You could point to the collapse of the ACC on Sunday and say the committee got the seeding very wrong this year. But if the committee was wrong, so was just about everyone else.

What the seeding process doesn't do well is take into account teams gelling late in the season. So two "hot" teams get #7 seeds and knock off #2's. If those two games go the other way on Sunday, 14 of the top 16 seeds would have advanced. Three P6 conferences would have had 3 teams each, the other three would have had 2 each, and #1 seed Gonzaga rounds out the field.

We, and the talking heads on ESPN and Fox, were all convinced the ACC was clearly the best conference this year. We, and they, and the selection committee were all looking at most of the same "metrics" to reach that conclusion. So, the question is were the metrics flawed, or did we just have two flukes in Louisville-Michigan and Duke-South Carolina that shattered our perception?

With conference expansion, there is very little room for significant inter-league play after New Year's. Whatever impressions were formed before that tend to dictate our analysis of what happens after that. And often, those impressions are based on the play of teams that barely resemble what they will have become by March. But the computer models can't factor in improvement over time. To them, a win in November counts as much as one in February.

Maybe that needs to change.