Nate Silver: Dear Media, Stop Freaking Out About Donald Trump’s Polls

[Bull Moose Base]

If no one were paying attention, the polls wouldn't be changing at all. Also, people who pay attention earlier logically seem like more likely voters. And Silver has himself in the past analyzed that early poll leaders in GOP presidential primaries have tended to win them. Of course, until 2012, that correlated with establishment support. But even in 2012, Romney maintained a stable polling lead in NH equivalent to Trump's, only losing it briefly when Gingrich surged about a month before the primary before quickly collapsing. So I buy that polls can fluctuate wildly up until voting but not that a candidate has no advantage over a rival he leads by 20 points.

[Bull Moose Base]

http://www.realclearpolitics.com/articles/2015/11/24/early-state_polls_arent_predictive_--_yet__128828.html

Here's another analysis based on the last 3 open primaries- both parties in 08, GOP in 12- that suggests IA and NH polls become much more predictive after Thanksgiving. But as it points out, since the elections are a month later this time, maybe after Christmas would align more with those past examples.

As far as analogies to last time, I'd go roughly with Trump=Gingrich, Cruz=Santorum, Carson=Cain, Rubio=Romney? (though if Bush, Kasich or even Fiorina or Christie beat Rubio in New Hampshire, they could snatch that role away from him).

Anyway, I don't think it makes sense to assume Trump would lose a one-on-one primary race. The last PPP for example showed him beating Rubio one-on-one. It remains an open question how much of his own money Trump will spend, but if he chooses to, he could compete with Rubio (or other establishment candidate) in a way that Gingrich could not.