I think with the lack of non-conference games you run into a fairness issue no matter what you do. The primary benefit of a computer is that it can be unbiased. Just like in the pros the criteria for getting a high postseason seed are known beforehand and everyone starts even. In college, teams create their own schedule and there is a vast discrepancy in SOS so it requires an algorithm that is more complex. But this year there is a problem with that.

Lets look at football because there is a more extreme example. The MAC, B10, and SEC have all played zero non conference games. Normally, the relative strength of each conference can be shown by its non-conference results. But if everyone starts equal then Buffalo's unbeaten record is approximately equal to Ohio State's right now. Most computers start with a preseason ranking that is based mostly on the results of recent years which is phased out and is dropped entirely half way through the year. There is just not enough data to do that this year and there might not be in basketball either. It looks like Army and Navy are the only Patriot League teams that intend to play non-conference games at all.

So you are left with applying the eye-test heavier than normal whereas resumes are usually more important. I have never been a fan of going too heavy on the eye-test. It tends to help the teams with a lot of five star recruits that always look good in trying to play even if they aren't playing all that well. IMO, if I'm unbeaten and I beat you then, barring some really extreme circumstance, I should be ranked higher than you no matter how I looked doing it.