Originally Posted by Pitt Gorilla
It's not that we don't understand them; it's that we wouldn't likely be able to reproduce them with 100% accuracy. However, it's not like his data points are great mysteries. We know (within an acceptable epsilon) what he's using and we can approximate his outcomes very accurately.
Yep. They are simple Monte Carlos with some adjustments. Nothing special. As many have mentioned, a model is only as good as its assumptions. Wednesday will (hopefully for our sake) let us know if the assumptions in the polls are close to the voting population.
On another note, not that any of you probably care, but I live in NJ and drove to Harrisburg, PA today for a Ryan rally. It was great. I'm ten minutes away from tomorrow night's Romney rally outside Philly. They are expecting tens of thousands. Pretty cool to be close to a sorta-swing-state. Just starting seeing the ads. Don't know how folks in OH, VA, CO, FL can stand it.