Blog Entry

## My Mathematically Correct Power Rankings

Posted on: October 4, 2009 1:10 am

These are a list of power rankings I came up with that have absolutely no opinion involved. Instead, I used a mathematical equation to determine them. Here's how it works...

Offensive points + total yards + passing yards + rushing yards + points allowed + rushing yards allowed + passing yards allowed + total yards allowed. Divide this total by 8. I would then look at their records and give that a rank as well. Teams that are 3-0 received a rank of 1, 2-1 a rank of 2, 1-2 a rank of 3, and 0-3 a rank of 4. I took the sum of the first equation and added the win/loss rank to their total. The team with the lowest score is the "best" team.

For example, the Ravens are the number one ranked team. Here is their equation:

Baltimore: 2+2+5+5+11+7+17+1 = 50 / 8 = 6.25 + 1 = 7.25

The Ravens are ranked 2nd in the league in offensive points, 2nd in overall yards, 5th in passing yards, 5th in rushing yards, etc.

Here's how the they all turned out:

1. Baltimore: 7.25

2. Denver: 7.625

3. New York Giants: 8.25

4. New Orleans: 8.75

5. New England: 10.875

6. Indianapolis: 11.75

8. New York Jets: 12.125

9. Minnesota: 12.375

10. Dallas: 15.75

11. Pittsburgh: 16.5

12. Seattle: 16.5

13. San Diego: 16.625

14. Chicago: 18

15. San Francisco: 18.625

16. Cincinnati: 18.75

17. Green Bay: 19.125

18. Washington: 19.25

19. Tennessee: 20

20. Arizona: 21.875

21. Atlanta: 22.125

22. Jacksonville: 22.125

23. Miami: 22.25

24. Buffalo: 23

25. Houston: 25

26. Detroit: 26.5

27. Oakland: 27.125

28. Carolina: 27.25

29. Kansas City: 27.875

30. Tampa Bay: 30

31. St. Louis: 30.125

32. Cleveland: 32.875

Category: NFL

Since: Oct 6, 2009
Posted on: October 6, 2009 6:09 pm

## My Mathematically Correct Power Rankings

This doesn't take into account the teams they have played. Pretty good, but not perfect.

Since: Aug 14, 2009
Posted on: October 6, 2009 8:15 am

## My Mathematically Correct Power Rankings

Your reliance on ranks is a plus over opinion, but ranks suffer from a basic problem:

You actually have real yards, etc.. so that the gap between rank 4 vs. 5 may be way smaller than between 1 vs. 2.
Why give up that precision?

You could convert the various metrics into Z scores, so as a common metric, they add.
(score-mean)/standard dev

That way, the different ranges are scaled to match.
In addition, those metrics are not all equally related to either points or winning...
tho you "weight" them equally.  How do you account for turnovers?

I get something more like this, using multiple regression...

NOR  17
DEN  15
NYG  13
BAL  11
PHI  10
IND  10
MIN   8
SNF   7
PAT   6
GRB   3

(the # scale is "points better than an average team")

Glad to see other math models, hope to confer down the road.

Since: Oct 14, 2006
Posted on: October 4, 2009 10:42 pm

## My Mathematically Correct Power Rankings

We all know how terrible a computer formula system has worked in the BCS, but I've seen the case made before that using a BSC-esque system would work better in the NFL because of the longer season and smaller number of teams. This means more common opponents and a better guage of who is really better than who.

I think someone should sit down and come up with a serious formula that provides a BCS ranking for the NFL teams. I would want to see how those rankings after the regular season would compare to the playoff results.

Since: Sep 3, 2006
Posted on: October 4, 2009 10:35 pm

## My Mathematically Correct Power Rankings

HAHAHAHA!!!!, Man Big Guns, you made me laugh!!  In a good way!!!  I needed that after a tough day.  100000000!!!

(Offense(75)+Defense(90) +Strength of sched(85))/3 = 83.33=  Still got a chance at the super bowl whew hooooo!!!

Math and Football yay!!!!  Uhm was this after the Raven's lost?  They diserve some consideration for bad calls too lol.

Giants  (90)+(80)+(70)=240/3=80!!!  Jets are better man!!

Since: Aug 31, 2008
Posted on: October 4, 2009 10:17 pm

## My Mathematically Correct Power Rankings

E=mc²
= 0.111 x 300,000,000 x 300,000,000
= 10,000,000,000,000,000 Colts Win!

Since: Sep 3, 2006
Posted on: October 4, 2009 9:53 pm

## Mathematical??? Nope

Math starts with rigorous proof.  Your formula is based on a theory about how teams should be ranked, which in turn is your opinion.  People who have degrees in math(me) hear this crap all the time.  Formula's must be proven.  Then they are placed in the back of a book for future use.   You can use the quadratic formula all you want, but it takes pages to prove it for all cases.  You should call it a physics ranking system.  Lot's of exceptions to the rules in physics.

Since: Apr 9, 2007
Posted on: October 4, 2009 8:45 pm

## My Mathematically Correct Power Rankings

My thoughts exactly.  You have to factor in strength of schedule.

Since: Dec 21, 2008
Posted on: October 4, 2009 7:26 pm

## The problem is strength of opponent

The numbers are not the be all and end all.  If you played soft teams in your first 3 games it would skew the results.  Overall it looks pretty good and as the season goes on it will get more accurate.  Your #1 was not a winner today and while Miami and Buffalo were almost tied in your numbers they were not close on the actual field of play.   Denver played three weak teams, that resulted in bad numbers, there is not a chance in H they are the second best team in the NFL.

Since: Jun 28, 2007
Posted on: October 4, 2009 12:53 pm

## My Mathematically Correct Power Rankings

oh and you could also add turnover  differential in there..

Since: Jun 28, 2007
Posted on: October 4, 2009 12:49 pm

## Great Formula but...

I think that  this is a real creative way to rank teams without the bias that  is subtly( and sometimes blatant *cough cowboys lol)  placed in the media but shouldn't there be a way that are well rounded more(in this case less) points than teams that are Great in one category and average or above average in other categories??? just a suggestion

The views expressed in this blog are solely those of the author and do not reflect the views of CBS Sports or CBSSports.com