Why Nate Silver predicts an Obama win.

Why Nate Silver predicts an Obama win. (Full Version)

All Forums >> [Community Discussions] >> Dungeon of Political and Religious Discussion

Message

DarkSteven -> Why Nate Silver predicts an Obama win. (11/4/2012 9:25:17 PM)
Obama's ahead in Ohio. That's it. Very simple. You take all the states that will predictably vote Democratic, like California and New York, and you add in the swing-states-that-will-vote-blue, like Wisconsin and Michigan, and then Ohio will be enough for a win. That's it. Unfortunately, Silver's under attack by a lot of folks for the various reasons: 1. They don't understand probability. It amazes me that otherwise intelligent people are math-illiterate. When Silver says that there's an 80% chance of an Obama win, he is NOT saying that Obama will get 80% of the delegates. 2. They don't like what he has to say. If you support Romney (and about 50% of the population does, or at least prefers Mitt to Obama), then you don't like Silver's conclusions. 3. They can't understand a word he says. To me, talk about systemic bias and reducing sampling error is natural, because I've worked as a data guy for years. Lots of folks have no clue. They don't understand that Silver's model simply processes poll results. They think that he's introducing bias within the model and don't understand his explanations. 4. He terrifies them. Assuming that his model does as well in 2012 as it did in 2008, then they're not going to be able to justify simple bluster and gut feel any more. Some of these guys probably struggled with basic algebra, and they do NOT want to have to become conversant with statistics and math should the field morph into a math-heavy discipline.

slvemike4u -> RE: Why Nate Silver predicts an Obama win. (11/4/2012 9:34:13 PM)
The guy is a genius,pure and simple...he looks at numbers and see's probability's He has no bias to introduce because numbers have no bias...and that is all the guy cares about,it's all he ever cared about. His work in baseball changed the way prospects were looked at and measured.

FMRFGOPGAL -> RE: Why Nate Silver predicts an Obama win. (11/4/2012 9:40:21 PM)
Nate Silver is by far and away the most clinical and careful Psephologist in the entire field. And if you're into nerdy guys. He's cute too. Not that I would ever suggest you're into nerdy guys, Steven.

TheHeretic -> RE: Why Nate Silver predicts an Obama win. (11/4/2012 9:43:09 PM)
Yep. And Silver is bright enough to know it's garbage in, garbage out, and hedges his bet with that percentage, in case the pollsters aren't calling the right assortment of people who are going to show up on the big day. The straight fundamentals (the shit economy, which is rated as the most important concern, and the tendency of undecided voters to break to the challenger) say Obama will lose, except for the elephant in the room, and nobody will know until Tuesday night how that plays itself out. Nice thing about the western time zones, though. We'll either know, or know we aren't going to know for a while, in time to get to bed at a decent hour.

DarkSteven -> RE: Why Nate Silver predicts an Obama win. (11/4/2012 9:45:15 PM)
Mike, the point is that he is NOT a genius. He's simply a probability and statistics guy who's methodical and accurate. He takes the polls and combines them. The only true artistry is in determining the weighting factors and understanding the numbers. Very simply - if a poll has proven to be inaccurate before, Silver will still incorporate it, but will rely on it less than more reliable polls, by multiplying its results by a smaller number than the more accurate polls. The beauty is that by combining the poll results, he reduces sampling bias.

FMRFGOPGAL -> RE: Why Nate Silver predicts an Obama win. (11/4/2012 10:59:19 PM)
I'm sorry... while I don't feel authoritative enough to call him a "genius". I am really comfortable rolling my eyes at those who are either oversimplifying or indicating with reliability, that he's wrong.

focalss -> RE: Why Nate Silver predicts an Obama win. (11/4/2012 11:44:41 PM)
Help me out on this one. I read the article and it says the polls have Obama 48 - 47, ahead of Romney by 1%. (I can't tell if it is a national poll or tally of the electoral votes from the states.) From this Silver comes up with Obama getting a 75% chance of winning. Here is my map and it all comes down to Ohio, I am giving Romney Florida, VA, NC and Colorado. I am giving Obama Iowa, Wisconsin, Michigan, NH. With Ohio in play I see it as almost irrelevant if one or the other loses Colorado or NH. ( http://www.270towin.com/2012_election_predictions.php?mapid=biXo ) So back to the 10 poll example for Ohio only (if I am interpreting that correctly). If 10 polls in Ohio put Obama ahead by 1% in Ohio and all of them are within the margin of error I don't see how that translates to 75%. Maybe he is counting the polls in other states and not allocating things the way I am for VA and FL. However, in principle, if a poll is within the margin of error I see it as closer to 50-50. Granted if 10 of them for one state all put Obama ahead maybe that takes it outside the margin of error but not by that much. So either Silver is giving Obama more states than I am or I don't get how he comes up with a 75% chance. Lastly, I am writing off PA for Romney. I am pro Romney but its too late in PA except for one thing. I was encouraged by the University of Colorado model which gave Romney Pennsylvania a few weeks ago (and the election) if the unemployment rate is over 7.8% but I just haven't seen anyone pick up on that. Then the Halloween mask sales, Redskins losing at home, .... I don't count them for anything. The one thing in PA (or few things) all come down to who shows up to vote on election day in Ohio and PA for example. The second thing or maybe the same thing as the first thing is that the undecided people tend to either stay away or break for the challenger, that's a rule of thumb but who knows what will happen yet I am not seeing how Romney can win in PA due to the last polls I saw. 4% margin favoring Obama. Lastly I am optimistic (for Romney) because of the Colorado Model, the rule of thumb and because I feel the country will go down the drain with another four years of Obama so maybe that's wishful thinking on my part. On the other hand again, Obama has worked over the last few years for Ohio to have a better economy and it seems Kaisich? helped things get better there too. As to the popular vote I am predicting Romney winning.

tweakabelle -> RE: Why Nate Silver predicts an Obama win. (11/5/2012 12:10:40 AM)
Viewed from Down Under, Silver's blog at the NYT was a clear, helpful guide to the intricacies of US politics for this reader. Silver sets out his methodology clearly and precisely. He takes his predictions from the models he has developed over time and explains them lucidly. I found him an invaluable guide to the complicated process the US has adopted to elect its President. At no time could I identify any bias towards one candidate or the other, or towards one party or another. Any predictions Silver offered were backed up with numbers, analysis and common sense. Silver set out his reasons for adopting the positions he adopted with clear reasoning, and empirical evidence. He is quite happy to compare his model to others of a roughly equivalent level of sophistication, and open to discussing possible flaws within his model. IOW he adopted the scientific method, and followed it rigourously. For me it is easy to see why Silver's model and predictions have helped build his reputation as America's leading election analyst. Over time I far preferred Silver's analyses to those at polltracker.talkingpointsmemo.com, the main other poll tracking site I followed - Silver is far more lucid, rigourous and for mine, persuasive. I might add that, FWIW, I still haven't the faintest idea which way Silver will cast his personal ballot, or even if he is going to vote at all.

AstralDreams -> RE: Why Nate Silver predicts an Obama win. (11/5/2012 2:45:51 AM)
Great post and such an interesting site, sadly though despite who wins little difference is made on current American foreign policy.

DomYngBlk -> RE: Why Nate Silver predicts an Obama win. (11/5/2012 3:25:48 AM)
The problem with the polls is the reliance on "likely" voters rather than "registered" voters. This isn't a likey voter election. The resultant victory for Obama will show that to be the case.

DarkSteven -> RE: Why Nate Silver predicts an Obama win. (11/5/2012 4:57:05 AM)
quote: ORIGINAL: DomYngBlk The problem with the polls is the reliance on "likely" voters rather than "registered" voters. This isn't a likey voter election. The resultant victory for Obama will show that to be the case. DYB, the reverse is true. Polls of registered voters is a starting place, but the Dems traditionally have more issues with turnout than GOP. That's why the "likely voters" polls are so valuable because the polls of registered voetrs will have a Dem bias. focalss, my compliments - you seem to have the gist of things down well. However, I haven't seen polls that show Obama with a 1% lead. It's currently 2.4% in Ohio. The CU model is crap. It has never predicted an election. The model used previous election data to calibrate it. Silver himself trashed it here. He says: 1. The U. of Colo. model fits the equivalent of 7 unknowns to 8 elections. That's not a good idea. 2. The Colo. model also assumes huge effects from unemployment if incumbent is a Dem., but none if he's GOP. Hard claim to defend. 3. Also, it's false advertising to claim CU model has predicted the last 8 elections right. It's a new model. Hasn't predicted anything yet. ************ Fitting 7 unknowns to eight data points assures a mathematical fit, but one that it completely invalid. Instead of a smooth line, you get an equation that careens all over things and looks ridiculous.

DomYngBlk -> RE: Why Nate Silver predicts an Obama win. (11/5/2012 5:04:08 AM)
DS, I understand all of that and in off year elections I'd say that it is more than true. However, I think that from the early voting we have seen that this is far from a "likely" voter election. So I personally think the data from registered is more likely the real case. We will know more tomorrow night of course

DarkSteven -> RE: Why Nate Silver predicts an Obama win. (11/5/2012 5:10:25 AM)
quote: ORIGINAL: DarkSteven Unfortunately, Silver's under attack by a lot of folks for the various reasons: 1. They don't understand probability. It amazes me that otherwise intelligent people are math-illiterate. When Silver says that there's an 80% chance of an Obama win, he is NOT saying that Obama will get 80% of the delegates. 2. They don't like what he has to say. If you support Romney (and about 50% of the population does, or at least prefers Mitt to Obama), then you don't like Silver's conclusions. 3. They can't understand a word he says. To me, talk about systemic bias and reducing sampling error is natural, because I've worked as a data guy for years. Lots of folks have no clue. They don't understand that Silver's model simply processes poll results. They think that he's introducing bias within the model and don't understand his explanations. 4. He terrifies them. Assuming that his model does as well in 2012 as it did in 2008, then they're not going to be able to justify simple bluster and gut feel any more. Some of these guys probably struggled with basic algebra, and they do NOT want to have to become conversant with statistics and math should the field morph into a math-heavy discipline. 5. Some people don't even understand that a win in the popular vote does not assure a win in the electoral college. After Bush-Gore, I'm surprised anyone could not understand this... 6. Some folks say that "Everyone I know is voting for Romney (or Obama)!" They don't understand the severe bias inherent in that. As we've become more polarized, we're coalescing into "pockets" ruled by groupthink. We are a tremendously heterogeneous group, and if we go by just what out acquaintances think, we get a very biased sample of people that are similar to us socieconomically.

DarkSteven -> RE: Why Nate Silver predicts an Obama win. (11/5/2012 5:18:54 AM)
quote: ORIGINAL: DomYngBlk DS, I understand all of that and in off year elections I'd say that it is more than true. However, I think that from the early voting we have seen that this is far from a "likely" voter election. So I personally think the data from registered is more likely the real case. We will know more tomorrow night of course DYB, I'm not following. If the actual voters reflected perfectly the makeup of registered voters, then the registered voters and likely voters samples would be essentially the same. The very fact that they are different means that the registered voters will not vote with equal Dem/Rep representation. Unless you think that the methodology to determine what are "likely" voters is flawed. What is your rationale for saying that this election will be different? I get the feeling that it's due to the makeup of the early voters - true?

DesideriScuri -> RE: Why Nate Silver predicts an Obama win. (11/5/2012 5:22:00 AM)
quote: ORIGINAL: DomYngBlk The problem with the polls is the reliance on "likely" voters rather than "registered" voters. This isn't a likey voter election. The resultant victory for Obama will show that to be the case. Actually, "likely voters" is a more accurate picture than "registered voters." Like voters are a subset of registered voters. The polls with the largest margin of error tend to be those that sample only the people who are of voting age. Since everyone isn't registered, applying that technique doesn't truly represent who is actually going to be voting. Even registering isn't necessarily as accurate as we would hope. Not everyone that is registered votes. Likely voter samples have the lowest margin of error coming from the sample makeup (all other things being equal). http://www.census.gov/prod/2010pubs/p20-562.pdf 2008 turnout was roughly 131M, approximately 64% of those eligible to vote. 71% of the voting age population was registered to vote. Almost 90% of registered voters polled actually voted. An interesting article is here.

mnottertail -> RE: Why Nate Silver predicts an Obama win. (11/5/2012 5:22:40 AM)
One of the things about questions (and this is huge) people LIE. Did you have sex with that woman? No. Did you vote? Yes. and so on. And those numbers are big, cuz everyone wants to be seen as a decent 'do the right thing' human. Polls are meaningless. Nothing at stake.

Yachtie -> RE: Why Nate Silver predicts an Obama win. (11/5/2012 5:44:12 AM)
quote: ORIGINAL: DarkSteven 4. He terrifies them. Assuming that his model does as well in 2012 as it did in 2008, How did his model work in 2010? What I found interesting is Silver's statement that Romney, having a statistical 16% chance, should he win, the only conclusion is bias in state polling. We shall see, beginning tomorrow :)

DarkSteven -> RE: Why Nate Silver predicts an Obama win. (11/5/2012 5:52:58 AM)
quote: ORIGINAL: Yachtie How did his model work in 2010? US Senate: 34 out of 36. House: He didn't predict individual races as far as I could tell, but predicted a 55 seat GOP gain, which was 8 less than the actual 63 seat gain but within his confidence interval. Governorships: 36 out of 37. Given here.

Moonhead -> RE: Why Nate Silver predicts an Obama win. (11/5/2012 5:52:59 AM)
quote: ORIGINAL: DarkSteven They don't understand probability. It amazes me that otherwise intelligent people are math-illiterate. When Silver says that there's an 80% chance of an Obama win, he is NOT saying that Obama will get 80% of the delegates. Selective innumeracy can be quite a useful tactic for blanking political arguments, particularly if the argument is bearing out a thesis that somebody refuses point blank to accept. There's often examples of that kind of stonewalling in here, as a matter of fact...

cloudboy -> RE: Why Nate Silver predicts an Obama win. (11/5/2012 7:47:39 AM)
Your map doesn't detail where its information is coming from. 538 bases its predictions on collating all the polls and recent polls. From the look of it, Romney is sinking like a stone. Rich likes to point out the white elephant in the room (the economy), but informed voters know (1) that austerity measures haven't worked (which have worsened economies in Europe) and (2) Obama isn't the actual cause of the economy's slow recovery (there's no quick fix for a great recession) and (3) the economy is bouncing back. One side is offering tired propaganda (the Republicans) and the other is offering a modest track record (the Democrats.)

Page: [1] 2 3 4 next > >>

Collarchat.com © 2025

Terms of Service

0.0625