We tend to think of RPI as strictly an ordinal ranking system, mainly because that is how the NCAA chooses to display it.
By that I mean it feels like the difference between #8 and #9 is the same as the difference between #9 and #10.
in actually, RPI is a score mathematically calculated by each team, and multiple teams can be in a tie. And statically, a bunch of teams can be meaningfully tied because this scores have only so much statistical significance.
this year, there is a large score separattion between #8 and #9 and again between #15 and #16. But the teams from #9 through #15 have no separation; they are close to tied.
this is what allowed multiple teams who did not even play to jump around the ratings a bunch the last couple of days, they were close enough that the butterfly effect of a former opponent’s opponent could effect their ordinal rankings.
so assuming the selection committee noted this fact, it allowed them to decide that other factors are more important than RPI for this particular grouping.
View attachment 1524315
I’m not sure what factor they chose to rule WF out, and I probably do not agree with it. But they put in an outright conference champion (UMass) over a shared conference champion in Wake. I guess that was it, but who knows?
And we can assume that the committee put weight on the UNC team gaining three key players and faring better later in the season. They put UNC at the top of all the ”tied“ teams.
with that said, we should not be the 8 seed over dook. the separation that dook earned should have put them 8th, dropping us to 10th. Not that playing at NW does us any favors.
