Rassmussen vs Nate Silver

[rockyct]

The only place where Silver "missed" was in the Senate by about 2 seats. He had an average of 52.5 seats for the Democrats. It's probably going to take at least one recount, but the Democrats look good to take 55 seats but you could average it as 54.5. Basically, all of the toss up seats went to the Democrat except for NV and AZ if you considered AZ being a toss up.

There is still a lot more counting to be done in MT but the Dem has a 3 point lead. That race was actually interesting in that the polls showed Tester winning by 1-2 but Nate's formula used "state fundamentals" to make the Republican the 2-1 favorite. Also, if Heitkamp keeps her 1 point win (3k votes), it will be Nate's worst loss. He had projected the Republican to win 92% of the time. I guess we can general say that smaller states are just harder to get strong polling numbers from.

On another note, I know a lot of us thought McCaskill was going to win (including Silver)..but not by over 15 points (except for SurveyUSA so credit them there).

As far as Presidential numbers, Nate was basically perfect.

 

[Doppel]

538 fvcking KILLED IT.

He got the popular, based on what I see at the moment, and down to exactly every single state as well, including the very marginal lean toward Obama, which is exactly what we're seeing now.

Anyway, I was wrong about the polls not Nate Silver.

Haha, fail. Just admit you had no idea what the hell you were talking about in recent weeks and it will be easier for you. Only if you admit complete defeat and that you were absolutely wrong can you actually learn from this, otherwise you learn nothing about it and about yourself.

 

[GarfieldtheCat]

I never said Silver was wrong in his methodology. If his model was incorrect it would be because of the polling data not because of the model itself. I thought the polls were over estimating Democratic turnout which turned out to be 100% wrong.

If we want to be picky. If Silver "predicts" 50 out of 50 states this is actually a failure of his model, at least of his percentage calculations, since there should have been some misses.

Colorado and Virginia both with 80% chance for Obama, 84% chance for Iowa and 50.3% chance for Florida to go Obama. If these percentages actually mean anything the odds are that one of them would have missed. There should have been a 27% chance of getting all of those states right.

If you take all of his probabilities he should have been right 10% of the time on every one of his picks.

His projections on probabilities will be tweaked, I'm sure of it. He's mentioned this in his blog.

Anyway, I was wrong about the polls not Nate Silver.

LOL, are you a Fox News commentator? You have the same anti-reality field that they have it seems.

How's the Romney victory that your "projections" predicted working out?

 

[SP33Demon]

Nate silver is a fucking bean counter.

So what does that make you and where is your awesome model? Nate's predictions were all within the margin of error and you look like a fucking dipshit. Go crawl back into obscurity with the rest of the Republicans for another 4 years.

 

[HomerJS]

A seventh grader could make a predictive system. It isn't magic, I'll give him credit for believing the polls because they ended up being right.

You mean Republicans were not undersampled??

 

[HomerJS]

I never said Silver was wrong in his methodology. If his model was incorrect it would be because of the polling data not because of the model itself. I thought the polls were over estimating Democratic turnout which turned out to be 100% wrong.

If we want to be picky. If Silver "predicts" 50 out of 50 states this is actually a failure of his model, at least of his percentage calculations, since there should have been some misses.

Colorado and Virginia both with 80% chance for Obama, 84% chance for Iowa and 50.3% chance for Florida to go Obama. If these percentages actually mean anything the odds are that one of them would have missed. There should have been a 27% chance of getting all of those states right.

If you take all of his probabilities he should have been right 10% of the time on every one of his picks.

His projections on probabilities will be tweaked, I'm sure of it. He's mentioned this in his blog.

Anyway, I was wrong about the polls not Nate Silver.

and the earth is still flat

 

[HomerJS]

Bucky is like Karl Rove last night. Even when their own bean counters go against his preconceived notions he's still in denial.

 

[HomerJS]

We need the picture of Muhammad Ali standing over Sonny Liston with captions. In case you didn't figure it out Nate Silver is Ali

 

[HomerJS]

Who will emerge as the most accurate? Fox News and the right have made a habit of always quoting Rassmussen claiming others underpoll Republicans. It will be interesting

As of now popular vote...

Rasmussen

Romney 49%
Obama 48%
Others 2%

Nate Silver

Obama 50.8%
Romney 48.3%

Popular vote results

Obama - 50%
Romney 48%

He even nailed that one!!!

 

[buckshot24]

LOL, are you a Fox News commentator? You have the same anti-reality field that they have it seems.

How's the Romney victory that your "projections" predicted working out?

I had no projections. Turns out I was wrong about the electorate and I never said that Silver's methodology was wrong. The data he was using turned out to be right. What is wrong about this "reality"?

 

[ohnoes]

I had no projections. Turns out I was wrong about the electorate and I never said that Silver's methodology was wrong. The data he was using turned out to be right. What is wrong about this "reality"?

I think the point was that if you understand stats, aggregating data reduces variance and increases accuracy. There could have been a systematic polling bias across all polls but that's very unlikely. You can believe that a few polls are biased, but thinking the whole system was biased was naive and due more to ideological reasons than anything based on reality.

 

[Charles Kozierok]

I had no projections. Turns out I was wrong about the electorate and I never said that Silver's methodology was wrong. The data he was using turned out to be right. What is wrong about this "reality"?

You're rewriting history.

You took the only two polls that showed you what you wanted to be true and repeatedly defended them against arguments from others that showed them to be outliers.

And by the way, what Silver does is a lot more complex and intelligent than just averaging polls together.

 

[Todd33]

I had no projections. Turns out I was wrong about the electorate and I never said that Silver's methodology was wrong. The data he was using turned out to be right. What is wrong about this "reality"?

Your about face is amazing. Karl Rove would be proud!

 

[trmiv]

So as Silver's methodology gets even more refined with more elections, will there ever be a mystery on elections again? Seems unless things are really, really, really close we'll pretty much know the results ahead of time from now on. Pretty cool.

 

[jackstar7]

Turns out I was wrong...

That's all you needed to say, bro.

Have a good one.

 

[randomrogue]

So as Silver's methodology gets even more refined with more elections, will there ever be a mystery on elections again? Seems unless things are really, really, really close we'll pretty much know the results ahead of time from now on. Pretty cool.

There's always October surprises and poor polling.

 

[jackstar7]

So as Silver's methodology gets even more refined with more elections, will there ever be a mystery on elections again? Seems unless things are really, really, really close we'll pretty much know the results ahead of time from now on. Pretty cool.

I think it will change the stakes for "game-changer" level moments. And next time there will not be a re-election campaign in the model. So unless one side can produce and elevate a candidate that produces the same raw level of excitement that president Obama produced in '08, I think things will start off close and stay that way.

But I'm assuming we aren't going to have 4 years of Republican rape talk. That could certainly skew things.

 

[SP33Demon]

I had no projections. Turns out I was wrong about the electorate and I never said that Silver's methodology was wrong. The data he was using turned out to be right. What is wrong about this "reality"?

Except Nate adds his own weighting to the data he receives and his weighting system is the best in the country. This is too hard for a mental midget like yourself to realize and give credit where credit is due.

 

[buckshot24]

You're rewriting history.

You took the only two polls that showed you what you wanted to be true and repeatedly defended them against arguments from others that showed them to be outliers.

And by the way, what Silver does is a lot more complex and intelligent than just averaging polls together.

I never ever argued that they were right and the others were wrong all I ever did is ask why you and others thought the two party ID polls were so wrong (when they had been fairly accurate up until last night). Now we know that they were full of shit. There was conflicting data which has since been resolved.

 

[Charles Kozierok]

I never ever argued that they were right and the others were wrong...

Yes, you did. You constantly defended Rasmussen's track record and his credibility, even though he was way, way outside the numbers of all of the other pollsters.

You also kept repeating theories about how the party ID splits of all the other polls were wrong.

Just admit it, and move on.

 

[alien42]

A seventh grader could make a predictive system. It isn't magic, I'll give him credit for believing the polls because they ended up being right.

 

[Pens1566]

Uh, so Silver was right? Couldn't have imagined that.

 

[buckshot24]

Except Nate adds his own weighting to the data he receives and his weighting system is the best in the country. This is too hard for a mental midget like yourself to realize and give credit where credit is due.

First of all FUCK YOU.

Secondly, Silver is BRILLIANT. That doesn't make him a wizard at predictions. Furthermore his results were too good for the percentages he gave each race he's written about this in his blog and he'll more than likely bring it up in an upcoming post.

Thirdly, a model could have been constructed using just polling aggregates that made the same predictions. Case in point.

 

[Todd33]

I never ever argued that they were right and the others were wrong all I ever did is ask why you and others thought the two party ID polls were so wrong (when they had been fairly accurate up until last night). Now we know that they were full of shit. There was conflicting data which has since been resolved.

Dude, you are so dishonest. You cherry picked all day long and fell into Karl Rove's propaganda. You wanted to win so bad you created a reality distortion field. You constantly had to think all polls had a bias while the few outlier you wanted to believe were right. Stop listening to the echo chamber, Fox News, it will only tell you what you want to hear.

Math wins, wishful thinking loses.

 

[Phokus]

I love XKCD