10/30 Nate's 538, OBAMA: 299, RMONEY: 239 - 77% to 23% --- Sam Wang Obama: 305 Romney: 233 (Links)

[Ellipsis]

Last edited Wed Oct 31, 2012, 01:27 AM - Edit history (1)

http://fivethirtyeight.blogs.nytimes.com/


http://election.princeton.edu/

As of October 30, 8:01PM EDT:
Obama: 305
Romney: 233
Meta-margin: Obama +2.36%

Probability of Obama re-election: Random Drift 93%, Bayesian Prediction 98%

[FVZA_Colonel]

1. Wow, what caused that jump?

It's wonderful, but I wasn't expecting that much.

[ProudToBeBlueInRhody]

2. Steady, positive numbers in VA and CO I would say....

...as well as holding the line in OH.

[Ellipsis]

4. Pulling out Gallup as an outlier I would think too and, perhaps? ... early voting idiocies.

Real Clear Politics Polls show a pretty boring last 24 hours.

[gcomeau]

7. Mostly 2 new Ohio polls.

And the new Quinnipiac +5 in Ohio isn't there yet. :-D

[Ellipsis]

11. ahhh.

[FVZA_Colonel]

14. With Quinnipiac to include, I can't wait to see what tomorrow looks like.

[reformist2]

3. Every day now with no real change in the polls ought to move the odds up a few percent.

I still long for the days when the "nowcast" was at 98% though.

[Quixote1818]

9. Thats exactly right

There seems to be no movement toward Romney yet. Lets hope it stays that way.

[reformist2]

5. FWIW, Intrade has the chances at 63.6% vs. 36.4% now. :)

[freshwest]

15. I don't understand these numbers. Or 'Random Drift 93%, Bayesian Prediction 98% '

Is Intrade saying Obama has 63.6% of the vote, although other polls are keeping it at around 50%?

That I understand, maybe. But the others from Princeton:

Random Drift 93% is 93% of what?

Bayesian Prediction is 98% of what?

The 93% or 98% can't be votes. What does that refer to?

Does it mean those are the chances Obama will get 63%?

And OBAMA: 299, RMONEY: 239 - 77% to 23% is not close to those numbers, either, but that would be the electoral votes, which is not the real votes.

I did not study statistics. Please dumb it down. A lot.

[Ellipsis]

18. 63.6% is VEGAS odds.

Random Drift 93%, Bayesian Prediction is 98% is a Professor of statistics, form Princeton University putting his ego on line... giving an assuredly that Obama will win.

In short... the odds are very good.

[freshwest]

19. So the Princeton figures are the electoral votes but the Intrade give less odds?

IOW, as you say, Princeton says it's a sure thing, although the media has made it look close.

And the Vegas people don't think it's that good, but is not based on votes.

This is getting far afield of what many of us think voting is.

I'm guessing they have certain vectors they are using to determine these figures, demographics, committed electors, registrations, early turnout, things like that?

Thanks.

[Ellipsis]

20. Nate Silver has a published formula which you can find on Wikepedia...

He was originally a baseball statistician. Will Pitt would adore him, because it was the formula that allowed the Boston Red Sox to win the world series for the first time in like a 100 years.

Intrade, what little I know of it, is publicly traded... the house always wins, so the odds are higher... about 70% to 30% lets say.

Princeton is academia ... science, algorithms etc.


..so yea, your analysis is right on.

[freshwest]

21. Yeah, I'd seen Nate Wikipedia page a while back... It was 'Greek' to me. Thanks.

[reformist2]

25. It's more like Nate's 538 is a scientific assessment, and Intrade/Vegas is the "gut" feeling.

So the gut feel says Obama is 2:1 favorite over Romney, but the scientific assessment of all the polls taken together (using Bayesian analysis) is actually a little higher at more than 3:1 now. The un-scientific Rethugs are screaming about Nate's inherent "bias", but the fact is they are silent about the Intrade number, which really isn't saying anything that different.

[LiberteToujours]

22. Here's some explanation

Is Intrade saying Obama has 63.6% of the vote, although other polls are keeping it at around 50%?

Intrade is saying there is a 63.6% chance of Obama winning the election (270 or more electoral votes); not that he will get 63.6% of the vote. Well, it's a betting website, so those are the odds on Obama winning, it's not statistical analysis.

The 93% and 98% are a separate analysis from some professor stating what he believes the odds are of Obama winning based on various statistical methods.

[freshwest]

23. Perhaps Intrade has a source on committed electors. The ballots haven't been counted yet.

During the 2008 primary the number of delegates commited to Obama exceeded the number necessary for nomination, and it was decided before the DNC, not the way it once was done. Romney did the same, with arguments continuing throughout the RNC in Tampa.

I'm old enough to remember state delegates voting at conventions, and the only real election number that mattered were the ballots. It's more complicated now but I wish it wasn't. Thanks very much.

[timlot]

6. Oh let it be.

Every time the numbers go up I feel like this crazy lady. Only I hope it come true.

<iframe width="640" height="360" src="

?feature=player_detailpage" frameborder="0" allowfullscreen></iframe>

[Ellipsis]

13. It will be. GOTV.

[Rowdyboy]

8. Thanks...I've gotten addicted to his site, as well as Sam Wang's....makes elections even more fun

[Ellipsis]

10. I edited the OP to add the "Wangster"

...just starting to appreciate him

[bluestateguy]

12. Hopefully Gallup does not come back

I think that is what helped the most.

[mil_5529dem]

16. woohoo

Fantastic news!

[JI7]

17. that Astrophysicist has it at Obam 278, Romney 256, Toss-Up 4

http://colleyrankings.com/election2012/

[Cali_Democrat]

24. Nate Silver and Sam Wang need to stop bombarding America with all these facts and hard numbers

It's just wrong. America just can't handle it at the moment.