Preseason Power Rankings

Calculated Power Rankings from preseason of the 2009 NFL season.  I use an engineering method of iteratively solving for best fit to the data (scores) to create the ranks and related applications that corrects for opponents played.I’m going to do a weekly power rankings series throughout the season. I am going to eschew subjective analysis and focus solely on a mathematical analysis. My primary variable will be scoring, with teams ranked 1 to 32 based on the net difference between offensive and defensive scoring.

As opposed to using raw scoring, the scoring offense and defense will be corrected for the strength of opponent, and the expected scoring against an average opponent calculated. Each week the program makes about 100,000 calculations to reach the final numbers. The net will be drawn from this expected scoring. This method of ranking is similar in concept to the BCS rankings I believe, however the parity in the NFL somewhat eliminates the need to blow out lesser teams to maintain ranking.

Lacking any live data, for the preseason rankings I am starting with the basic input of last years data. This is then modified for the preseason performance of the first string offense and defense (weeks 1 and 2 first quarter plus week 3 first half). After tallying scores, 10% of the difference between what the first team scored and gave up in these 4 quarters, and the expected amount (from last season’s post-week 17 data that covered every game last year), is then added to that expected score, then the net between offense and defense calculated to rank the teams. I capped this modifier at +/- 2, which only applied to one unit of one team, the Packers offense, which scored more than 40 points more than expected. A quick check of the preseason scoring average for the sum of these 4 quarters shows that the scores are in line with what is expected. A team scored on average 22.0 points in 2008, those 4 quarters saw an average score of 20.1 points.

Much like Football Outsiders and their power rankings, the preseason ranking data will be used through week 4 to smooth the live data, after week 4 only 2009 data will be used. The raw data from 2008 is corrected for the strength of opponent (I did a dry run of the program week by week using last season’s data to verify/tweak functionality), the preseason data from 2009 is not opponent corrected.

I should have the weekly rankings prepared by some time Tuesday morning throughout the season.

Preseason Power Rankings:
1. Baltimore Ravens
2. Pittsburgh Steelers
3. Tennessee Titans
4. New York Giants
5. New Orleans Saints
6. Indianapolis Colts
7. Green Bay Packers
8. San Diego Chargers
9. Philadelphia Eagles
10. Minnesota Vikings
11. Chicago Bears
12. Atlanta Falcons
13. New England Patriots
14. Carolina Panthers
15. Dallas Cowboys
16. Tampa Bay Buccaneers
17. Miami Dolphins
18. Houston Texans
19. New York Jets
20. Jacksonville Jaguars
21. Washington Redskins
22. Cleveland Browns
23. Arizona Cardinals
24. San Francisco 49ers
25. Buffalo Bills
26. Denver Broncos
27. Seattle Seahawks
28. Cincinnati Bengals
29. Kansas City Chiefs
30. Oakland Raiders
31. Detroit Lions
32. St. Louis Rams


Tags: , , , , , , , , ,

2 Responses to “Preseason Power Rankings”

  1. Week 1 Power Rankings « Where's Lambeau? Says:

    […] to individual games. 80% of the data this week is the calculated preseason power rankings (see here for calculation methodology), with 20% being calculted from the past weekends games. Week 1 data […]

  2. Week 2 Power Rankings « Where's Lambeau? Says:

    […] purposes, the rank in the DVOA equivalent, 2009 live data only, is presented in the bracket. (see here for calculation […]

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: