Who is going to win?

[DinoDoc]

Out of interest, would a Romney win hurt the cache Silver built last Presidential election cycle? I don't actually care given that he seems inoffensive but the near religious fervor with which some people (not you) have taken to defending him from criticism is a little wierd.

[giblets]

Ahahaha, If it were me - you'd be going on your usual shtick about how I'm obviously wrong.

If you had made Jaguar's post I would be very surprised.

[Ben Kenobi]

If you had made Jaguar's post I would be very surprised.

Why? That the reason for the discreprency between Ras and PPP is because PPP assumes that democrats have a built-in registration advantage? FFS, that's been known for like forever. I forget that Poly is a liberal paradise which is why this fact isn't well known by some polytubbies, but it's good to see Jag doing the Lord's work here.

We'll see who's right Ras was off by a couple come election day 4 years ago and the die-hard Obamaphiles that I talk to are all going on about how he's gonna win the EC and lose the PV. I'm actually more confident in Obama's showing this election than they are, which is strange.

[Jaguar]

Out of interest, would a Romney win hurt the cache Silver built last Presidential election cycle? I don't actually care given that he seems inoffensive but the near religious fervor with which some people (not you) have taken to defending him from criticism is a little wierd.

Yes, it would. People already made a hissyfit about how prediction markets "don't work" when Intrade gave a ~75% chance of the individual mandate being overturned and then it wasn't.

[giblets]

Why? That the reason for the discreprency between Ras and PPP is because PPP assumes that democrats have a built-in registration advantage? FFS, that's been known for like forever. I forget that Poly is a liberal paradise which is why this fact isn't well known by some polytubbies, but it's good to see Jag doing the Lord's work here.

We'll see who's right Ras was off by a couple come election day 4 years ago and the die-hard Obamaphiles that I talk to are all going on about how he's gonna win the EC and lose the PV. I'm actually more confident in Obama's showing this election than they are, which is strange.

PPP doesn't "assume" that there are more Democrats. They just find more Democrats when they call random people. You are retarded.

[Ben Kenobi]

People already made a hissyfit about how prediction markets "don't work" when Intrade gave a ~75% chance of the individual mandate being overturned and then it wasn't.

If I had had the conviction of laying down 1 grand on intrade for Rick Santorum on the trifecta of Colorado, Minnesota and Missouri, I'd be able to retire today. I figured he had two, but Minnesota?

[Ben Kenobi]

PPP doesn't "assume" that there are more Democrats. They just find more Democrats when they call random people. You are retarded.

Uh, yes, they do. Ras assumes that partisan registration is even. PPP has about a 5 point registration advantage for Democrats. Did you not understand Jag's post?

[giblets]

You obviously didn't value your own opinions that much.

[giblets]

Uh, yes, they do. Ras assumes that partisan registration is even. PPP has about a 5 point registration advantage for Democrats. Did you not understand Jag's post?

You're a moron.

[Ben Kenobi]

You're a moron.

Well, we've established that the issue seems to be your reading comprehension. You clearly did not understand what Jag posted. Wow. And here my first thought was, "gribbler agreed with the post?" Gribbler, must not have read it. Then you posted what you did and I thought, hrm.. Maybe I misjudged gribbler and he is amenable to reason and sense. But now the 'HATE KENOBI' "SMASH KENOBI" shields are down full force, so there's not much reason to discuss this with you further... Sad.

[Ben Kenobi]

You obviously didn't value your own opinions that much.

For all the word that gets out that I'm a blind fanatic, the truth is much closer to this.

[regexcellent]

I don't agree with Nate Silver's assessment but I think he is pretty much the only one doing things right among elections forecasters. By "right" I mean "in an even vaguely scientific manner."

[Jaguar]

The polling prognosticator who needs to lose his reputation over this election is Professor Sam Wang at Princeton. Like Nate Silver, he "called" every state but Indiana in 1998. I believe the RCP average also called 49 states that year.

The question isn't whether you can call things right - a lot of people can do that - the question is whether you can be trusted to provide people with useful information that other people haven't already provided. I can say that about Nate Silver. I can't say that about Sam Wang.

Like Silver, he uses state polls. Like Silver, he adjusts for house effects. Like Silver, he believes Romney needs to make up 2.1-2.3% in the key swing states to be the favorite. Like Silver, he publishes the confidence level in his predictions.

But unlike Silver, he publishes completely untrustworthy confidence intervals. If you took 100 independent situations where Nate Silver says he's 75% confident in his prediction, you'd find about 75 of those predictions come true. The same is not true at all for Sam Wang. His model predicts with 98% "certainty" that Obama will win.

This is because he doesn't account for covariance in the states and just multiplies out the polynomial as if each state's outcome is statistically independent from the next. Commenters have been yelling at him for this since 2008. His FAQ answer on the subject, written in 2008, shows that he didn't really understand the issue. It wasn't a big deal in 2008, because Obama had a 100% chance of victory. It's a big deal in any close election.

What's offensive is not his prediction that Obama wins, or his percentage chances for each state, or even his inability to estimate covariance between states. What's offensive is that he provided us with a virtually meaningless number. When Nate Silver says 75%, you can expect him to be right 75% of the time. When Sam Wang says 98%, you have no idea what that means. I might as well say that Obama has the lead in Jaguar Awesomeness Points.

[Hauldren Collider]

In 2008, unless you were an idiot you would be able to call about 47 states with no difficulty. The three really tricky ones were Missouri, Indiana and North Carolina. The others were just not that close.

This year, there are probably 10 states that will be pretty close.

[DinoDoc]

You have to admit though that Wang Approval Points has a better ring to it.