SRS: With Pittsburgh Steelers Loss, a New Team Is at the Top!

Wow, did one of my predictions completely fall apart.  Seattle at Oakland being close?  Ouch!  I did get my lock and upset picks correct, so two out of three ain’t bad!  SRS would like to welcome Kansas City and Miami to the top 10 list!  Since there were some questions on my college article, I’ll answer them partly here:  Quality wins are determined by being worth 10 percent or more of the No. 1 team’s total points.

How are these rankings formed?  Check out the bottom of the article.

R	Team	Record	LW	Result	Quality Wins	Losses
10.	Kansas City	(5-2)	--	W vs. Buffalo	@Cleveland	@Indianapolis @Houston
9.	New York Giants	(5-2)	9	BYE	@Houston	Tennessee @Indianapolis
7.	Atlanta	(5-2)	8	BYE	@Cleveland	@Philadelphia @Pittsburgh
7.	Miami	(4-3)	--	W @ Cincinnati	@Cincinnati	Jets New England Pittsburgh
6.	New York Jets	(5-2)	7	L vs. Green Bay	None	Baltimore Green Bay
5.	Houston	(4-3)	4	L @ Indianapolis	Indianapolis	Dallas Giants @Indianapolis
4.	Indianapolis	(5-2)	5	W vs. Houston	None	@Jacksonville @Houston
3.	Pittsburgh	(5-2)	1	L @ New Orleans	@Tennessee	Baltimore @New Orleans
2.	New England	(6-1)	6	W vs. Minnesota	Baltimore	@Jets
1.	Baltimore	(5-2)	2	BYE	@Pittsburgh @Jets	@Cincinnati @New England

Lock (1-0)

Green Bay over Dallas

Great Game (Averaging 30 point difference—sure hope that goes down)

Arizona @ Minnesota

Upset (1-0)

San Diego @ Houston

How SRS Is Created

The whole point of me starting a ranking system 15 years ago was to remove any and all bias for college football (that’s what they all say).  So I decided to try my list with the pros as well. 

Full disclosure: My favorite team is the St. Louis Rams. 

I've tried to come up with the rationale that every sports announcer I've ever heard brings up.  Conference rankings (check), road vs. away (check) and who did you play (double-check).  As I said before, I've been doing this privately for over 15 years, and every year there have been slight changes to it. 

Usually by the end of the season I see more improvements that I can make and I apply them to the following season.  By no stretch is it perfect, but I think it’s getting there.

So what are the specifics?  It starts by generating a ranking for each division.  This is done by generating a win-loss record for each division against the other divisions.  This is modified by determining who those wins are against.

This is mostly because of college football's big-time teams tending to play cream puffs to build up their record, so I'm going to use a college football example:  The Big Ten loses some of its power by scheduling a lot of MAC Teams.  I then take a ranking of that value. 

As of 10/12/10, the divisional rankings are as follows:

1. AFC North (LW: 1)
2. AFC South (4)
3. NFC East (2)
4. AFC East (5)
5. AFC West (3)
6. NFC North (7)
7. NFC South (4)
8. NFC West (8)

After determining the division rankings, I generate values that a team would receive for playing another team.  Values are generated for a home win, a home loss, a road win and a road loss.  Each of the values are derived from the home win, which is formed by taking the team's place in the division (first, second, third, etc.), inverting it so more points are awarded for beating the better team and multiplying it by the inverse conference ranking. 

For example, the first place team in the first place division is Baltimore.  If a team beats Baltimore on the road, it will earn 64 points.  This is because Baltimore is in first place (inverse = four multiplied by inverse conference ranking = eight, and multiplied by two because it would be on the road, eight times four times two equals 64.)

Losses are similar, though there is no inverting.  So if a team loses, it will only cause them to lose one point (place in conference = one, conference rank = one, one times one equals one).  A home win is the same as a road win, except the value is not multiplied by two, and a home loss takes the road loss and multiplies it by two.