WMU was #73 in final RPI. UConn, at #36, was the lowest RPI team to get an at-large bid. NC State, at #34, was left out. So WMU had no chance. Winning one more game would not change RPI rank by very much. Coming back to UNC, I noticed that they played almost all their games at home. This was probably a significant factor in their being seeded below their RPI rank by the committee. FWIW, I believe the committee's choices are essentially random. They go into a room, they can't possibly hold all the relevant data in their heads, they have no particular expertise in statistics, many have no particular expertise in soccer (though this shouldn't even be relevant: do we want the committee making decisions based on their impressions of how good a team is?). So, they basically toss around ideas like, "yeah, but UNC played so many games at home!" and "Akron did beat Georgetown, and we think the Hoyas should get a top-4 seed" and the like, and then come to some consensus, and then go to lunch.