2016 Postmortem

aaaaaa5a

(4,667 posts) Sun Nov 4, 2012, 08:45 PM Nov 2012

Why should you trust Nate Silver? Today's national polling has already proved him right.

Never bet against Math and Science!

For those of us that have followed Nate Silver's website, we know that all throughout the election he had Obama winning the popular vote roughly 50.4 to 48.5.

Of course for most of the last several weeks, Romney has been well ahead in the popular vote. But Nate's model didn't budge with regard to the election day forecast for the the popular vote winner.

Now at this late hour, national polls are pouring in. And has anyone noticed anything? The national polls are suddenly lining up exactly where Nate Silver had them in his election day "now" forecast months ago.

Today he tweeted this.

Nate Silver ?@fivethirtyeight
Obama had good nat'l polls today, however, the model was already anticipating that he would so most gains were priced in.

http://twitter.com/fivethirtyeight

And he is exactly right. Nate's model told us that despite the Gallup and Rasmussen numbers which often times showed Obama with a lead nationally as large as 7 points, that in the end Obama would win the national popular vote by about 2 points.

Now look at today's national polling. Add it together and do the math. Bingo! Nate was right on the money months ago.

I think its possible that people like Nate Silver and Sam Wang could change TV punditry and the way we look at political races. All of Nate's numbers are coming home just like his stats claimed they would. Its both good, and quite frankly, downright scary.

19 replies

= new reply since forum marked as read

Highlight:

Why should you trust Nate Silver? Today's national polling has already proved him right. (Original Post) aaaaaa5a Nov 2012 OP

Trust but verify. grantcart Nov 2012 #1

Another recent tweet. aaaaaa5a Nov 2012 #2

Which is why kurt_cagle Nov 2012 #3

Post removed Post removed Nov 2012 #4

How awful. Jackpine Radical Nov 2012 #6

Sorry "dead on" means on target. Hoyt Nov 2012 #11

Sure, but you tapped into Jackpine Radical Nov 2012 #15

I think you need to change the wording. Two different meanings. LiberalFighter Nov 2012 #10

I took it as a threat. RomneyLies Nov 2012 #12

You guys need to chill out. It means I hope he's exactly right and Obama wins. Hoyt Nov 2012 #13

Yeah, but the context could also mean RomneyLies Nov 2012 #16

Yep, threatening is not this posters style. He's a snark master, not a crazy person. rDigital Nov 2012 #18

I think he has access LibbyD Nov 2012 #5

That was for 2008 only RomneyLies Nov 2012 #8

was talking to my friend about this today - analogy of the accuracy of the forecast of Sandy's path NRaleighLiberal Nov 2012 #7

I make a living doing statistical analysis RomneyLies Nov 2012 #9

Go Nate! I hope you are right!!! Kteachums Nov 2012 #14

A 2% popular vote margin would be huge BlueStreak Nov 2012 #17

Trust no one: Get Out The Vote! eallen Nov 2012 #19

grantcart

(53,061 posts)

1. Trust but verify.

Reply to aaaaaa5a (Original post)

Sun Nov 4, 2012, 08:46 PM

Nov 2012

Where Nate uses mainstream polls, even Rasmussen, he is fine.

When he allows the RW to flood a bunch of polling hacks into distort the result, ala North Carolina. you have to call it out.

aaaaaa5a

(4,667 posts)

2. Another recent tweet.

Reply to aaaaaa5a (Original post)

Sun Nov 4, 2012, 08:47 PM

Nov 2012

Nate Silver ?@fivethirtyeight
Everything gelling a bit: 1) US polls catching up to state polls 2) bad polls for BHO in noncompetitive states help explain remaining spread

Both Nate and Sam Wang wrote blogs on this explaining the reason behind Romney's popular vote numbers. (Even as vastly overstated as they were.) Wang even wrote a blog discussing how the more accurate state polls could be used to better determine national results, when compared to doing national surveys alone.

Once again, nerds rule!

kurt_cagle

(534 posts)

3. Which is why

Reply to aaaaaa5a (Original post)

Sun Nov 4, 2012, 08:48 PM

Nov 2012

the MSM will do everything in their power to shut Nate Silver down in the next election cycle. Hard to control the narrative when the quants consistently prove that you're wrong.

Response to aaaaaa5a (Original post)

Post removed

Jackpine Radical

(45,274 posts)

6. How awful.

Reply to Post removed (Reply #4)

Sun Nov 4, 2012, 08:51 PM

Nov 2012

I hope he's alive & well on Tuesday.

Hoyt

(54,770 posts)

11. Sorry "dead on" means on target.

Reply to Jackpine Radical (Reply #6)

Sun Nov 4, 2012, 08:54 PM

Nov 2012

Jackpine Radical

(45,274 posts)

15. Sure, but you tapped into

Reply to Hoyt (Reply #11)

Sun Nov 4, 2012, 09:03 PM

Nov 2012

my weakness for cheap shots.

LiberalFighter

(51,094 posts)

10. I think you need to change the wording. Two different meanings.

Reply to Post removed (Reply #4)

Sun Nov 4, 2012, 08:53 PM

Nov 2012

Better if you say something like "Hope Silver is on the mark on Tuesday"

RomneyLies

(3,333 posts)

12. I took it as a threat.

Reply to LiberalFighter (Reply #10)

Sun Nov 4, 2012, 08:55 PM

Nov 2012

Seeing how he probably meant it, now.

Hoyt

(54,770 posts)

13. You guys need to chill out. It means I hope he's exactly right and Obama wins.

Reply to RomneyLies (Reply #12)

Sun Nov 4, 2012, 08:59 PM

Nov 2012

In fact, I hate using dictionaries, but Webster defines "dead on" as "exactly correct, accurate. "

RomneyLies

(3,333 posts)

16. Yeah, but the context could also mean

Reply to Hoyt (Reply #13)

Sun Nov 4, 2012, 09:08 PM

Nov 2012

you wanted Nate dead on Monday.

Now, had you said "Hope Silver's numbers are dead on Tuesday" it would have been clearer.

Just saying, your original post wasn't as clear as it could have been and seeing as how I'm pretty new here, I haven't seen enough posts from you to know much about you.

Things are, of course, much clearer now.

BTW, some of my favorite authors are as terse as your post (Roger Zelazny comes to mind). Had I been thinking more about the context of the overall thread, I'd probably have caught your meaning immediately.

rDigital

(2,239 posts)

18. Yep, threatening is not this posters style. He's a snark master, not a crazy person.

Reply to Hoyt (Reply #13)

Sun Nov 4, 2012, 09:25 PM

Nov 2012

Agree all should commence chilling.

LibbyD

(34 posts)

5. I think he has access

Reply to aaaaaa5a (Original post)

Sun Nov 4, 2012, 08:49 PM

Nov 2012

to internal polling as well? Didn't he sign a confidentiality agreement with the Obama campaign in 2008?

RomneyLies

(3,333 posts)

8. That was for 2008 only

Reply to LibbyD (Reply #5)

Sun Nov 4, 2012, 08:52 PM

Nov 2012

Welcome to DU

Enjoy your stay.

NRaleighLiberal

(60,021 posts)

7. was talking to my friend about this today - analogy of the accuracy of the forecast of Sandy's path

Reply to aaaaaa5a (Original post)

Sun Nov 4, 2012, 08:52 PM

Nov 2012

computers, math, science - the running of many models, and convergence. The weather services do it...and Nate and Sam and the like are doing it. Times have changed....as the MSM will now learn, to its (deserved) detriment.

RomneyLies

(3,333 posts)

9. I make a living doing statistical analysis

Reply to aaaaaa5a (Original post)

Sun Nov 4, 2012, 08:53 PM

Nov 2012

Nate's one of the best in the biz.

Kteachums

(331 posts)

14. Go Nate! I hope you are right!!!

Reply to aaaaaa5a (Original post)

Sun Nov 4, 2012, 09:00 PM

Nov 2012

BlueStreak

(8,377 posts)

17. A 2% popular vote margin would be huge

Reply to aaaaaa5a (Original post)

Sun Nov 4, 2012, 09:10 PM

Nov 2012

That isn't "neck and neck".

Bush in 2004 was 2.5%
Bush in 2000 lost the popular vote
Carter in 1976 was 2%
Nixon in 1968 was under 1%
Kennedy in 1960 was way under 1%

Add to that this is the first Presidential election under Citizen's United rules, and this forced Obama to use a swing state strategy exclusively. There are lots of states where Obama has virtually no presence, ran virtually no ads, and had no campaign events. Considering there are states where the Citizen's United factor means Obama will lose by double digits, a 2% national win is equivalent to a 5-8% spread in previous elections.

eallen

(2,955 posts)

19. Trust no one: Get Out The Vote!

Reply to aaaaaa5a (Original post)

Sun Nov 4, 2012, 10:29 PM

Nov 2012

I want Nate Silver to be wrong.

I want us to do better than his estimates.

Get. Out. The. Vote.

Reply to this discussion