.2 34.1
1976 19.4 37.2
1977 18.4 43.4
1978 21.9 48.2
1979 22.8 46.4
1980 25.9 48.2
1981 26.8 40.6
1982 26.2 48.4
1983 24.1 47.3
1984 25.9 42.4
1985 17.0 38.8
1986 21.4 47.3
1987 19.0 47.1
1988 27.7 50.4
1989 24.6 47.8
1990 24.1 43.3
1991 25.4 50.0
1992 21.4 39.3
1993 23.7 46.9
1994 26.8 51.3
1995 25.4 47.9
1996 19.6 45.4
1997 27.9 46.2
1998 20.8 47.1
1999 25.8 46.4
2000 24.6 44.0
2001 25.0 48.8
2002 24.6 49.2
2003 23.4 48.4
2004 23.8 45.3
2005 23.4 44.5

Finally, here is the yearly standard deviation of team's ratings from the simple rating system. Using the simple rating system instead of records should, at least theoretically, remove any scheduling bias. Remember, the standard deviation tells you how spread out the observations are. So a big number would indicate low parity and a small number would indicate high parity.


Year St.Dev.
================
1970 7.2
1971 6.0
1972 7.5
1973 7.9
1974 5.9
1975 7.5
1976 8.6
1977 6.3
1978 4.6
1979 5.4
1980 5.4
1981 5.3
1982 5.6
1983 5.4
1984 6.7
1985 6.1
1986 5.9
1987 5.7
1988 4.7
1989 5.7
1990 5.9
1991 7.0
1992 6.0
1993 5.2
1994 5.0
1995 5.3
1996 5.4
1997 5.1
1998 6.8
1999 6.3
2000 6.4
2001 5.4
2002 5.5
2003 5.4
2004 6.2
2005 6.5

This entry was posted on Wednesday, August 16th, 2006 at 5:06 am and is filed under General. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.