Nate Silver Is Unskewing Polls — All Of Them — In Trump’s Direction

[UCmeNdc]

Why doesn't someone suggest to Nate Silver to post both poll number crunching results on his 538 web site?

Nate Silver’s 538 model is giving Donald Trump a heart-stopping 35 percent chance of winning as of this weekend.

He ratcheted the panic up to 11 on Friday with his latest forecast, tweeting out, “Trump is about 3 points behind Clinton ― and 3-point polling errors happen pretty often.”

The short version is that Silver is changing the results of polls to fit where he thinks the polls truly are, rather than simply entering the poll numbers into his model and crunching them.


http://www.huffingtonpost.com/entry/nate-silver-election-forecast_us_581e1c33e4b0d9ce6fbc6f7f

1. So, Nate Silver, why not put the straight poll numbers in the computer and crunch them then post that result.

2. Then post the results of his "belief adjustment to the polls" to fit where he thinks the polls truly are.

Post both numbers on the web site:

Then the reader can see which model is ultimately correct when the election is over. Someone should suggest that to him.

[PeteSelman]

1. It's s either gong to be a Clinton win or a Trump landslide.

Hillary is almost certainly going to win but if she doesn't it's going to be a Reaganesque landslide for Trump. It won't go any other way.

[jcgoldie]

2. Wut?

Sorry that makes zero fucking sense.

[sarcasmo]

3. Trump landslide, might be the funniest thing I've heard on this site.

[regnaD kciN]

5. I'd put it the other way around…

If Trump manages to grab practically every swing state, he would still only get a little past 270. If there's going to be a landslide, it'll be for us.

[MoonRiver]

6. ???

[Demsrule86]

14. Trump will not win.

He has to run the board and Hillary needs one state...she is at what she needs right now.

[qdouble]

17. Even in Nate's conservative, Trump leaning measurements, the chances of a Trump landslide are just

0.4%. Not sure if you're serious.

[joeybee12]

4. I'll take Nate Silver Ober the Huffington Compost any day...nt

[regnaD kciN]

9. Except that HuffPost's analyst, Drew Linzer…

…has a perfect record for both 2012 and 2014.

[Demsrule86]

15. But not this year...primary was awful,

and he messed up Brexit too...and some say the Chicago/Cleveland games.

[pnwmom]

11. And I'll take Prof Wang over a Princeton Election Consortium.

He's explained why Silver gets those big swings in poll results -- he's double counting by using both state and national polls.

Wang only uses state polls.

[UCmeNdc]

10. I do not get why Nate Silver is getting so defensive

Post both of his predictions to his web site:

Post the old way with the straight numbers being crunched......

Then say I feel this polling method might not account for xyz, so I believe the real results are.....

Then he would not have to defend anything and just say let's see what happens on Tuesday.

[NoGoodNamesLeft]

12. He's probably trying to factor in that non-existent "hidden Trump voter" data...BUT

He appears to be completely ignoring the fact that Hispanic voters (many who register Republican) are NOT going to be voting Trump. So that counters any hidden Trump vote and files it into Hillary's column. And that number needs to be at least doubled, tripled in some places because the Latino Americans are feeling kinda spicy and it's payback time!

[wildeyed]

13. Did you follow the links in the article?

The assumptions in Silver's model were published in June. Nobody was bitching about it then. And he is not "unskewing" polls. He weights the polls the way he has ALWAYS done, by looking at past performance of the poll, giving the company a grade and weighting the HISTORICALLY more reliable polls.

This is a total hit piece on Silver and he is justifiably enraged by it. Who is Ryan Grimm anyway and why do we care what he says? Bio says he used to work for the Washington City Paper, which is one of those freebie rags you pick up at the coffee shop or bookstore. Hmmmm.....

Look, I don't like what Silver's model is saying either. But anyone who wasn't complaining about it when Clinton was way ahead, shouldn't be complaining now. It is THE SAME system. And if you think it is wrong, the ignore it. But anyone accusing him of 'unskewing' and comparing him to Dean Chambers is guilty of some serious clickbait BS.

[Demsrule86]

16. I think that Silver has slipped this year and not because the polls are not to my liking

The other day...I notice a Utah poll make Hillary's chance go down...which is ridiculous and indicates a bad model.

[marybourg]

20. I think you're confusing Silver

with pollsters. He's not a pollster; he's an aggregater. As explained up-thread, he decided on a formula for weighting polls at the beginning of the season and has consistently used it.

We may not like the results of the polls, but that's not Silvers fault. And he still has her winning on EC votes at 290.