Really interesting list, but some of it is 'luck of the draw' kind of stuff in terms of using RPI - you can have the exact same record year over year, and even play the exact same teams, and have your RPI number skyrocket with opponents and their opponents winning one or two more games over the course of 25-30 games. That is especially true because you aren't looking at the raw number but the relative number compared to other teams so moving from 320 to 150 is not a significant change in raw number, just moving over a bunch of other low numbers.
And when you look at a team like Miss State - they are clearly a better team, but a big driver of their improved RPI is at the back end of their schedule: 2016 101-200 RPI 11 games vs 7, 201+ RPI 6 vs. 4, and on the plus side 51-100 RPI 3 vs. 7. I assuming you are using the full year RPI so they actually have a more RPI 1-50 games last year 16 vs 15 (Chattanooga #42, MichSt #18, and Uconn #1 were NCAA games) In fact in just the OOC it is a pretty identical year over year in terms of top 50 - TX both years (L in 2016, W in 2017) and wins USF #20 in 2016 vs Oregon #35 and ArkLR #44 It is harder for them to move up from RPI 20 - RPI 5 just because they are moving above teams that already have relatively strong RPI numbers, but like MD having such a bad RPI is driven by their scheduling and not by the quality of their team, the shift of games in OOC from 200+ RPI to 50-100 RPI is probably driving most of Miss States move.
