terps while technically its possible for a team who went 10-0 vs the top 100 to go 0-10 in their next 10 games, its just not probable. seeding must be done with probability of success in mind. technically the best team in the country could be 0-30. how do you know in a series of 1000 games, they wouldnt go 970-30, while the 30-0 team goes 30-970? you dont know, but come on, what are the odds of that? you have to make the decision based on the information you do have.
of course, SOS should count for something, but isnt the real example here 8-1 vs 8-4? clearly 8-1 is better, theres no question there. its not like a 2-0 vs 10-2, saying well the first guy is 100% WP and the second is, whatever, 84% or so. that 10-2 is better, because with a larger body of work, you are more confident that is a really good team. but when you have equal wins and a much better WP, there is no argument, it seems to me. that SOS would give you the edge if it was close - but 8-4 and 8-1 simply are not that close.