I agree with your post regarding the use of KRACH vs. RPI and about Duluth.
The only reason, IMO, that Duluth even deserved consideration was because of the way they played "down the stretch" compared to St. Cloud. But the argument by USCHO regarding Duluth and SCSU clearly indicates that they did not deserve to make the NCAA field.
Just because Duluth was playing better the last few weeks of the season, doesn't mean that they were the better OVERALL team. Duluth did in fact win more head-to-head matchups (4-3) this season, but it's not like they beat them all 7 times. Here's the season series between UMD and SCSU (without the 1st round of the WCHA playoffs):
11/22/2002, St. Cloud State 3 @ Minnesota-Duluth 2
11/23/2002, Minnesota-Duluth 4 @ St. Cloud State 2
01/24/2003. Minnesota-Duluth 2 @ St. Cloud State 3
01/24/2003, St. Cloud State 1 @ Minnesota-Duluth 3
Before the playoffs, they split the games, then Duluth won the playoff series 2-1. Does it logically make any sense for us to consider these games less than we do the games in March? No, it doesn't. Fact of the matter is that those teams playing well down the stretch deserve no more consideration than teams that play well at the beginning of the year.
The NCAA playoffs should be for those teams that play well ALL year, not just for part of it. And yes, Duluth played well all year (22-15-5), but their non-conference schedule included Notre Dame (twice), Lakehead, Bemidji State (twice), Union (twice) and Rensselaer (twice). In these games they went 5-3-1, so they didn't exactly "light 'em up" in non-conference play.