Sorry but the ppp polling performance is mixed:
The last ppp poll was good (
% between Obama and romney) in:
CO, FL, VA, MN, MI, PA, Michigan, NM and OR
but was wrong (>3%) in NV, IA, NH, AZ, MT, WI, WA, ME, MA and CT and sometimes very wrong.
For the most important state (OH), PPP was off by 3, giving Obama+5.
For NC, where PPP is based, PPP was off by 2, giving a tie when everybody knew that Romney winning.
And I don't speak about the Missouri senate race... last poll: akin losing by 4. Final result: losing by 15 !
So the conclusion is that PPP is sometimes pretty good, sometimes pretty bad and it's stupid to apply a D or R house effect.