Division Power and Super Bowl Winners

Division Power and FWDP

APR already gives a way to compute a power index for individual teams. Division power can be computed by averaging teams power index on a per-division basis.

Just as team power indexes can be computed for each week of the regular season, division power indexes can be computed for each week as well. For the purposes of analyzing the division power of Super Bowl winners, I will use the division powers from the final week of the regular season.

For the purpose of future discussion, I will abbreviate Final Week Division Power as FWDP.

Super Bowl Contestants by Division Power

The the list of the strongest divisions by FWDP since 1960 has the divisions of lot of Championship and Super Bowl winning teams on it. Which naturally suggests there might be a correlation between strong FWDP and winning the Super Bowl.

Looking at the seasons from 1990 to 2007, there appears to be a very strong correlation between strong FWDP and winning the Super Bowl:

Year	High FWDP	Result	Low FWDP
1990	New York Giants	beat	Buffalo Bills
1991	Washington Redskins	beat	Buffalo Bills
1992	Dallas Cowboys	beat	Buffalo Bills
1993	Dallas Cowboys	beat	Buffalo Bills
1994	San Francisco 49ers	beat	San Diego Chargers
1995	Dallas Cowboys	beat	Pittsburgh Steelers
1996	Green Bay Packers	beat	New England Patriots
1997	Green Bay Packers	lost to	Denver Broncos
1998	Atlanta Falcons	lost to	Denver Broncos
1999	Tennessee Titans	lost to	St. Louis Rams
2000	Baltimore Ravens	beat	New York Giants
2001	New England Patriots	beat	St. Louis Rams
2002	Oakland Raiders	lost to	Tampa Bay Buccaneers
2003	New England Patriots	beat	Carolina Panthers
2004	New England Patriots	beat	Philadelphia Eagles
2005	Pittsburgh Steelers	beat	Seattle Seahawks
2006	Indianapolis Colts	beat	Chicago Bears
2007	New York Giants	beat	New England Patriots

That's a 14-4 record (77.7%) for that span of games, which suggests that FWDP is a strong indicator of which team wins the Super Bowl.

Something Happened in 1990

I chose the cut-off year of 1990 advisedly. The results from 1970 to 1989 are quite different:

Year	High FWDP	Result	Low FWDP
1970	Dallas Cowboys	lost to	Baltimore Colts
1971	Dallas Cowboys	beat	Miami Dolphins
1972	Washington Redskins	lost to	Miami Dolphins
1973	Miami Dolphins	beat	Minnestoa Vikings
1974	Pittsburgh Steelers	beat	Minnesota Vikings
1975	Pittsburgh Steelers	beat	Dallas Cowboys
1976	Minnesota Vikings	lost to	Oakland Raiders
1977	Denver Broncos	lost to	Dallas Cowboys
1978	Pittsburgh Steelers	beat	Dallas Cowboys
1979	Pittsburgh Steelers	beat	Los Angeles Rams
1980	Oakland Raiders	beat	Philadelphia Eagles
1981	Cincinnati Bengals	lost to	San Francisco 49ers
1982	Washington Redskins	beat	Miami Dolphins
1983	Washington Redskins	lost to	Los Angeles Raiders
1984	San Francisco 49ers	beat	Miami Dolphins
1985	New England Patriots	lost to	Chicago Bears
1986	Denver Broncos	lost to	New York Giants
1987	Washington Redskins	beat	Denver Broncos
1988	Cincinnati Bengals	lost to	San Francisco 49ers
1989	San Fracisco 49ers	beat	Denver Broncos

That's a 11-9 record (55.0%) for that span of games, which suggests that there is little or no correlation between FWDP and which teams won the Super Bowl during that era.

What Happened in 1990?

If you've followed the NFL for any length of time, you know that they are constantly changing the rules. Often to protect the safety of the players, or to make the game more exciting.

A number of significant changes were made during this era. Free agency and the salary cap were introduced. The post-season was expanded from 5 playoff teams per conference to 6.

I would like to have a definitive answer to put here. However, given the small sample space, the number of rule changes, and the fact that there's no way to "replay" old seasons with different rules, it's probably impossible to say what caused the change.

I would also like to be able to make some claim that the 1990-2007 correlation will continue in the future. But with no specific cause to point to, there's no way to make the case that some subsequent change won't undo (or hasn't already undone) the effect.

The Future

Following the final week of the 2008 season (and for each future season I keep FSPI going), I will post the FWDP for all 8 divisions. In the pick result post following the Super Bowl, I will link back to the FWDP post, note whether or not the team with the better FWDP index won, and update the 1990-present tally.

NFL: Worst (and Best) divisions since 1960

Recently, Football Outsiders posted an article pondering the Worst Division Ever. The actual article (on ESPN's site) uses non-division records and combined net points; no attempt is made to put the divisions in order.

APR already gives a way to compute a power index for individual teams. This can easily be extended. Division power can be computed by averaging teams power index on a per-division basis.

By computing the division powers at the end of each season, APR provides a numerical ordering of divisions, much like the weekly team power rankings.

Some Disclaimers

I only have score data going back to the 1960 season, so the ranking of those 1930's divisions will languish in obscurity.

It should also be noted there's been a lot of changes in the NFL since 1960: the merger of the AFL and NFL, the introduction of a number of new teams, the introduction of sudden-death overtime for regular season games, playoff seeding, the salary cap and free agency, the division realignment in 2002, and rule changes too numerous to mention.

Comparison of division (or team) power indexes from year to year (or between the AFL and NFL prior to the merger) should not be taken too seriously.

APR's Worst Divisions, 1960-2008 (week 11)

Of course, I'm not going to let any of that stop me. Here is the top 20+ worst divisions since 1960, including the 2008 season up to week 11. Obviously (right?) the end-of-season rankings for the 2008 divisions are subject to change.

Here are the 20 worst (and a few more) divisions in the data set, ordered from worst to... not worst. Hopefully the table is pretty self explanatory; divisions marked with a '*' were mentioned in the linked article.

Rank	Year	Division	Power
1	1969	NFL Century	1.1406	*
2	2007	NFC West	1.1542
3	2004	NFC West	1.1544	*
4	1991	AFC East	1.1565
5	2008	NFC West	1.1565	*
6	1965	NFL East	1.1594
7	1979	NFC Central	1.1662
8	1982	NFC West	1.1700
9	2008	AFC West	1.1760	*
10	1982	AFC Central	1.1765
11	1976	NFC West	1.1770
12	1998	AFC Central	1.1789
13	1975	NFC West	1.1790
14	1984	AFC East	1.1800
15	1985	AFC Central	1.1813
16	1961	AFL West	1.1821
17	2002	NFC North	1.1826	*
18	1976	AFC East	1.1847
19	2007	NFC South	1.1848
20	1967	AFL East	1.1866	*
Other divisions mentioned:
27	1969	AFL East	1.2023	*
73	1968	AFL East	1.2342	*
76	1979	NFC West	1.2367	*
109	1984	AFC Central	1.2528	*
111	2008	NFC North	1.2541	*

By APR's measure, the 1969 NFL Century is the worst of the lot by a pretty significant margin. The 2007 NFC West doesn't look that bad from the Fooball Outsider's metric (non-division record was 14-26 (35%) with a combined net points of -213; marginally better than those mentioned). But both the Rams and 49ers had miserable seasons in 2007 (very much like 2008, so far) that are (perhaps) not completely reflected in their records or net points.

APR's Best Divisions, 1960-2008 (week 11)

Of course, the same data set yields the best divisions, too. Here are the top 10 divisions, ranked from the best, descending.

Rank	Year	Division	Power	Playoff Notes
1	1965	NFL West	1.4513	Packers won NFL Championship
2	2004	AFC East	1.4441	Patriots won Super Bowl XXIX
3	2007	NFC East	1.4356	Giants won Super Bowl XLII
4	2006	AFC East	1.4154	Patriots lost AFC CG
5	2006	AFC South	1.4104	Colts won Superbowl XLI
6	1978	AFC Central	1.3953	Steelers won Superbowl XIII
7	1979	AFC West	1.3919	Chargers, Broncos 1-and-done
8	1979	AFC Central	1.3916	Steelers won Superbowl XIV
9	1970	NFC Central	1.3896	Vikings, Lions 1-and-done
10	1992	NFC East	1.3865	Cowboys won Super Bowl XXVII

The highest-ranked 2008 divisions are the NFC East (currently ranked #20), and the AFC South (currently ranked #34). I have another post coming that looks at the correlation of division power and Super Bowl winners. Stay tuned!