2016 Postmortem
Related: About this forumDuer's, I would advise IGNORING Nate Silver.
He clearly was a one trick pony. He has gotten so much wrong this year he should consider retirement.
olddots
(10,237 posts)litlbilly
(2,227 posts)Scootaloo
(25,699 posts)All he really does is compile other people's data and make projections on that. If that data is flawed, what he outputs will be flawed, too. It's a polling thing, not specifically a Nate Silver thing. Maybe he ought to stick to running the numbers at the ponies or whatever his main gig is. At least until polling techniques catch up with technology and demographics.
LonePirate
(13,386 posts)Perhaps not for the primaries but I suspect for the general he will consider other variables or he might apply restrictions or weights or other changes to the data he uses.
What he does is not magic and it never has been. The fools in the press simply are bad at math and don't invest the time he has to compile the results.
We're definitely witnessing first hand the reason why Gallup stated they were essentially abandoning polling this cycle.
Cleita
(75,480 posts)I knew he would be missing the trends and he did.
MellowDem
(5,018 posts)he does great statistical analysis. He's a poll aggregator, if the polls are way off he will be as well, which does happen, and causes pollsters to re-examine how they poll. Michigan was kinda strange in that there was no 2008 primary to go off of for modeling and before that it was Caucasus.
He's been very accurate so far this election. He predicted the 2008 and 2012 elections very well.
And his analysis for why the polls got it so wrong will be interesting reading as well. You'll want him as a resource for the general election too.
I'm a stats junkie and a Sanders supporter. I know it sucks seeing the hard math being against my candidate, but the facts are the facts, and I'm ok with it. I definitely hope Sanders is able to pull more upsets and show some of these polls as wrong, but we'll see!