PPP vs. Nate Silver and potentially cooked polls

[Beet]

Nate Silver wasnt the most accurate aggregator in 2012; Josh Putnam, Hugh Jackman, and Drew Linzer were all more accurate according to an analysis of Brier scores. He just got the most attention because his NY Times soapbox got him a lot of sh**t from conservatives. And it's not that he had some secret sauce that allowed him to be right on all 50 states; it's that he applied rigorous probability theory to generate precise forecasts. Him getting all 50 states right was the cherry on top. But it's true that if you believe what he was saying the whole campaign, even if he judged the probability right, he still could have ended up "predicting" wrong. A coin that has a 5 percent chance of landing heads will still land heads one in 20 times; it doesn't become impossible. Besides, even if the Florida poll was cooked, what does Nate Silver have to do with it? Garbage in, garbage out. His weighting of the various pollsters seemed to based on objective performance.

To me, the meat of this article is the criticism/accusation thrown at PPP is that they don't scientifically determine the white share of the vote. But the main evidence presented is that the white vote share decreased from 69 to 65 percent. To me it certainly seems unlikely that such a thing could have happened IRL; especially as you would have expected Romney's supporters, predominantly white, to have become more energized/likely to vote after the first debate. But pollsters arent supposed to make any assumptions so... this is circumstantial evidence at best. Is there a rule that says the demographic weightings can't change by more than x percent from poll to poll?

[Beet]

Nate Cohn offers more criticism of PPP here:

http://www.newrepublic.com/article/114769/ppp-methodology-results-arent-defense

essentially arguing that it looks like PPP uses ad hoc weighting to get their results closer to the polling average, in cases where other pollsters have already polled the race.

I've had this thought before.

1. Start a fake polling company.
2. Release bogus polls that simply report results close to averages of previously published polls.
3. Be right most of the time (and when you're wrong, everyone else is too).
4. Establish a great reputation as the most accurate pollster.

(Leap of faith)

5. Make money.

This is pretty much how Wall Street works. If everyone's committing the same kind of fraud, then the aggregate results are nobody's "fault."