- Conference champions
- Regular season conference champions if they didn't win the tournament
- Top non champions and non-conference schools, as selected by the sorting algorithm below
My sorting algorithm, used both to select the extra schools and for seeding and placement, would sort the schools based on the following:
- Record against other Division I NCAA Schools
- Overall record
- Overall point differential (i.e. points scored - points scored against)
- The full name of the school in standard order
However, the last criteria -- the school name -- would be ignored for selecting the top teams. Instead if there was a tie for the 64th slot, there would be play-in games utilizing these teams, and possibly the 63rd place team to select the last position.
Once the 64, or more, teams have been selected, they would be sorted using the algorithm above (this is where I needed a consistent, but arbitrary, tie breaker). Then the sorted list is grouped into sets of 4.
The four regions will be given a geographic location -- either the center of a geographic region, where each region has an equal number of division I schools, or the location of the regional final, I haven't decided which. Starting with the top four schools, each school will be sent to the available region that is closest to their location. This causes the 4th seed to become the top seed in the left over region. This process repeats with each group of 4 down the list, with the added adjustment that if there are more schools from the same conference in the region a school would otherwise go to, then in another region, the school will go to the next closest region. So, for example if we are placing the number 9 school -- who is from the Big 10 -- and they would normally go to the "East" region, but there is already a Big 10 school there, and the "Midwest" region doesn't have one yet, they would go to the "Midwest" region instead.
One effect of this is that there would be no drama. At any point schools will have a good idea of where they will end up based on how they finish the season. Of course I could see a coach knowing this choosing to run up the score in a late season game to get a better seeding -- which might be a bit of a problem. The other problem would be making sure that schools don't deliberately choose to play known weak teams just to enhance their record. This could, however, end up benefiting the weak teams by making them stronger.