I've said for years I think you need to scale the teams better. There is practically no difference between a team ranked 90 and a team ranked 125. You also put way too much emphasis on every game. On any given schedule there are only about half to 9 games that truly matter i.e. games that could be realistically close (they all aren't always close and some are close that shouldn't be, but that is a different thing). I mean is anyone going to pay attention to Ohio State's game against Buffalo to start the year. The game just doesn't matter in the scheme of things (at least the 999,999 times out of a million that OSU does what it should). There is no way that game should have an equal bearing on Ohio State's ranking as the game at Michigan to close the year. You seem to have this notion that every game counts, which just isn't true. Only the losses and the good wins count, the **** game wins don't (unless virtually your whole schedule is **** games, call that the Boise State problem).
I think you could capture this if you just treated every teams worst 3 wins the same. Lets say Buffalo, Florida A&M and San Diego State end up as the three worst wins for Ohio State (they may they may not, but just for the sake of argument). Just treat them the exact same way you treat Colorado St., Georgia State, and Chattanooga (the likely 3 worst wins for Alabama). And on and on. Now sure this might pose problems when you start getting the crap teams at the end of the year, but the reality is no one really cares about rankings past the top 25-50 teams any way, so who really cares, especially for a ranking like yours that is supposed to be used for determining playoff teams.