I agree with @PadrosWindup below that what we have now is probably as good as it gets. There's just no great (and certainly no "optimal") answer to the problem. You've got way too many teams involved with not enough games to think that there's some magic set of objective, mathematical criteria is "right." Have people take a look at any computer rankings and many are going to say "how the hell is XYZ State ranked above ABC U.?!?!" For any 2 teams from different conferences, there's likely zero overlap in their schedules. Heck, even within conferences now that they've grown stupid big, you can take 2 teams and they may not have much overlap. It's pretty hard to compare teams in that circumstance. The committee can make use of whatever objective criteria/rankings they want (and I'm sure they do), but at the end of the day choosing the "best" or "most deserving" teams and ordering them is a task that involves opinions. Any individual algortimic ranking also kinda involves opinion...it's opinion applied the same way across all teams, but there wouldn't be so many computer rankings if there weren't varying opinions on what's valuable in ranking a team and how an algorithm gets built.