I think there may be a problem with the way you calculate you power rankings. By subtracting potential winning percentage from actual winning percentage you are devaluing successful teams (or rather penalizing them for success).
Contrived example (4-team league):
Team A is 4 - 0, 1.000, PR 12 - 0, 1.000
Team B is 2 - 2, 0.500, PR 10 - 2, 0.833
A = 0.000
B = -0.333
A is clearly the better team and should have a lower result (higher P.R.).
The stat that I subtract the potential from the acutal is not for me "Power Ranking" stat it is for my "Luck Factor or Achivement Stat", I may have been unclear up above. The "Luck Factor" stat really just shows if teams are over or under achieving. Your example is perfect for this. The team that is 4-0/12-0 has a 0.000 luck factor rating because their actual performance matches there potential performance. While Team B has a luck rating of -.333 becuase they are under achieving or their acutal perfomance is less than their potential.
For the power rating I just use the potential performance number.