Welcome to DU! The truly grassroots left-of-center political community where regular people, not algorithms, drive the discussions and set the standards. Join the community: Create a free account Support DU (and get rid of ads!): Become a Star Member Latest Breaking News General Discussion The DU Lounge All Forums Issue Forums Culture Forums Alliance Forums Region Forums Support Forums Help & Search

2016 Postmortem

MSMITH33156

(879 posts) Wed Nov 7, 2012, 11:57 AM Nov 2012

No one will notice

but it looks like Nate's Senate forecast might need some work. He had 92.5% chance that Berg would beat Heitkamp, and 65.6% chance that Tester would lose Montana. I think modeling senate races is infinitely harder than modeling the Presidential race, and I am not knocking him at all. Just interesting that no one has really noticed this.

3 replies

= new reply since forum marked as read

Highlight:

No one will notice (Original Post) MSMITH33156 Nov 2012 OP

I noticed. VWolf Nov 2012 #1

You have to EXPECT a few to go wrong when there are so many races. Silent3 Nov 2012 #2

possibly because they are just so many more national polls and presidential polls? faithfulcitizen Nov 2012 #3

VWolf

(3,944 posts)

1. I noticed.

Reply to MSMITH33156 (Original post)

Wed Nov 7, 2012, 12:28 PM

Nov 2012

Here's the thing. I'd rather be pleasantly surprised than unexpectedly disappointed. So I'll cut him some slack on this.

Silent3

(15,259 posts)

2. You have to EXPECT a few to go wrong when there are so many races.

Reply to MSMITH33156 (Original post)

Wed Nov 7, 2012, 12:33 PM

Nov 2012

92.5% is not 100%. 65.6 is much further from 100% -- you can only expect a guy with a rating of 65.6% to win 2 out of 3 times in similar circumstances.

Being "right" in the kind of work Nate Silver is doing doesn't mean that everyone race you rate at over 50% wins and every race rated under 50% loses.

faithfulcitizen

(3,191 posts)

3. possibly because they are just so many more national polls and presidential polls?