| Forum Index > General Forum |
Hey guys! We'll be closing this thread shortly, but we will make an American politics megathread where we can continue the discussions in here. The new thread can be found here: http://www.teamliquid.net/forum/viewmessage.php?topic_id=383301 | ||
|
Defacer
Canada5052 Posts
On November 04 2012 07:56 paralleluniverse wrote: Given all the denial of Nate Silver, this deserves reposting: http://www.teamliquid.net/forum/viewmessage.php?topic_id=330491¤tpage=1140#22796 LOL. <3 What an epic post. | ||
|
Feartheguru
Canada1334 Posts
On November 04 2012 07:56 paralleluniverse wrote: Given all the denial of Nate Silver, this deserves reposting: http://www.teamliquid.net/forum/viewmessage.php?topic_id=330491¤tpage=1140#22796 "Avert your gaze, liberals: Nate Silver admits he's simply averaging public polls and there is no secret sauce." I honestly don't know whether to laugh or cry at this. Guess I'll settle with a smile. | ||
Souma
2nd Worst City in CA8938 Posts
I'm surprised I missed the post the first time. btw paralleluniverse, are you a native Australian living in Australia? An American living in Australia? An Australian living in America? Or what? =o Just curious because you like to use "we" in your posts a lot when discussing U.S. economic policy. | ||
|
jdseemoreglass
United States3773 Posts
| ||
|
paralleluniverse
4065 Posts
On November 04 2012 08:03 Souma wrote: I'm surprised I missed the post the first time. btw paralleluniverse, are you a native Australian living in Australia? An American living in Australia? An Australian living in America? Or what? =o Just curious because you like to use "we" in your posts a lot when discussing U.S. economic policy. Australian living in Australia. I use we all the time, it's a habit. | ||
|
radiatoren
Denmark1907 Posts
On November 04 2012 07:52 paralleluniverse wrote: Because RCP uses an unweighted average that is statistically wrong. Weighting by sample size reduces the standard error of an estimator. Nate Silver weights by sample size and reliability. What you're doing is just denialism and anti-intellectualism. RCP is not very good, but neither very bad. Again it is a question of bullshit in bullshit out and how the best intentions with a correlation of the data based on groupings. The math is technically not a problem in itself, but if your sample is scewed by representation-biases it is less accurate than what you would hope for mathematically. Given that time also factors into the end-result, you end up with MoEs not being compatible because of changes over time. I see Nate Silvers numbers as very professionally handled and he seems to know a lot about the finer techniques of modelling, making his methodology far better than RCP. BUT when the polls themself are carrying +-4% MoEs you truely have to be cautious about how you use them and how much you base on other factors. In the end you cannot really say that states with a percentage chance of a candidate winning in the Nate Silver model is going to be a win for either candidate and that is where the predictability of the polls and aggregations stop and that is where the experience kicks in. A report, here, pointed out that the prediction of simply finding the median of the aggregation has been the best method to predict the winner of an election in the past. Nate Silvers model is very elaborate and if he gets the numbers calibrated enough it has a far better potential, but so far scepticism is warrented. The typical modellers dilemma is in play. Too many variables and your model suffers from their inaccuracies, too many constants and the model will be very hard or impossible to balance correctl | ||
|
[UoN]Sentinel
United States11320 Posts
On November 04 2012 08:07 jdseemoreglass wrote: I feel kind of bad for the conservatives who are doing psychological jumping jacks trying to avoid what has been obvious for a while now, and I'm glad I'm not emotionally invested in this election at all. The first time Obama came center stage over 4 years ago I knew he'd be a 2 term president immediately. He did some things I really disagree with, but he's not the anti-Christ. I dare say he was better than W at least. This. I'm a Republican and I disagree with a lot of Obama's policies, but I think he's the man who can at least somewhat restore the economic state of this country. Romney could too and either recovery will be insane, but Obama's got his plan more laid out. | ||
|
Darknat
United States122 Posts
On November 04 2012 08:20 [UoN]Sentinel wrote: This. I'm a Republican and I disagree with a lot of Obama's policies, but I think he's the man who can at least somewhat restore the economic state of this country. Romney could too and either recovery will be insane, but Obama's got his plan more laid out. I can't help but think you're black and are really voting based on race(like Colin Powell). User was banned for this post. | ||
|
oneofthem
Cayman Islands24199 Posts
| ||
|
[UoN]Sentinel
United States11320 Posts
On November 04 2012 08:27 Darknat wrote: I can't help but think you're black and are really voting based on race(like Colin Powell). I'm as white as they get and supported McCain in 2008. | ||
|
mynameisgreat11
599 Posts
On November 04 2012 08:27 Darknat wrote: I can't help but think you're black and are really voting based on race(like Colin Powell). Get this post in the hall of fame pronto. | ||
|
ticklishmusic
United States15977 Posts
| ||
|
kwizach
3658 Posts
On November 04 2012 08:27 Darknat wrote: I can't help but think you're black and are really voting based on race(like Colin Powell). Are you actually being serious here, or are you sarcastically poking fun at the idiotic statements of some Republicans regarding Powell's endorsement? | ||
|
Jumbled
1543 Posts
On November 04 2012 08:34 kwizach wrote: Are you actually being serious here, or are you sarcastically poking fun at the idiotic statements of some Republicans regarding Powell's endorsement? Normally I'd think it was a joke, but having seen some of his previous posts I'm not so sure. | ||
|
Feartheguru
Canada1334 Posts
On November 04 2012 08:32 ticklishmusic wrote: Alas Darknat, we hardly knew him. lol I was totally gonna post those exact words. | ||
|
Jaaaaasper
United States10225 Posts
On November 04 2012 08:27 Darknat wrote: I can't help but think you're black and are really voting based on race(like Colin Powell). Best post in the thread right there. Just because someone supports Obama does not mean they are black or are voting based on race if they are. I didn't want to vote for Obama at first, but Romney appeases his opponents and allies far more than I like. Compromise is one thing, changing your opinion depending on who your talking to is something else entirely. | ||
|
frogrubdown
1266 Posts
On November 04 2012 08:01 Feartheguru wrote: "Avert your gaze, liberals: Nate Silver admits he's simply averaging public polls and there is no secret sauce." I honestly don't know whether to laugh or cry at this. Guess I'll settle with a smile. Someday the Bayesians will be in charge. Someday... | ||
|
oneofthem
Cayman Islands24199 Posts
bayesians have it as obama by a landslide, volcano slide? something big | ||
|
TheTenthDoc
United States9561 Posts
On November 04 2012 08:20 radiatoren wrote: RCP is not very good, but neither very bad. Again it is a question of bullshit in bullshit out and how the best intentions with a correlation of the data based on groupings. The math is technically not a problem in itself, but if your sample is scewed by representation-biases it is less accurate than what you would hope for mathematically. Given that time also factors into the end-result, you end up with MoEs not being compatible because of changes over time. I see Nate Silvers numbers as very professionally handled and he seems to know a lot about the finer techniques of modelling, making his methodology far better than RCP. BUT when the polls themself are carrying +-4% MoEs you truely have to be cautious about how you use them and how much you base on other factors. In the end you cannot really say that states with a percentage chance of a candidate winning in the Nate Silver model is going to be a win for either candidate and that is where the predictability of the polls and aggregations stop and that is where the experience kicks in. A report, here, pointed out that the prediction of simply finding the median of the aggregation has been the best method to predict the winner of an election in the past. Nate Silvers model is very elaborate and if he gets the numbers calibrated enough it has a far better potential, but so far scepticism is warrented. The typical modellers dilemma is in play. Too many variables and your model suffers from their inaccuracies, too many constants and the model will be very hard or impossible to balance correctl Ah, but as Silver pointed out in his newest blog post his model incorporates the possibilities that the polls are systematically in error. Plus he has poll data since 1968 to boost his numbers. I mean, as he points out in the post-let's ignore all the models. The day after a candidate leads in 19/22 polls of swing states, you shouldn't say the race is a toss-up unless you have a cogent argument the polls are systematically biased (which the media sure as hell doesn't have, yet they constantly call it a toss-up). | ||
|
frogrubdown
1266 Posts
On November 04 2012 08:46 oneofthem wrote: http://election.princeton.edu/ bayesians have it as obama by a landslide, volcano slide? something big Eh, they're not the only Bayesians in town. Personally I'd put more credence in 538. | ||
| ||
StarCraft 2 StarCraft: Brood War Dota 2 League of Legends Counter-Strike Other Games |
|
KCM Race Survival
IntoTheTV X SOOP
SHIN vs GuMiho
Clem vs Classic
OSC
WardiTV Invitational
GSL
Epic.LAN
Replay Cast
Sparkling Tuna Cup
Epic.LAN
AI Arena Tournament
[ Show More ] |
|
|
EPT
