First, the BCS was not all computer formulas. Human polls were used as well.
Second, "computer formulas" do not generate themselves. A human being designs and constructs them using subjective value judgments to decide what factors to employ and how to weigh those factors relative to one another. The fact a computer does the math does not make it reliable, valid or objective. The only true positive about models is that they are internally consistent in producing the results for every team. Consistency and true accuracy are very different things.
Third, it's rather silly to think that any system, polls, computer models, a combination of the two or a selection committee can ever approach perfection. It's also fairly easy to see that people will support whatever particular method happens to coincidentally produce the results with which they agree.
The only truly "fair" way to decide would be impossible in practice. Every team can't play every other team so that "last man standing" is the indisputable champ.
The problem now is quite simply that 4 teams is not enough most years to include every team for which someone can make a plausible argument that it's worthy.
Having people who know the sport well, have watched all the games involving potential teams and are committed to selecting the "best" teams free from cheerleading interests is the smartest way to decide. Stats and models are at most somewhat useful in making very close calls.
There will always be a cut-off and a last team out, and people who support that team or its conference will always be unhappy.
If that team questionably 8th-9th (or higher) its exclusion will be more easily justified than if the question is whether it's 4th-5th.
