The following is a response to a friend who wondered about my reaction to this. It's germane to subjects that have dominated the blog for the last several months so it's worth sharing, keeping in mind that it's more free-wheeling than I generally (try to) allow myself to be here. I've added a few relevant links.
So Silver announces the launch of 538's statistical models for the 2016 general election. At the time of launch they give Trump a 20% chance. As new polls come out, the odds will change. Unless you're trading on the prediction markets, it's just a dog and pony show.
Funny that it goes live following several weeks of a Trump polling slump. Where was he a month ago when Trump had a slight edge over Hillary in aggregate? That plus economic conditions, which Silver says increase Trump's odds relative to polling alone, suggest his model may have had Trump as the odds-on favorite then [the model picks up almost three weeks after Trump had the narrow lead--at that point it shows Hillary 66%, Trump 34%].
Nate Silver was spectacularly wrong about the primaries. It's lame that he tries to save face by saying his poll model predicted 52 of 57 states or whatever. BFD, so did the RCP's averaging of recent state polls (minus the unreliable ones from dubious sources) prior to a primary taking place. Silver's being prudent here, copying what the betting markets show and then dressing it up quantitatively. And for the record, there were people--not just amateurs like yours truly, but also significant media figures like Scott Adams--who got it right.
Also, to say Trump underperformed polling by looking at state wins is misleading. He underperformed early and then outperformed late. In Arizona and then in every state after Wisconsin he blew the polls out of the water, even when the race was still on (in NY and the 5 NE states + Indiana). I have no idea how that'll translate to the general election, but it's worth pointing out that, based on polling averages, Remain was expected to win by 2 points. Instead, it lost by the same.
That said, it seems plausible, I guess--as plausible as the local news station's meteorologist making a weather forecast for the Thursday four weeks from tomorrow. None of the debates, which will be the most watched in history, have happened. Conventions haven't happened. VP nominations haven't happened. Trump will blow the sub-$2 million fundraising figure out of the water in June. He's soliciting hard.
On the other hand, as I've said all along, the demographics just keep getting worse for Republicans. They got crushed in '08 (which was predictable enough) but then lost pretty bad in '12, which should've been an easy win based on traditional metrics (basically everything was against Obama except for the fact that he was an incumbent). A year ago the markets had it something like 60%-40% that the Democrats would win the White House in '16. I don't think my prediction that we may never see another Republican president elected in our lifetimes is hyperbolic.
It is nice to see Silver putting pressure on what I've been saying for awhile about the electoral college. Trump's national polling deficit relative to '12 is relegated almost exclusively to the deep red central corridor states where he's going to win anyway. If his margins of victories in Utah, Texas, Kansas, and Idaho are in the single-digits instead of the double-digits but he wins Pennsylvania, Florida, and Ohio, it's worth it.
Just in time, too, Quinnipiac--which is one of the best--comes out showing Hillary with a 2 point lead nationally, while Rasmussen, the only one included in the current RCP average that is of likely voters (always better than registered voters) has Trump up by 4.
Trump has a problem getting traditional Republicans on board. But Hillary has a problem motivating traditional Democrats, who aren't excited at all about her.
Again, a month ago Trump was slightly ahead. With things swinging that much month from month, it seems foolhardy to use current polling to confidently project a November outcome. The Gravis one is "weighted by anticipated voting demographics". Is Gravis anticipating a 10%+ drop in black turnout?
Parenthetically, a Hillary presidency still makes the outcome of the '16 election cycle better than I would've conceived of a year ago.
To have someone as nakedly corrupt and devoid of any notable achievements throughout the course of decades in 'public service', running on keeping the immigrant floodgates open and disarming the American public while kowtowing to racialist thugs like Black Lives Matter, beat a man who not only refused to apologize for considering the interests of white middle Americans but even had the gall to advocate in their favor at the expense of those outside the inner concentric circles of national loyalty--well, it's difficult to think of much that could do more to push us towards the long-term goal of secession and the disintegration of the empire.
Anything that causes Middle America--and white men specifically--to withdrawal consent from an American empire that plunders the fruits of their labor while treating them as moral reprobate for being white men is a good thing.
Five years ago Brexit would've been laughed off as a pipe dream. Yes, the new government will try to gum it up so that it doesn't happen, but no one said the globalists would go down without a fight!
The best line of attack is to have state governments simply refuse to enforce federal laws. The tenth amendment center does yeoman's work on this front. They pushed for Utah to refuse to provide water for the federal government's NSA facility in the state (it needs a bazillion gallons a day to cool all the servers). The federal government doesn't have the apparatus to run everything--anything, really--so if the anti-federalist push comes from the state level, where it is doable, then we're talking about something more than just individual preppers standing up to leviathan.
The federal government has the military, yes, but history is replete with examples of some central authority ordering its military to do X, Y, and Z when that military, sympathizing with its supposed targets, ends up flipping sides.