Sagrin Ratings - Through Week 5

[UNH Fanboi]

1. Georgia Southern
2. Wofford
3. Northern Iowa
4. James Madison
5. North Dakota State
6. New Hampshire
7. Montana State
8. Maine
9. Appalachian State
10. Furman
11. Delaware
12. Sam Houston State
13. Richmond
14. Montana
15. Lehigh
16. Harvard
17. William & Mary
18. McNeese State
19. South Alabama
20. Southern Illinois
21. Indiana State
22. Southern Utah
23. Massachusetts
24. Weber State
25. Youngstown State

[danefan]

I never can understand why the Ivy's get so inflated by computer polls?

[Bogus Megapardus]

Harvard is way too high. Johnnies lost soundly to Holy Cross. Beating up on my hapless, apocalyptically-inept Pards doesn't make you good.

[UAalum72]

Sagarin still using starting rankings. By next week all teams should be connected, then we'll see how it looks.

[bojeta]

I never can understand why the Ivy's get so inflated by computer polls?

My guess would be.... Don't play anyone outside your conference and therefore you don't rack up losses to FBS schools etc.

[danefan]

My guess would be.... Don't play anyone outside your conference and therefore you don't rack up losses to FBS schools etc.

But how does that account for strength of schedule? I mean Harvard lost to Holy Cross. HC is a very good team, but they don't even get into Sagarin's Top 25, yet the team they beat is 16th?

[youwouldno]

The "rating" number is a combination of two separate measures- one only looks at wins and losses, the other only looks at score. The latter is far more accurate. All the metrics become much more accurate as the season progresses.

Because the Ivies play a limited range of out-of-conference opponents, they are more likely to be either over- or under-rated as a group. With respect to Harvard, the ratings don't weight one game more than any other, so a reasonably close loss often doesn't do too much damage if the opponent is decent.

[danefan]

The "rating" number is a combination of two separate measures- one only looks at wins and losses, the other only looks at score. The latter is far more accurate. All the metrics become much more accurate as the season progresses.

Because the Ivies play a limited range of out-of-conference opponents, they are more likely to be either over- or under-rated as a group. With respect to Harvard, the ratings don't weight one game more than any other, so a reasonably close loss often doesn't do too much damage if the opponent is decent.

So they do or do not account for strength of schedule?

[youwouldno]

So they do or do not account for strength of schedule?

Sure, I mean the formula is basically (Strength of Opponent) X (Margin of win/loss vs. Opponent). A team's overall rating is the addition of that formula for every game they've played.

Beating a good team by 10 is far better than beating a weak team by 10- in the latter case the winner's rating could even go down.

[Squealofthepig]

So they do or do not account for strength of schedule?

Indirectly, through scores; as scores are tallied between more and more teams, the relative scores hold a bit more weight (though obviously this has statistical limitations, especially early in the season).

On the Ivies, one additional item that may skew their relative computer love is that several (like Penn) don't even play until the third game of the year, so the few games they play are more heavily weighted (in effect, due to a smaller n).

[FurmanWins!!]

2011 NCAA College Football Jeff Sagarin computer ratings by conference:

SoCon is the top rated league as of now
http://www.usatoday.com/sports/sagarin/fbc11.htm

[Squealofthepig]

2011 NCAA College Football Jeff Sagarin computer ratings by conference:

SoCon is the top rated league as of now
http://www.usatoday.com/sports/sagarin/fbc11.htm

Can't quibble with that - amazing how close the top 5 conference rankings are.