Nate Silver releases his 2016 U.S. Election Forecast

Post Reply New Topic

http://projects.fivethirtyeight.com/201 ... -forecast/

The local bookies are giving 3 to 1 odds favoring Ms. Clinton, as well.

A closer examination of FiveThirtyEight's estimates also shows that it will be significantly easier for Clinton to focus her campaign than Trump...

In the 11 states with the highest projected margin (20 percentage points or more), 8 are Democratic-led and only 3 are Republican-led.

In the 11 states with the smallest projected margin (5 percentage points or less), 6 are Republican-led and 5 are Democrat-led.

In other words: The Democrats have more "safe" states than the Republicans, meaning that they can focus their resources on crucial battleground states like Ohio and Florida.

Another state that one ought to pay attention to is Utah, as it may yield a surprise in this election.

It is a traditionally GOP state (Utah has voted Republican in presidential elections since 1968). But it is also majority Mormon, and the Mormons *loathe* Donald Trump. Although not the most likely scenario, it wouldn't be too far fetched for Gary Johnson to act as a spoiler for Trump in Utah, thus letting Clinton win the state.

In fact, the latest poll from early June in Utah actually has Clinton beating Trump by a small margin, largely because of Gary Johnson.
http://projects.fivethirtyeight.com/201 ... cast/utah/

I don't know what's happening in the US but in the UK polling organisations have quite a bit of egg on their face. The vast majority were wrong about the last general election and the recent referendum on EU membership. If the polls say Clinton will win, might be time to place a large bet on the Trumpenführer.

Oh, thank the Lord! Even still with such a good sized winning margin Silver gives to Clinton, I have to think America will just barely miss the bullet.

You guys still believe the baseball nerd Nate Silver still has any credibility after these primaries? That's cute, not the oracle you/he thought he was.


wont post this one http://i.imgur.com/dpPrnau.jpg?1



polling is only as good as its methodology, sampling, and neutrality

Jacoby,

No one's perfect, and the 2016 presidential primaries were weird by any standard.

In addition, Silver has been entirely forthcoming about his wrong prediction:

http://fivethirtyeight.com/features/how ... ald-trump/

I think Nate Silver's still got it.

I don't remember Silver's exact formula but I am pretty sure he uses an aggregate of polls so his predictions are only as good as the polls and the polls have completely discredited this year on both sides of the aisle and the Atlantic. Pollsters have this idea that voters are going behave like they did in 2012 or like in some other typical US election and they're clearly not, America has changed and these are just wild guesses at this point.

Trump has a better chance of winning Massachusetts than Hillary has winning Utah, it's not gonna happen.

I don't believe Gary Johnson and Jill Stein will get over 5% of the vote combined just because people are going to want their vote to count and neither are going to qualify for the debates. Hillary is losing independents big and a lot of these independents that these polls are saying are going to vote third party won't.

Honestly the polling business isn't a good one to get into with the way the globe is right now, polling works when there is stability and thus predictability. I don't think we're going to be seeing any calm seas for a long time now.

Trump was ahead in the polls of Republican voters from his announcement until Iowa. He had already said things on a daily basis would have ended candidacies of any other canididate in any other year. Once the primary season got started in earnist his voting percentage steadly went up and he was still saying supposidly candidacy destroying things on a daily basis.

So I think the mistakes were not so much that the data was wrong was that the data was read through then lens of conventinal wisdom. The biggest one was that a candidate with a populist us vs them appeal and appeal to bullying nature could not win in America even in a time of social and economic upheavel. The statistics show the nation has been in recovery since 2010, the polls show people in thier daily lives are not feeling it. Politicians with similar appeals to Trump in times of large social and economic change have won in every era in every region of the world but somehow this was not supposed to happen in America because it has not happened to this degree dispite a populist streak in the electorate that has always been there. The closest we have had to this in the last 100 years was Wallace in '68 and maybe Perot in '92. Polls have shown the level and intensity of dissatisfactions has grown since the bailouts and the non recovery. We are in a internet driven world where a story from an hour ago might as well have happened a year ago. The bailouts were eight years ago the idea that people are still PO'd by what that represented was beyond the pundits, after all the same anger and fury was around in 2012 but to put it mildly a candidate to appeal/exploit that was not around.

Political correctness had to seem like a real tired old issue to the experts, conservatives have bern complaining about it since the late 1980's. Pollsters were not even asking about this. The creeping real world negative effects combined with changing demographics and people more and more bieng intimidated from saying the wrong thing caused this issue to stay undergound. The Supreme Court gay marriage decision light the smoldering embers and Trump poured gasoline on the fire so to speak.

Human action cannot be modeled but understanding and knowledge of human action is required for any model. The politicians, the pundits, the pollsters, the think tanks, big business, the coastal elitists, the 1% have all so isolated themselves from the rest of the country that it is no surprise that they cannot fathom the appeal of candidate like Donald Trump or the Brexiteers. The establishment advocates on their own behalf telling the rest of the country it is for their own good and that they're too stupid to control their own destinies, they don't care or understand the the consequences of their views/actions but only their own self-interest. Mass immigration might be great for big business and those that play identity politics, the same with free trade since it makes the rich people in this country richer, but on main street it is a totally different. Communities have been destroyed, our culture is under constant attack, our jobs are gone, our neighborhoods aren't safe, our schools are failing, our social safety net is overburden and failing, simply put they don't really care about us but rather just our money and our votes. With the rise of new media, the disconnect between the elite and the masses has come boiling on both sides of the aisle but this discontent has always been there simmering in the background.

Take Silver seriously. He is an expert as using Bayesian methods to make predictions.

If you have the mathematical cojones please see http://www.stat.cmu.edu/~genovese/talks/ipam-04.pdf, a primer on Bayesian inference.

FiveThirtyEight just released an interesting update to the forecast, comparing the 2016 polls in red, blue and purple (swing) states with the 2012 election results...





Source: http://fivethirtyeight.com/features/ele ... ame-thing/

As demonstrated above Clinton has a general edge over Trump wrt. red states, as several of these (Utah, Missouri and Arizona) now are polling much closer than in the 2012 election. She also seems to have an edge over Trump in swing states, as her polling margin has increased compared to Obama in 2012. Clinton is polling slightly worse than Obama in blue states, but here the Democratic margin remains at 6 percent or higher in all states.

Furthermore, the states expected to flip based on the above (Kansas, North Carolina and Nevada) would result in a net gain of 15 delegates for Clinton (Kansas and Nevada cancel each other out with 6 delegates each, while North Carolina adds 15 delegates to to the Democratic camp).

Also note - as previously mentioned - the massive erosion of the Republican margin in Utah of almost 45 percentage points from 2012 to the 2016 polls... almost 4 times as large as the worst development for the Democrats in Connecticut at 11.3 percentage points...

the polls 5 months out literally mean nothing especially now that they have the candidates under 40%, based on nothing.

they say what the pollster wants them to say, they're meaningless. totally changed the methodology and now there is no more primaries for them to wear egg on their face until November.

Silver's formula can't work if the polls don't work and they haven't been working for a while now

Yup. I can't believe how bad the reporting has been on the changes taking place in the American electorate. Trump has a lot of angry, poorly educated, white males on his side, but they come from both parties. It's a new coalition, and I don't think anyone has a very clear idea of exactly who is in it.

People are fed up. I don't think their thinking is particularly well informed, and I don't think they're blaming the right things, but we seem to have achieved a breaking point with regards to our tolerance for the status quo.

Essentially correct.

However, local bookies went from nearly even odds to about 5 to 3 favoring Ms. Clinton (down from 3 to 1). Of course, the odds seem to change every day; but they depend on which way the money goes, and not on any questions other than "Who ya bettin' on?" and "How much ya bettin'?"

I'll trust an honest bookie over a dishonest pollster any day!

the bookies thought Marco Rubio was going to be the nominee