I knew it was that or the other. I just ran, after an hour of data prep, yay me and yay my life, a logistic regression, home away effects, with a very small penalty to make sure things converge. A bit out of whack as an 8-0 Holy Cross gets ranked 26th but overall on the thing I just ran I put UConn 117 nationally out of NCAA-1. Sagarin has us at 133 and Massey at 139.
I don't think we're bottom 25 like the other two would imply. I think they use last year's info as part of the calculation.
edit: I hand count 20 FCS schools ahead of UConn so that puts us 97 overall in whatever I did. This makes more sense to me.
Losses at #11 (UMich), vs #21 (Syr), #23 (NCSt), at #110 (UtSt), at #133 (them). All teams which are ranked higher if you factor in home field.
Wins vs 90 (FrSt), at #156 (FIU), vs #164 (BC), #259 (CCSU). Fresno state is the only win ranked higher than our own ranking even account for home-field.
----
The reason I don't produce a method is that I don't see any I like that "solve" for traditional problems. There are a couple of mathematical computing headaches involved.
Also the BCS always used methods that left analysis contained to the current season. The computer rankings are no longer obligated.
I also haven't found any score model with my worry. More philosophical issues on my part. I'd like to use a thing that models scores.
----
Sincerely,
Your local Ph.D. Statistician.