The real winner last night was Nate Silver. As in 2008, he nailed it.
A perfect 1.000; 50 for 50 on calling the individual states.
For the popular vote, he predicted 50.8% to 48.3%. The actual result? 50.3% to 48.1%. Impressive.
I love it when raw data wins. How 'bout them apples, pundits?
http://fivethirtyeight.blogs.nytimes.com/
..
..
..
Or, put another way:
Nate Silver 50 | Joe Scarborough 0
^ You’re entitled to your own opinions, you're just not entitled to your own facts
Based on my calculations, Silver had a 12.96% chance of predicting 51 for 51 (including DC)
He hit everything except Florida (went to Romney) in July when he started the model.
That's both impressive, and ironic in that $6 billion in spending money changed nothing.
(11-07-2012 10:11 AM)bronconick Wrote: [ -> ]That's both impressive, and ironic in that $6 billion in spending money changed nothing.
Absolutely. That was the money quote on last night's live Daily Show, in my opinion...
Quote:"Two years, three billion dollars, and we are clearly in the same f***ing place we were when we started."
- Jon Stewart
My brother was CONVINCED Romney was going to run away with it. I sent him Silver's link and told him I thought otherwise. He called him an idiot. I sent him a text the moment Obama won and said "what do you think of silver now?" He never responded hahah!
stdatwmu
Quote:The real winner last night was Nate Silver. As in 2008, he nailed it.
A perfect 1.000; 50 for 50 on calling the individual states.
For the popular vote, he predicted 50.8% to 48.3%. The actual result? 50.3% to 48.1%. Impressive.
I love it when raw data wins. How 'bout them apples, pundits?
http://fivethirtyeight.blogs.nytimes.com/
Not exactly earth shattering. A trained monkey could pick 45 states, picking the final 5 doesn't sound quite as amazing as saying he picked all 50.
I'll slow down just for you.
It was about picking 4 or 5 states, not 50. NH, NC, VA, FL, OH.
The point is that picking the winner in 5 states isn't nearly as difficult as you make it sound when you say he picked all 50.
He isn't picking anything. He is aggregating polling data to shrink the error term. If you collect enough data that +/- shrinks to a level where you can be really accurate. Nate Silver provides a valuable service by warehousing polls, and analyzing the cross tabs and past results to weigh the strength of various polls. So he doesn't over count prolific yet inaccurate polling firms the way a real clear politics or pollster.com would. He is still only as good as the pollsters though who in this case were very accurate on the balance.
The bad predictions are always people trying to extrapolate conclusions from a favorable poll that match their hopes or experience.
Never argue algorithms or numbers with an Aspergers guy. Serious, they crush you every time just like Nate did.
(11-08-2012 12:56 AM)DesertBronco Wrote: [ -> ]Never argue algorithms or numbers with an Aspergers guy. Serious, they crush you every time just like Nate did.
You of all people mocking "Aspergers."
Once again. Everyone on the planet knew who would win 45 of the 50 states. They were the same states that went either Rep. or Dem. for the last gagillion years. So it came down to picking the 5 swing states. It's just not that impressive when you are picking among 5 states, but sounds amazing when you say he got all 50.
Quote:The bad predictions are always people trying to extrapolate conclusions from a favorable poll that match their hopes or experience.
Kind of a no brainer don't you think, at least in a close election.
I'm not mocking it at all, it's a fact. They're money that way.
(11-08-2012 01:22 AM)DesertBronco Wrote: [ -> ]I'm not mocking it at all, it's a fact. They're money that way.
Had a kid once who could name the number of every NHL hockey player.............but couldn't tie his shoes.
But he didn't "pick" them he just took all the polls in every state giving weights to them based on past performance, sample size etc., factoring in well established economic data and ran state by state monte carlo simulations. If the polls are accurate he is going to be accurate because the error term falls to insignificant levels. He didn't invent the math, conduct the polls, right any software, or use any subjective criteria. Its just aggregation and since there are fewer swing states and more polls in the field than ever before, he is going to be really really accurate. Unless hundreds of polls are systemically wrong and all in the same direction for reasons we don't see until after the election.
Alrighty. Nate's the man. Picked all 5 swing states, simply incredible.
Note to self: Is the success of some guy named Nate Silver really worth debating? Probably not.
Considering Fawks called him everything but a human being because of his prediction, I think it's a big deal.
Did anyone see Rove in complete denial last night? How does he face those big check writers to Crossroads this week? "Uhhhhh, that was other kids, yeah yeah"
![[Image: beavis-and-butthead-beavis-and-butthead-...45-559.jpg]](http://images5.fanpop.com/image/photos/28000000/beavis-and-butthead-beavis-and-butthead-28017839-445-559.jpg)
I'm kind of with you on this one CD. I'm not that impressed by what he does. He is like a chef with really fresh ingredients and hungry customers. He doesn't over complicate things and he produces something very easy to understand but he isn't a political guy he is a numbers guy. I think its kind of significant that he is becoming a media icon and a hero to the left for reassuring everyone for months of Obama's inevitable victory. But the idea that he has any kind of magical insight is crazy. Anyone with enough free time to compile spread sheets and $55 for stata can replicate what he does. Its the pollsters who model the electorate and accurately sample it that deserve any credit for being "right" IMO.
You guys mock him AFTER the fact like you knew it all along, he was out there on his own calling this in the face of strong headwinds, other pollsters saying otherwise and in complete denial. They got personal with him, he was confounded (act, trust me), but in the end, he CRUSHED EVERYONE.
I'm starting to think that Rove's magic math of finding 50%+1 in 2004 was just self promotion on his part and he really wasn't trying for a narrow victory so W wouldn't be beholden to so many people in his 2nd term. He has been staggeringly wrong since then and wasted a lot of peoples money on some bad assumptions. He has been the loudest voice for exciting the base while the base has been shrinking. I think hes a one trick pony. He had a candidate in W he had a group of evangelicals who were ripe to be mobilized and he had a country divided by the war in a way that favored his guy.
