Nate Silver to Joe Scarborough: Wanna Bet? - Yahoo! News

 

Thought Clyde might find this interesting considering it's his "go to" site. Silver seems pretty confident, I guess he has a reason to be...

 

Political polling guru Nate Silver is so confident in his statistical models that he just offered to bet MSNBC's Joe Scarborough $1,000 that Barack Obama will win re-election. Scarborough, you may recall, criticized Silver's math earlier this week, saying that "Anybody that thinks that this race is anything but a tossup right now is such an ideologue ... they're jokes." He was specifically talking about Silver's FiveThirtyEight website, which shows Mitt Romney with just a 1-in-4 chance of becoming president.

Nate Silver to Joe Scarborough: Wanna Bet? - Yahoo! News

 

Thought Clyde might find this interesting considering it's his "go to" site. Silver seems pretty confident, I guess he has a reason to be...

 

Political polling guru Nate Silver is so confident in his statistical models that he just offered to bet MSNBC's Joe Scarborough $1,000 that Barack Obama will win re-election. Scarborough, you may recall, criticized Silver's math earlier this week, saying that "Anybody that thinks that this race is anything but a tossup right now is such an ideologue ... they're jokes." He was specifically talking about Silver's FiveThirtyEight website, which shows Mitt Romney with just a 1-in-4 chance of becoming president.

 

I agree wholeheartedly with Scarborough. Anytime with these polls you have one that is completely out of whack with the rest, it's almost always the outlier. Gallup and Pew are much more reliable sources than a blogger for the NYT IMO.

Money talks.

I agree wholeheartedly with Scarborough. Anytime with these polls you have one that is completely out of whack with the rest, it's almost always the outlier. Gallup and Pew are much more reliable sources than a blogger for the NYT IMO.

 

538 is not a pollster, so he creates no data to be an outlier. He simply works with multiple poll numbers and various weighting and runs probability calculations. I guess if anyone has issues with him it would have to be with his weighting. We'll all know in less than a week.

Money talks.

 

Obviously Obama has the easier road to 270.

 

Mitt's still left with some ground to make up in some key states. Interesting Obama is headed to Wisconsin (Obama's), Nevada (Obama's) and Colorado (Mitt's, barely) today.

I agree wholeheartedly with Scarborough. Anytime with these polls you have one that is completely out of whack with the rest, it's almost always the outlier. Gallup and Pew are much more reliable sources than a blogger for the NYT IMO.

 

But they're not a pollster, they analyze an average of various polls as I understand it.

But they're not a pollster, they analyze an average of various polls as I understand it.

 

But he is, because he equates things and gives weight to things, like which party is more likely to turnout, that he has no true clue on. If he uses the same %'s as 2008, then he's going to be way off IMO.

Look, I'm not saying President Obama isn't going to win, I've long stated I think he will, but to predict it's going to be another landslide like Silver is, is asinine.

Silver is saying that neither one of them have much of a chance to win the election without Ohio and he thinks b/c of the polling numbers that Ohio goes to Obama. If Ohio does go to Obama I have a hard time believing Romney has much of a statistical chance to win. However just like in sports math something could change the course. The NYG at one point Sunday with a 23 point lead had a 99% chance to win the game according to the stats experts. After Dallas went on the furious storm they were at one point favored to win by the same experts. Silver is saying the serious storm has to be Romney carrying Ohio. I can't disagree with that. I think Romney has a better chance in Ohio than Silver is giving him however b/c of the early voting numbers we're hearing and I don't think Obama has close to the support he did on 08.

Look, I'm not saying President Obama isn't going to win, I've long stated I think he will, but to predict it's going to be another landslide like Silver is, is asinine.

 

I don't think he's predicting a landslide as much as as he's giving Romney a very low percentage chance to win Ohio. He said last night if Romney carries Ohio he's only giving Obama an 8% chance to win the election. To me that isn't saying Obama in a landslide.

I don't think he's predicting a landslide as much as as he's giving Romney a very low percentage chance to win Ohio. He said last night if Romney carries Ohio he's only giving Obama an 8% chance to win the election. To me that isn't saying Obama in a landslide.

 

Look at his predicted voting breakdown.....About a 70-85 Electoral vote win for the President. To me, that constitutes a landslide. He's saying even if Romney takes Ohio, he still loses.

Criticism of Nate Silver has been deeply misinformed and has coincided with the election year carp about polls - they are wrong if they don't show my guy winning. I think people should understand Silver is offering a probability (which has a precise statistical meaning which should not be conflated with certainty) of a particular outcome based on publicly available polling data. While his model is a bit more sophisticated than that as it includes economic data and the national polls (and thus makes it a bit more bearish on the president's chances than mere aggregators), that's the gist of it. If you think its probabilities are wrong, then you also think available poll numbers are also wrong. They, of course, may well be, but you can't build a statistical model around your "gut feeling."

Criticism of Nate Silver has been deeply misinformed and has coincided with the election year carp about polls - they are wrong if they don't show my guy winning. I think people should understand Silver is offering a probability (which has a precise statistical meaning which should not be conflated with certainty) of a particular outcome based on publicly available polling data. While his model is a bit more sophisticated than that as it includes economic data and the national polls (and thus makes it a bit more conservative on the president's chances than mere aggregators), that's the gist of it. If you think its probabilities are wrong, then you also think available poll numbers are also wrong. They, of course, may well be, but you can't build a statistical model around your "gut feeling."

 

He's including stats like "electorate makeup" and things of those nature in. Where is he seeing/getting that information? Since the election hasn't happened yet, he has to be using 2008 data as an estimate for those numbers, correct?

Look at his predicted voting breakdown.....About a 70-85 Electoral vote win for the President. To me, that constitutes a landslide. He's saying even if Romney takes Ohio, he still loses.

 

I sort of agree with you except that he stated Obama only has an 8% chance to win if he loses Ohio. He basically is stating Ohio is the only swing state that matters. I think I agree with that. I'm like you though in that I don't think Ohio is as easily Obama as he does b/c I don't think the D's will vote nearly as much as they did in 08 and the R's are going to come out and vote.

He's including stats like "electorate makeup" and things of those nature in. Where is he seeing/getting that information? Since the election hasn't happened yet, he has to be using 2008 data as an estimate for those numbers, correct?

 

No, not necessarily. He can use current state polls to get a breakdown of partisan distribution in each state, can use approval ratings of each candidate by state, can use voter registration tallies (which I'm sure he controls for against previous elections), and can use state fundraising totals for each candidate all to get a sense of what state electorates look like. This much he says on his site.