Do you have adblock enabled?
 
If you can read this, either the style sheet didn't load or you have an older browser that doesn't support style sheets. Try clearing your browser cache and refreshing the page.

(Examiner)   Unskewed Pollman: "Nate Silver can't be trusted because he's thin and might be gay, too"   (examiner.com ) divider line
    More: Dumbass, Baseball Prospectus, career development, Fantasyland, swing vote, Dick Morris, swing states  
•       •       •

4757 clicks; posted to Politics » on 27 Oct 2012 at 10:51 AM (3 years ago)   |   Favorite    |   share:  Share on Twitter share via Email Share on Facebook   more»



381 Comments   (+0 »)
   
View Voting Results: Smartest and Funniest

Archived thread

First | « | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | » | Last | Show all
 
2012-10-27 09:31:07 PM  

FormlessOne: the GOP can't have the country successful under the obviously tyrannical rule of a half-black secret Muslim socialist Kenyan Manchurian Candidate intellectual Democrat, now, can they?


This attitude is so prevalent in the republican party and it's absolutely despicable. IMO it's the biggest difference between the two party's supporters. You dont see this kind of obstructionism* among the Democrats.

*obstructionism for the purpose of avoiding success, as opposed to obstructionism due to political idealogy
 
2012-10-27 09:38:11 PM  
He doesn't look gay. he looks like a Jewish math nerd.
 
2012-10-27 09:38:44 PM  

abb3w: This clown apparently misses at least three things.

1) Nate's weighing mechanism isn't ultimately based on how much the polls disagree with "his own average", but on how their predictions differed from the election outcome in 2008. As happens, the polls on average were biased below the measurement threshold, but some pollsters consistently gave results further from the final outcome than others, and/or consistently gave results consistently biased D or biased R. Nate normalizes to cancel out this, based on past measurement from electoral outcome.

2) The odds he assigns are different from the polling percentages, because the odds are derived from the polls combined with their uncertainty measure. A poll running 60%R 40%D with the usual ±3 or so does not mean the Democratic candidate has a 40% chance of winning, it means he has a chance of winning less than a Powerball ticket.The ± is the 95% confidence interval, which is around two standard deviations either side of the mean; winning with a 20 point difference requires results results 13 sigma above the mean, which means "Ain't Happening" (without impact from new news not considered by those polled). However, if it was instead a R 50.7% D 49.2% race with the same uncertainty, a D win would only require a result about one sigma above the mean, yielding a circa 15% chance (vs 85% for the R), despite a race "inside the margin of error".


Could I perhaps trouble you to rephrase this second point in plainer English?
 
2012-10-27 09:47:44 PM  
My favorite Nate Silver picture

www.details.com

Details Magazine. Hmmm....
 
2012-10-27 09:56:43 PM  

Fista-Phobia: [i1199.photobucket.com image 831x143]


i1199.photobucket.com
i49.tinypic.com
fim.413chan.net
 
2012-10-27 09:57:00 PM  

Huggermugger: Three Crooked Squirrels: Nate Silver can't be trusted because when he talks, he kinda sounds gay. But Dick Morris can be trusted because in this guy's world, Dick Morris doesn't sound gay? I know he was busted for sucking a female hooker's toes, but that guy really sounds gay.

Dick Morris has some of the weirdest, most effeminate mannerisms I've ever seen on an allegedly heterosexual man. I have honestly never understood why he was given such a prominent role in Fox News. He's not physically attractive, his voice is annoying, he doesn't say anything intelligent - it's all overblown belligerence, he's morally corrupt, and he has no credibility outside that Fox bubble. They could have found any number of superior personalities to fill his slot, so I'm assuming that he's entirely coasting on his past association with the Clintons, because he's got absolutely nothing else going for him. He's very repulsive.


I remember right after Bush gave that prime time speech from New Orleans after Katrina, Morris declared that Bush's popularity "Would SOAR."
 
2012-10-27 09:58:01 PM  

Fista-Phobia: [i1199.photobucket.com image 831x143]

[i.imgur.com image 400x300]


My favorite comment to his post: ...Hey fatass, unskew your weight
 
2012-10-27 10:19:34 PM  

cloud_van_dame: My favorite Nate Silver picture

[www.details.com image 620x430]

Details Magazine. Hmmm....


Heh.
 
2012-10-27 10:28:04 PM  
Oh geeze, I'm almost crying from laughing so much with all the posts with the pictures of the Unscewed Polls fool and the accomponing comments with them. 😂Thanks Farkers.
 
2012-10-27 10:34:46 PM  

Altitude5280: Oh geeze, I'm almost crying from laughing so much with all the posts with the pictures of the Unscewed Polls fool and the accomponing comments with them. 😂Thanks Farkers.


Read through the comments in TFA, they had me rolling too
 
2012-10-27 10:49:15 PM  

log_jammin: I have a feeling we will never hear anything else about "UnSkewedPolls.com" after November.


I'll be sure to submit something in December.
 
2012-10-27 11:39:26 PM  
1.bp.blogspot.com
would be gay, but just can't get into that kinda shape.
 
2012-10-27 11:55:13 PM  
I just reread the article in Louis Anderson's voice. could really take that on the road.

www.starscolor.com
 
2012-10-28 12:02:07 AM  

Altitude5280: Huggermugger: Three Crooked Squirrels: Nate Silver can't be trusted because when he talks, he kinda sounds gay. But Dick Morris can be trusted because in this guy's world, Dick Morris doesn't sound gay? I know he was busted for sucking a female hooker's toes, but that guy really sounds gay.

Dick Morris has some of the weirdest, most effeminate mannerisms I've ever seen on an allegedly heterosexual man. I have honestly never understood why he was given such a prominent role in Fox News. He's not physically attractive, his voice is annoying, he doesn't say anything intelligent - it's all overblown belligerence, he's morally corrupt, and he has no credibility outside that Fox bubble. They could have found any number of superior personalities to fill his slot, so I'm assuming that he's entirely coasting on his past association with the Clintons, because he's got absolutely nothing else going for him. He's very repulsive.

I remember right after Bush gave that prime time speech from New Orleans after Katrina, Morris declared that Bush's popularity "Would SOAR."


LOLWUT?
 
2012-10-28 12:23:37 AM  

PsiChick: he might turn out to be an atheist or even some eclectic religion you've never heard of.


Great. Hipster religions.
 
2012-10-28 12:29:06 AM  
upload.wikimedia.org
 
2012-10-28 12:35:48 AM  

Cuthbert Allgood: Dude, break that down for some conservatives and report your findings.


Sorry, I don't speak moron. The best I can do is post to Fark, and hope someone with a troll alt can provide a translation.
 
2012-10-28 01:02:31 AM  

abb3w: However, if it was instead a R 50.7% D 49.2% race with the same uncertainty, a D win would only require a result about one sigma above the mean, yielding a circa 15% chance (vs 85% for the R), despite a race "inside the margin of error".


Do you have that right? Because if one goes up 1% the other usually goes down 1%. Meaning the +- error should be treated as double (basically). My calculations give a 50.7% to 49.2% with a 3% margin of error about a 66% 33% chance. (I can post my R code if you want it - the results are pretty close to Silver's)

Still, your point is valid.
 
2012-10-28 01:23:53 AM  

mistersnark: Could I perhaps trouble you to rephrase this second point in plainer English?


Probably in the morning, anyway.

impaler: Do you have that right? Because if one goes up 1% the other usually goes down 1%.


I think so; the ± is usually for the value of any candidate, not the margin between.
Contrariwise, it seems incidental to the basic point of "candidate polling percentages are not election victory odds"; nohow, a simple factor-of-two math mistake at some step is well within my ability to screw up.
 
2012-10-28 01:49:33 AM  

d-fens99: Lionel Mandrake: [i159.photobucket.com image 401x271]

UnSkewed Polls guy can't be trusted because he's a blob of pus and serves greasy pork sandwiches in dirty ashtrays.

You know, I checked the computing power circa 1985, and it turns out my iphone 4 is way more powerful than the computers used to make the girl in that movie. So, where's my make a babe app?

[2.bp.blogspot.com image 430x640]


I have the Weirdest Science boner right now.
 
2012-10-28 01:57:10 AM  

crab66: Anyone posted this yet?

[static5.businessinsider.com image 400x300]


3.bp.blogspot.com
 
2012-10-28 02:04:18 AM  

Blowmonkey: Conservatives are flailing because it's becoming apparent they are going to lose again and are creating a narrative that forestalls the inevitable.


What I think you really mean is that they're creating a narrative that allows them to delegitimize the President. *Especially* if he manages to lose the popular vote but still win the EV vote. That's pretty much been their MO from Day One. If he wins, it's because the American people saw through the bullshiat, realized he's a failure, and voted Romney. If Obama loses, it's because either (a) The American people are stupid, easily fooled Takers, or (b) The election was stolen by ACORN and illegal Mexicans voting. 

You can see it everywhere - he's a Marxist! He's not a citizen! Voter Fraud!

In either case, he's not really a legitimate President. They did it with Bill Clinton, too.

/loved Digging Your Scene, btw.
 
2012-10-28 03:49:53 AM  

entitygm: I want to see a map like this. Early call in Ohio and Fox claim it's a lock for Romney, celebration starts early.... then they cry themselves to sleep.

[i46.tinypic.com image 755x493]


Naw, won't happen. Obama has Ohio pretty well sewn up and out of the remaining toss-ups, the only state where it's likely Romney will pull off a win is Florida. Still leaves him like 40 EV short on the night.
 
2012-10-28 06:45:30 AM  

log_jammin: I have a feeling we will never hear anything else about "UnSkewedPolls.com" after November.


Until 2016
 
2012-10-28 07:57:17 AM  
Mommy! The gay man is looking at me again! He's giving me certain urges that I don't understand! Mommy! Make him stop!

Moooommmmmmmmmy!
 
2012-10-28 11:00:02 AM  

Precious Roy's Horse Dividers: Ok which one of you Farkettes is Megan Mills from the comment section? I want to marry you.


FTFCMy dog ate 2lbs of crisco and took a dump that looked remarkably similar to Dean, then my dog got hungry and re-ate it. Second time around was a real charm, thanks for the memories Dean.
 
2012-10-28 02:12:52 PM  
Today I was bored so I did something unscientific, but interesting to me. I took the recent state-by-state polls, and chose those that were least favorable to Obama, and plugged them into each state. I used the 2008 results to project a vote total for each state, and for the states that didn't have polls, I downgraded Obama's votes (and upgraded Romney's votes) by the average change from 2008 in the states that did have polls, and then calculated a national vote total.

The result is that Romney wins the popular vote 50.4%-48.5%, and leads the electoral college 267-227, with Ohio, Michigan and Wisconsin tied (all of which could break for Obama). Remember, that's the *worst* recent polls for Obama. (The "best" polls has Obama winning 51.5%-47.2%, and winning the electoral college 343-180 with NC tied).

So I think Nate's percentage seems right-on to me.
 
2012-10-28 03:26:28 PM  

abb3w: 2) The odds he assigns are different from the polling percentages, because the odds are derived from the polls combined with their uncertainty measure. A poll running 60%R 40%D with the usual ±3 or so does not mean the Democratic candidate has a 40% chance of winning, it means he has a chance of winning less than a Powerball ticket.The ± is the 95% confidence interval, which is around two standard deviations either side of the mean; winning with a 20 point difference requires results results 13 sigma above the mean, which means "Ain't Happening" (without impact from new news not considered by those polled). However, if it was instead a R 50.7% D 49.2% race with the same uncertainty, a D win would only require a result about one sigma above the mean, yielding a circa 15% chance (vs 85% for the R), despite a race "inside the margin of error".

mistersnark: Could I perhaps trouble you to rephrase this second point in plainer English?


So, the ± refers to the confidence interval; it's a function of sample size. There's a decent entry in wikipedia on it. Most polls report one based at the 95% confidence standard, so a poll showing a race of 60%R 40%D and ±3% means that there's statistically about a 47.5% chance of the republican getting 60.0-63.0% of the vote, and 47.5% chance of the republican getting 57.0-60.0% of the vote. The ±3% is typical for polls asking on the order of 1000 people randomly sampled from the whole population.

One of the most common curves in statistics is the Gaussian distribution; one of the parameters of the usual expression of the distribution function is the standard deviation, denoted by a lower case sigma, which effectively scales the axis. The distribution is symmetric, but decreases with distance; about 34% of the distribution falls within zero and one sigma above the midpoint, but only about 13.6 between one and two.

upload.wikimedia.org



Conveniently for rule-of-thumb work, within two sigma on either side of the mean gives 95.4% of the distribution. The actual distribution for polling results isn't exactly a Gaussian (which trails off to infinity, whereas a vote outcome of 107% for one candidate is a bit hinky), but it's something similar enough for rule-of-thumb work in polling where the race isn't worse than 2:1.

So, cut the ±3% in half, and you get 1.5% is the standard deviation "sigma".

Which brings up...

impaler: My calculations give a 50.7% to 49.2% with a 3% margin of error about a 66% 33% chance.


Yeah, in the cold grey light of afternoon and sobriety, sounds right.

For a (60:40)±3 race, anything 50.001% or more for the underdog is 10 percent off from survey, which is 6.6 sigma out -- which is not 13ish, but still one-in-ten-billion Powerball territory. (Lucky me for understating how bad that was the first time.) For a (50.7:49.2)±3 race, that victory threshold is only 0.8ish off from the poll, which is about 0.5 times the sigma distance of 1.5%, which is (yes) somewhere about 30-40% chance for the underdog. Using a statistics package like R will give more precise answers and allow more exact calculations; but knowing a few points on the gaussian cumulative distribution (and not slipping in a careless factor of two) gets to rule-of-thumb understanding that Nate Silver is not pulling these numbers out of his backside.

There's other factors -- the odds of how undecideds break, the flexibility of the electorate's attitudes with time and new news, and so on. Nate's worked most of those into his model, and written up explanations on what he's done, for the curious to read if they care to improve their understanding.

However, political conservatives tend not to value curiosity. Maybe it's just too hard for them.
 
2012-10-28 03:42:44 PM  
OK - he's chubby, but honestly Dean Chambers sends out LOTS of gay vibes. He looks the type that DESPERATELY wants to be dominated by a "daddy" type, who then likes to "loan" his boy out to his buddies.
 
2012-10-29 12:54:59 AM  

Frederick: Osomatic: Frederick: log_jammin: Yeah. If Obama wins, they'll wake up the next day angry at the world but will never consider the possibility they were lied to and lied to them selves. I don't get how people go through life like that.

It's a baffling phenomenon to me; the people who can successfully lie to themselves. In my experience there is a correlation between those people and early life religious indoctrination.

You may have something there, but... what about people who get no religious indoctrination? We've been raising our child as, basically, "nothing." We're not religious, but we don't say religious people are bad. We tell him that it's up to him what to believe without telling him what we (his mom and I) believe, which is, well, nothing. When he came back from his pre-school filled with ideas about God, we said "well, that's what some people believe." I like to think we're raising him to create his own beliefs, but... sometimes I wonder. We've also tried to show him and talk to him about what science says, without being dogmatic about it. I fear we may have been a bit dogmatic about stuff like psychics and astrology, though. But I feel okay about that, really. That stuff is just stupid - I don't think anybody would consider us to be "indoctrinating" him if we warned him off of false advertising, for instance. Still, we've done our very best to be hands-off with religion.

And I guess I wonder... this is a *really* smart kid, but one who has been actually diagnosed with at least mild ADD or ADHD... what is he going to end up believing? I wish I knew. We tried so hard to just leave it open but I'm not sure we succeeded.

When as a kid I began to question obvious flaws in the Christian bible and other seemingly bright kids around me were not, I made note of it. I never understood how they could reconcile logical flaws with religious dogma.

I was raised in a similar environment to your son. Both my parents grew up in the church and then had a ...


Not necessarily, but... I still don't know what this kid is going to end up being.
 
2012-10-29 09:32:55 AM  

abb3w: abb3w: 2) The odds he assigns are different from the polling percentages, because the odds are derived from the polls combined with their uncertainty measure. A poll running 60%R 40%D with the usual ±3 or so does not mean the Democratic candidate has a 40% chance of winning, it means he has a chance of winning less than a Powerball ticket.The ± is the 95% confidence interval, which is around two standard deviations either side of the mean; winning with a 20 point difference requires results results 13 sigma above the mean, which means "Ain't Happening" (without impact from new news not considered by those polled). However, if it was instead a R 50.7% D 49.2% race with the same uncertainty, a D win would only require a result about one sigma above the mean, yielding a circa 15% chance (vs 85% for the R), despite a race "inside the margin of error".
mistersnark: Could I perhaps trouble you to rephrase this second point in plainer English?

So, the ± refers to the confidence interval; it's a function of sample size.


Wow. Thanks for putting it in terms even a History major can understand.
 
Displayed 31 of 381 comments

First | « | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | » | Last | Show all

View Voting Results: Smartest and Funniest


This thread is archived, and closed to new comments.

Continue Farking
Submit a Link »
On Twitter






In Other Media


  1. Links are submitted by members of the Fark community.

  2. When community members submit a link, they also write a custom headline for the story.

  3. Other Farkers comment on the links. This is the number of comments. Click here to read them.

  4. Click here to submit a link.

Report