Posted on 06/29/2016 12:14:30 PM PDT by detective
“He’s not predicting a Hillary win in November. His totals are for his projection of what would happen if the election were today.”
That’s a little nuanced for me... What good is that sort of caveat on his projection? Just brews discontent, etc. (Which, of course, is the purpose...)
I don’t remember his models predicting Trump wins.
I don’t follow him. Did he publish the predictions of his models and reject them publicly?
Seems I would have noticed that.
Not at all to the point. Just an informative and very early aggregation of Trump’s hurdles. Doesn’t reference any results from his methodology at all.
It’s hard for me to believe he never used his methodology at any time during the primaries.
And there are strong reasons for him to protect it’s ‘reputation’.
Again, that is saying he never used any of the statistical analysis that is his very forte!
But he did, and does, no worse than most IMO.
Every pollster is getting predictions significantly left-leaning from the actual results these days.
I think issues now are just not so easily divided along the race-age-gender metrics pollsters use.
You’re joking, right?
Why does he have “the best methodology”?
His “methodology” has been consistently wrong about Trump. He said that Trump was unpopular and would not be taken seriously. He said that Trump would not get the Republican nomination.
The MSM calls Silver “freakishly accurate”. That is because he is one of their favorites.
Other posters have noted that his reputation is based on having inside information from Obama in the 2008 and 2012 elections. Probably information on voter fraud.
“his previous predictions against Trump were not based on polling numbers or other statistical sources, but rather on his “subjective” opinion of Trump’s chances. His statistical models had Trump winning all along (because Trump was leading in the primary polls, pretty much all along).
Now, he has moved back towards a statistically-based analysis, and it is THAT analysis that is resulting in his 80%-20% chances.
I tend to think Silver is correct here”
All polling is subjective. It is based on assumptions about randomness of the sample, selection criteria, eventual turnout, reliability of respondent’s answers, bias in wording the questions and bias of the questioner.
Most polls available to the public are not meant to be accurate. Their purpose is to shape opinion, not to measure it.
We shall see if Silver is correct.
I noted that in my post. Nate Silver noted it as well.
So why do I like Nate Silver's "Methodology"? Partly and maybe mostly because in 2012 he made the call on all 50 states and got ever single one of them right. Nobody else did that.
I also like him because he is not a Pollster. He uses other peoples polls and picks them apart and weighs their value to build an overall model of the election. He compares the sample set for each of the polls vs. what he expects for the turnout of D vs R vs I and makes adjustments to each poll. Then he weights them based on his view of their relative value and then averages them.
IMHO that is as good as you can do given the biased crap we have for polling. The Pollsters are mostly Dem Favoring Crooks (IMHO). Nate Silver is certainly a Lib but he was good in 2012 and I expect him to be good in 2016. I also expect Trump to win.
He is way off on the GOP primaries, way off on the last World Cup soccer tournament. Did he expect Leicester City to win the EPL this year? His models do not hold up.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.