I'm open for discussion - but I think any matchup changes will have to be for MUSDUC 3.
If the seeding method becomes more closely aligned with actual strength of the teams I'm not sure how being a 1 or a 2 seed would or should be any tougher than being a 5 or a 6 if you play all the other 1s or 2s and 5s or 6s. The idea is for the teams to face opponents of similar relative strength, ostensibly to determine which is the best conference that season - acknowledging of course that in a head to head single game situation anything can happen of course...
Here's this season's #1s, how they did in the MUSDUC, (closest games) and projection, rpi and sos. I don't know that this tells us anything though, since the seeding method used was the 2 season average which even I admit is a poor barometer of the caliber of the team this season. For the most part though teams played some close games that could have gone either way. I expect that as the seeding method zeroes on on something that more closely resembles actual team strength (assuming that we achieve that) this would be even more true. Maybe the seeding thing is not really a true measure and I'm off base, like I said I'm trying to consider everything, but I still think that in order to have the most overall "fair" tournament that like seeds should face like seeds, and that we should work to make the seedings as accurate as possible to that end.
TAMU, Commerce 1-4 (2 pt loss, 4 pt loss) 78 Proj, 75 RPI, 55 SOS
Grand Valley St 4-1 (1 pt win, 3 pt win) 7 Proj, 8 RPI, 24 SOS
Stonehill 2-3 (6 pt loss, 7 pt loss) 42 Proj, 42 RPI, 27 SOS
NC Central 4-1 (1 pt loss, 2 pt win) 3 Proj, 4 RPI, 23 SOS
Cal, Davis 3-2 (5 pt win, 3 pt loss) 45 Proj, 68 RPI, 75 SOS
Incarnate Word 2-3 (5 pt loss, 4 pt win) 67 Proj, 59 RPI, 36 SOS
I do see the compromise in rogelio's proposal, and I am intrigued by it some. Why isn't it mirrored though ABC and DEF? Does it not work out if that's the case? (I haven't parsed everything) ie, why not have the ABC #1 play the DEF 234 (they seem to have the 245 now)? If it is possible that seems fairer...