[M T]

I wish I understood RPI better.  How can UConn be 7 in RPI when it plays so many patsies? 

Strength of schedule is:
1 Virginia
2 South Carolina
3 Indiana
4 Maryland
5 Oregon
6 Oregon State
7 Ohio State
8 Baylor
9 Stanford
10 Duke
11 Iowa
12 UConn

I'm not sure where/when you got SoS from.  It looks suspect to me (as to you).
I'll use RealTimeRPI (These are SoS rankings, not their RPI rankings) as of Monday night.
I'm going to calculate something easier to calculate:  Winning percentage of opponents (including games against the team).  I don't claim this is better, just that it is easier to comprehend.  (Call it "SoSo".)
I will calculate two numbers: one for opponents in games already played, one for opponents on schedule (past or future). (If an opponent is played twice, its win-loss record is used twice.)
1. Virginia 72.2% (68.4%)
7. Oregon State  63.1% (64.4%)
10. Oregon 61.9% (64.6%)
12. So. Carolina 60.4% (63.1%)
16. Stanford 58.7% (62.7%)
30. UConn 55.3% (58.9%)
35. UCLA 57.8% (61.7%)

By however RealTimeRPI calculates SoS, it does appear that UConn is more highly ranked that this more simplistic estimation of opponents' strength.

Note that if you looked only at the PAC-10 conference games in 1995-1996 when Stanford went 18-0 in conference, the SoSo for Stanford calculated as I did would have been
    (conf games not won by Stanford)/ (2 teams/game * conf games) = (18*9)/(2* 18*10) = 45%
The in-conference SoSo for a hypothetical team that went 0-18 that year would be exactly 50%.
SoSo and, to a lesser extent, SoS are weak measures when a lot of games are played amongst your opponents.

[M T]

I wish I understood RPI better.  How can UConn be 7 in RPI when it plays so many patsies? 

Strength of schedule is:
1 Virginia
2 South Carolina
3 Indiana
4 Maryland
5 Oregon
6 Oregon State
7 Ohio State
8 Baylor
9 Stanford
10 Duke
11 Iowa
12 UConn

I'm not sure where/when you got SoS from.  It looks suspect to me (as to you).
I'll use RealTimeRPI (These are SoS rankings, not their RPI rankings) as of Monday night.
I'm going to calculate something easier to calculate:  Winning percentage of opponents (including games against the team).  I don't claim this is better, just that it is easier to comprehend.  (Call it "SoSo".)
I will calculate two numbers: one for opponents in games already played, one for opponents on schedule (past or future). (If an opponent is played twice, its win-loss record is used twice.)
1. Virginia 72.2% (68.4%)
7. Oregon State  63.1% (64.4%)
10. Oregon 61.9% (64.6%)
12. So. Carolina 60.4% (63.1%)
16. Stanford 58.7% (62.7%)
30. UConn 55.3% (58.9%)
35. UCLA 57.8% (61.7%)

By however RealTimeRPI calculates SoS, it does appear that UConn is more highly ranked that this more simplistic estimation of opponents' strength.

Note that if you looked only at the PAC-10 conference games in 1995-1996 when Stanford went 18-0 in conference, the SoSo for Stanford calculated as I did would have been
    (conf games not won by Stanford)/ (2 teams/game * conf games) = (18*9)/(2* 18*10) = 45%
The in-conference SoSo for a hypothetical team that went 0-18 that year would be exactly 50%.
SoSo and, to a lesser extent, SoS are weak measures when a lot of games are played amongst your opponents.