Nate Silver: 2016 is a tossup; most conventional wisdom analysis is flimsy

[Mister Mets]

http://www.nytimes.com/2011/11/06/magazine/nate-silver-handicaps-2012-election.html

Is Obama Toast? Handicapping the 2012 Election
by Nate Silver

NOV. 3, 2011

You must be logged in to read this quote.

You must be logged in to read this quote.

You must be logged in to read this quote.

You must be logged in to read this quote.

Nate does a fine job of adding 10 polls together and dividing by 10 to find the average, but his analysis is no better than the wish-wish media pundits he claims to hate. "This election is going to be an epic Florida 2000 tossup! Please subscribe!"

I'm outraged that an year before the election, he suggested that the guy who went on to win by four whole points was a modest underdog.

[Mister Mets]

I don't see how it's a toss up when all the Democrats need are the Kerry states + VA, NM, NV.  

New Mexico and Nevada are practically freebies at this point.   That just leaves Virginia, which is probably one of the most obvious trending states in the country.    Even if the national vote shifts 3% toward the GOP from 2012, Virginia would still be winnable.  

It's just not a workable map for the Republicans.   They NEEEEEED a realignment.

There tend to be large swings in elections.

From 1972 to 1976, the party that won by 23 points lost by two. So a lot of states swing hard.

From 1976 to 1980, the party that won by two lost by 9.8, and that was reflected in the states. Ohio, which Ford narrowly lost by less than a percent, when to Reagan by ten points. North Carolina, which Ford lost by 11 points, went to Reagan by two.

1988 was a good presidential cycle for Republicans, but Bush lost a lot of ground from Reagan's reelection. Reagan won Wisconsin by nine points. Bush lost it by 3.62.

Bush loses Wisconsin by half a point in 2004. Obama wins it by nearly 14 points in 2008. Bush wins Indiana by 20 points in 2004. Obama wins it in 2008.

States that got by more than ten points to a party aren't guaranteed for that party in the next cycle.

[Mister Mets]

Clarifying what the polls are saying is a reasonably valuable thing to do, and was pretty low-hanging fruit in 2008 and 2012. But it's low hanging fruit because it's easy.

Has Nate Silver ever made actual hard predictions about the outcome of races? More than that, has he made predictions well in advance of the actual races? Has anybody gone back and run an analysis on this?

He has two conflicting messages.

First, he was credited with predicting the outcomes of elections by state in 2012.

However, he focuses on odds and ambiguity rather than certainty. It's an admirable position, given how many blowhards make promises without caveats which end up being wrong. But it contradicts the way he gets credit.