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.

(RedState)   A critique of Nate Silver that doesn't include skinny, gay, or a durty lie-beral? Not even any attacks on math and facts? A reasonable conservative view on current election polls? Yes, yes, and yes   (redstate.com) divider line 159
    More: Interesting, level of difficulty, poll average, Markos Moulitsas, Mondale, Double-A, Jonathan Chait, The Big Short, MBS  
•       •       •

2290 clicks; posted to Politics » on 01 Nov 2012 at 12:01 PM (2 years ago)   |  Favorite    |   share:  Share on Twitter share via Email Share on Facebook   more»



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

Archived thread

First | « | 1 | 2 | 3 | 4 | » | Last | Show all
 
2012-11-01 10:46:14 AM  
Reasonable from Redstate? They declared themselves a fact free zone.
 
2012-11-01 10:50:16 AM  
Wow, the mods at redstate are like trolls with a freaking toy sheriffs stars pinned to their shirts.
 
2012-11-01 11:00:39 AM  
Essentially, the article is refuting Asimov, encouraging its readers to believe their ignorance is, in fact, as good as anyone's knowledge.
 
2012-11-01 11:01:49 AM  
Wow... leftist circle jerk from the start
 
2012-11-01 11:16:55 AM  

Spaz-master: Wow... leftist circle jerk from the start


cry moar
 
2012-11-01 11:20:35 AM  
"Gee, Silver's been right on just about everything in the past, but, uh....this time he's going to be wrong and I think that guy from unskewedpolls.com is on the money. This is because of "poll fundamentals" which I'm only going to hint at but not actually describe. Plus because this time around people are embarrassed to tell pollsters they're voting GOP. I don't know why that didn't happen last time, but trust me, that is what's going on now! So, I'm not going to actually show any numbers to describe this--because as we all know every single poll, even the right-leaning ones like Rasmussen, are totally off--but me and my GOP buddies are absolutely sure Obama is going down to a double-digit defeat."
 
2012-11-01 11:24:37 AM  

The Third Man: "Gee, Silver's been right on just about everything in the past, but, uh....this time he's going to be wrong and I think that guy from unskewedpolls.com is on the money. This is because of "poll fundamentals" which I'm only going to hint at but not actually describe. Plus because this time around people are embarrassed to tell pollsters they're voting GOP. I don't know why that didn't happen last time, but trust me, that is what's going on now! So, I'm not going to actually show any numbers to describe this--because as we all know every single poll, even the right-leaning ones like Rasmussen, are totally off--but me and my GOP buddies are absolutely sure Obama is going down to a double-digit defeat."


I hate doing this, but... um... I have to defend the guy. He says that Unskewed is correct to look at more than just the topline, but then criticizes the entire method he's doing so as "methodological hash."

Honestly, this is the most even-handed and non-screamy attempt at criticizing 538 that I've ever seen. I think he's wrong, but he makes decent arguments throughout the entire piece.
 
2012-11-01 11:25:32 AM  

FlashHarry: Spaz-master: Wow... leftist circle jerk from the start

cry moar


I won't be the one crying Tuesday night
 
2012-11-01 11:26:26 AM  

Rincewind53: The Third Man: "Gee, Silver's been right on just about everything in the past, but, uh....this time he's going to be wrong and I think that guy from unskewedpolls.com is on the money. This is because of "poll fundamentals" which I'm only going to hint at but not actually describe. Plus because this time around people are embarrassed to tell pollsters they're voting GOP. I don't know why that didn't happen last time, but trust me, that is what's going on now! So, I'm not going to actually show any numbers to describe this--because as we all know every single poll, even the right-leaning ones like Rasmussen, are totally off--but me and my GOP buddies are absolutely sure Obama is going down to a double-digit defeat."

I hate doing this, but... um... I have to defend the guy. He says that Unskewed is correct to look at more than just the topline, but then criticizes the entire method he's doing so as "methodological hash."

Honestly, this is the most even-handed and non-screamy attempt at criticizing 538 that I've ever seen. I think he's wrong, but he makes decent arguments throughout the entire piece.


I agree. This is the best critique I have read of Silver's model. And n Red State?
 
2012-11-01 11:26:33 AM  

Spaz-master: Wow... leftist circle jerk from the start


Translantion: Why won't you jerks accept our invented reality yet? *pout*
 
2012-11-01 11:31:15 AM  

vartian: Spaz-master: Wow... leftist circle jerk from the start

TranslantionTranslation: Why won't you jerks accept our invented reality yet? *pout*

FTFY

It's actually more fun watching them dance in the circle... The likelihood of some major leftist butthurt on Tuesday continues to become more likely
 
2012-11-01 11:32:34 AM  
And yes... I used variations of likely twice in a sentence highlight the likelihood of said butthurt
 
2012-11-01 11:34:03 AM  

Spaz-master: vartian: Spaz-master: Wow... leftist circle jerk from the start

TranslantionTranslation: Why won't you jerks accept our invented reality yet? *pout* FTFY

It's actually more fun watching them dance in the circle... The likelihood of some major leftist butthurt on Tuesday continues to become more likely


Well, no. I think Nate's right. "All the polls are wrong" is a pretty bad argument. The only reason I said this guy made decent arguments because he proposed plausible explanations for why all the polls might be wrong. But it's still a bad argument at its heart.
 
2012-11-01 11:36:12 AM  

Spaz-master: FlashHarry: Spaz-master: Wow... leftist circle jerk from the start

cry moar

I won't be the one crying Tuesday night

 

i386.photobucket.com
 
2012-11-01 11:36:44 AM  
I just realized something. Conservatives really hope that Romney will win and Obama will lose. They're ignoring all polls that say otherwise because their hope is so strong. They're audacious in their denial of any data that doesn't fit their ideal.

They've mastered the audacity of hope.
 
2012-11-01 11:38:04 AM  

Rincewind53: The Third Man: "Gee, Silver's been right on just about everything in the past, but, uh....this time he's going to be wrong and I think that guy from unskewedpolls.com is on the money. This is because of "poll fundamentals" which I'm only going to hint at but not actually describe. Plus because this time around people are embarrassed to tell pollsters they're voting GOP. I don't know why that didn't happen last time, but trust me, that is what's going on now! So, I'm not going to actually show any numbers to describe this--because as we all know every single poll, even the right-leaning ones like Rasmussen, are totally off--but me and my GOP buddies are absolutely sure Obama is going down to a double-digit defeat."

I hate doing this, but... um... I have to defend the guy. He says that Unskewed is correct to look at more than just the topline, but then criticizes the entire method he's doing so as "methodological hash."

Honestly, this is the most even-handed and non-screamy attempt at criticizing 538 that I've ever seen. I think he's wrong, but he makes decent arguments throughout the entire piece.


The problem is that he's legitimizing the wrong things. He says, "But if history teaches us anything, it's that the more abuse that's directed towards skeptics, the greater the need for someone to play Socrates," while 9/11 Truthers, Apollo Moon-Landing Hoaxers, and Flat-Earthers nod vigorously in agreement. He explicitly says that the people who were right on the mortgage crisis were not people whose modeling and scientific analysis showed an alternative and bleaker future, but people who just FELT like it was gonna go down. Replace the nouns, and you can use his argument to show why a turtle picking the outcomes of NFL playoffs is as legitimate as Jaworski's.

It's a nice try, and it is probably a genuine attempt at explaining why predictions are predictions and not always right, but he does so by legitimizing quackery because, after all, sometimes the quacks were right and sometimes the experts were wrong. Again- it's exactly the problem Asimov pointed out.
 
2012-11-01 11:39:14 AM  

Spaz-master: vartian: Spaz-master: Wow... leftist circle jerk from the start

TranslantionTranslation: Why won't you jerks accept our invented reality yet? *pout* FTFY


Ty for the spellcheck. Waiting for my meeting to start and farking on my iPad do not mix.

And sweetie, no one cares. If you are so sure your guy is going to win, wonderful. Bully for you. But if you were that sure, you wouldn't be on here biatching about us because we don't agree with you.
 
2012-11-01 11:41:03 AM  

Rincewind53: Spaz-master: vartian: Spaz-master: Wow... leftist circle jerk from the start

TranslantionTranslation: Why won't you jerks accept our invented reality yet? *pout* FTFY

It's actually more fun watching them dance in the circle... The likelihood of some major leftist butthurt on Tuesday continues to become more likely

Well, no. I think Nate's right. "All the polls are wrong" is a pretty bad argument. The only reason I said this guy made decent arguments because he proposed plausible explanations for why all the polls might be wrong. But it's still a bad argument at its heart.


I agree. I think the media has placed too much weight on specific models
 
2012-11-01 11:42:02 AM  

vartian: Spaz-master: vartian: Spaz-master: Wow... leftist circle jerk from the start

TranslantionTranslation: Why won't you jerks accept our invented reality yet? *pout* FTFY

Ty for the spellcheck. Waiting for my meeting to start and farking on my iPad do not mix.

And sweetie, no one cares. If you are so sure your guy is going to win, wonderful. Bully for you. But if you were that sure, you wouldn't be on here biatching about us because we don't agree with you.


Awww... he called me sweetie
//batting my eyes
 
2012-11-01 11:46:44 AM  

Spaz-master: Rincewind53: Spaz-master: vartian: Spaz-master: Wow... leftist circle jerk from the start

TranslantionTranslation: Why won't you jerks accept our invented reality yet? *pout* FTFY

It's actually more fun watching them dance in the circle... The likelihood of some major leftist butthurt on Tuesday continues to become more likely

Well, no. I think Nate's right. "All the polls are wrong" is a pretty bad argument. The only reason I said this guy made decent arguments because he proposed plausible explanations for why all the polls might be wrong. But it's still a bad argument at its heart.

I agree. I think the media has placed too much weight on specific models


And by "specific models" you mean pretty much all of them except Unskewed Polls and Dick Morris.
 
2012-11-01 12:01:49 PM  

Spaz-master: vartian: Spaz-master: Wow... leftist circle jerk from the start

TranslantionTranslation: Why won't you jerks accept our invented reality yet? *pout* FTFY

It's actually more fun watching them dance in the circle... The likelihood of some major leftist butthurt on Tuesday continues to become more likely


It is actually becoming less likely. I like that this guy is making a reasoned argument from a statistical perspective and I think that he also makes some decent points. The trouble is, however, the fact that Silver's weighting system does apply corrections to top line results based, partially at least, on expected voter turnouts. He weights in factors not directly related to polls; economic results, jobs along with poll-specific items: recency, sample size and historical performance. Add to this, he uses FEC donation reports, Gallup voter identity results and a bunch of other measurements to enhance reliability of his model.

In the end, back when Obama was getting a >90% chance in his model, Silver's response was "everybody doing statistical analysis agrees that it will be O by +2%"

Now that voters are making up their minds, less than a week out, this appears to be a very accurate prediction. Look at the methodology section and a lot of the arguments the author of TFA makes in this piece and you'd find he is really reaching.
 
2012-11-01 12:05:44 PM  
Fun fact: Nate's calling out Joe Scarborough on Joe's assertion that nobody can predict the election right now.

Nate said, okay, smart guy. We'll bet on it. If Obama wins, you donate $1,000 to the Red Cross. If Romney wins, I do.
Joe said, here's a better idea, how about we both donate $1,000 to the Red Cross right now?
Nate said, here's an even better idea, we'll make the bet $2,000.
 
2012-11-01 12:06:45 PM  
I guess it was a reasonable argument much the same way saying "The moon landing was faked" is more reasonable than "SQUIRRELS HARVEST THE SOULS OF UNCIRCUMCISED CHILDREN WHEN THE MOTHERSHIP COMES YOU'LL SEE!"

But "all the polls are wrong" is still a fundamentally flawed critique of polling aggregators. Don't people realize that all the "unprecedented" things they say will have to happen for Obama to win are about as "unprecedented" as the massive, fundamental, across-the-board systematic inaccuracy of state polling they're rooting for?

So if something's gotta give in this election, I'm going to go with Occam's Razor here. Partisanship is at an all-time high but math remains neutral.
 
2012-11-01 12:09:13 PM  
i309.photobucket.com
 
2012-11-01 12:11:20 PM  
The author is a homophobic, sexist, racist fascist.

/In Nate Silver We Trust
 
2012-11-01 12:12:02 PM  
Back in 2008 a farker was so sure McCain was going to win he bet farkers a month of total fark he'd beat Obama. Are any farkers making a similar bet for Romney?
 
2012-11-01 12:12:52 PM  

Rincewind53: The Third Man: "Gee, Silver's been right on just about everything in the past, but, uh....this time he's going to be wrong and I think that guy from unskewedpolls.com is on the money. This is because of "poll fundamentals" which I'm only going to hint at but not actually describe. Plus because this time around people are embarrassed to tell pollsters they're voting GOP. I don't know why that didn't happen last time, but trust me, that is what's going on now! So, I'm not going to actually show any numbers to describe this--because as we all know every single poll, even the right-leaning ones like Rasmussen, are totally off--but me and my GOP buddies are absolutely sure Obama is going down to a double-digit defeat."

I hate doing this, but... um... I have to defend the guy. He says that Unskewed is correct to look at more than just the topline, but then criticizes the entire method he's doing so as "methodological hash."

Honestly, this is the most even-handed and non-screamy attempt at criticizing 538 that I've ever seen. I think he's wrong, but he makes decent arguments throughout the entire piece.


He also asserts (falsely) that Romney is ahead in the National polls and heavily implies that only 538 is calling the race for Obama. He even sites RCP, but fails to mention that it also is calling the race for Obama, as well as every betting site and nearly every polling site.
 
2012-11-01 12:13:17 PM  

cameroncrazy1984: Spaz-master: Spaz-master: Elandriel: Wow, that mod MoeLane at that site is amazingly stupid. Someone says "Also professor Wang (snerk) has 93% chance, interesting. We will see what happens on election night", mod presumes guy is speaking on behalf of site, guy explains himself, mod says YOU DIDN'T DO WHAT I TOLD YOU, GUESS WHAT HAPPENS NOW? BANT.

That is just exactly what happens when someone challenges your decent analysis but not very well supported article by citing other decent analysis and slightly though not altogether better supported articles.

Also Spaz-Master, please do us all a favor and quit channeling Phil Herup. Your special brand of trololol is really tiresome, as was his.

You'll be joining us on here Tuesday night then?

I have a feeling you won't be here.


Look: you really can't blame them. People talk trash before the big game in case they can't talk trash after. It's like being int he stands with some drunk Rams fans who've come to watch their team take on the visiting Patriots, insisting how much of a thrashing they'll put on Brady and how the Pats fans are all going to cry when the game is over. By the end of the game, they're no longer around, having left the stadium after the Patriots' lead reached 30 points. So when Ohio and Virginia are called, Spaz will be nowhere to be found.
 
2012-11-01 12:13:22 PM  

Spaz-master: I have no horse in the race other than to watch the fall of prognosticators. A lot of people have gone all in on a weak hand


Yes, that's it. Every prognosticator is wrong. I guess you're not familiar with Occam's Razor.
 
2012-11-01 12:13:32 PM  
I think the best point made in this article, that was shockingly coherent for RedState, is the fact that models can go rapidly off the rails when a key underlying assumption fails to hold true. Mortgage backed securities being a very salient example.

My main takeaway is that I still think 538 maybe the best model out there given what we think we know, but if Romney were to win decisively, it's dumb to conclude that their must of been widespread fraud because 'THE MODELS SAID OBAMA WOULD WIN!!!!111!!!' -

its infinitely more likely that there was an assumption in Nate's models that was flawed than a grand conspiracy of a rigged election
 
2012-11-01 12:13:54 PM  
I love the line of reasoning that since the national polls and the state polls don't agree, that every single state poll must be flawed and the national polls are correct. How hard do I have to bang my head on the concrete to see that as a rational evaluation?
 
2012-11-01 12:14:41 PM  
TL:DR

You know, this is really the first election where i've paid some pretty close attention to the right-wing Derpoverse. I live in a red state, and the radio guys here have gone all-in on Romney, as in A.) Romney is certain to win, B.) If Obama were to win it will be the end of America and things will come crashing down catastrophically.

Do they do this every time? If we wake up on the 7th and Obama is still the HNIC, what do they do then?
 
2012-11-01 12:16:29 PM  

vygramul: cameroncrazy1984: Spaz-master: Spaz-master: Elandriel: Wow, that mod MoeLane at that site is amazingly stupid. Someone says "Also professor Wang (snerk) has 93% chance, interesting. We will see what happens on election night", mod presumes guy is speaking on behalf of site, guy explains himself, mod says YOU DIDN'T DO WHAT I TOLD YOU, GUESS WHAT HAPPENS NOW? BANT.

That is just exactly what happens when someone challenges your decent analysis but not very well supported article by citing other decent analysis and slightly though not altogether better supported articles.

Also Spaz-Master, please do us all a favor and quit channeling Phil Herup. Your special brand of trololol is really tiresome, as was his.

You'll be joining us on here Tuesday night then?

I have a feeling you won't be here.

Look: you really can't blame them. People talk trash before the big game in case they can't talk trash after. It's like being int he stands with some drunk Rams fans who've come to watch their team take on the visiting Patriots, insisting how much of a thrashing they'll put on Brady and how the Pats fans are all going to cry when the game is over. By the end of the game, they're no longer around, having left the stadium after the Patriots' lead reached 30 points. So when Ohio and Virginia are called, Spaz will be nowhere to be found.


I will be here through the night. And will you stay if they are announced in a surprise manner?
I will be here the entire night. Will you?
 
2012-11-01 12:17:30 PM  

ELF Radio:

Do they do this every time? If we wake up on the 7th and Obama is still the HNIC, what do they do then?


Yes, this election has actually been more civil than the last few.
 
2012-11-01 12:18:01 PM  

ELF Radio: TL:DR

You know, this is really the first election where i've paid some pretty close attention to the right-wing Derpoverse. I live in a red state, and the radio guys here have gone all-in on Romney, as in A.) Romney is certain to win, B.) If Obama were to win it will be the end of America and things will come crashing down catastrophically.

Do they do this every time? If we wake up on the 7th and Obama is still the HNIC, what do they do then?


They'll go nuts just like they did four years ago when they said McCain was certain to win. Eventually, they'll win one again. I just hope when they do, it's someone not horrible.
 
2012-11-01 12:19:22 PM  
Anybody up for another trace challenge, like the octsurprise thing? patdollard is now claiming that he's got documentary evidence from a sitting us congressman of something regarding a Libya stand down order. The interview was with "Moon colony Newt", referenced on dailycaller, so we know already it's fake, but there's a ref to a sitting us congressman, that I just don't have time (or inclination) to search for. Any takers?
 
2012-11-01 12:19:39 PM  
The REAL reason the projection is wrong is that it's a proven fact that white people won't vote for a black person. They might say they will, but when they get in the privacy of the voting booth, they just can't do it.
It's a fact. You can look it up.
 
2012-11-01 12:20:00 PM  

cameroncrazy1984: Spaz-master: I have no horse in the race other than to watch the fall of prognosticators. A lot of people have gone all in on a weak hand

Yes, that's it. Every prognosticator is wrong. I guess you're not familiar with Occam's Razor.


Quite. But polls aren't what matters.
Only Tuesday does
 
2012-11-01 12:20:29 PM  
I don't understand. He says the poll "internals" tell a story that Obama is getting clobbered by independents, but wouldn't that be reflected in the topline then? Party affiliation is fluid and is an independent answer on a poll, not a qualifier. Pollsters just take a random sample and they glean party affiliation from answers, just like they glean the candidate they are voting for.
 
2012-11-01 12:20:50 PM  

Spaz-master:

I will be here through the night. And will you stay if they are announced in a surprise manner?
I will be here the entire night. Will you?


You think they'll have a winner election night? I'm betting we don't know Ohio's results until Wednesday morning maybe even Wednesday night.
 
2012-11-01 12:21:33 PM  

jsteiner78: I think the best point made in this article, that was shockingly coherent for RedState, is the fact that models can go rapidly off the rails when a key underlying assumption fails to hold true. Mortgage backed securities being a very salient example.

My main takeaway is that I still think 538 maybe the best model out there given what we think we know, but if Romney were to win decisively, it's dumb to conclude that their must of been widespread fraud because 'THE MODELS SAID OBAMA WOULD WIN!!!!111!!!' -

its infinitely more likely that there was an assumption in Nate's models that was flawed than a grand conspiracy of a rigged election


Another mistake: the model didn't say that Obama will win. It said Obama is more likely to win. It's like the weatherman saying there's a 30% chance of rain today. That's not a license to leave your umbrella at home.
 
2012-11-01 12:21:46 PM  

Elandriel: Wow, that mod MoeLane at that site is amazingly stupid. Someone says "Also professor Wang (snerk) has 93% chance, interesting. We will see what happens on election night", mod presumes guy is speaking on behalf of site, guy explains himself, mod says YOU DIDN'T DO WHAT I TOLD YOU, GUESS WHAT HAPPENS NOW? BANT.

That is just exactly what happens when someone challenges your decent analysis but not very well supported article by citing other decent analysis and slightly though not altogether better supported articles.

Also Spaz-Master, please do us all a favor and quit channeling Phil Herup. Your special brand of trololol is really tiresome, as was his.


In his defense, MoeLane is a self-hating gay man. It's likely he's just directing a bit of that impotent rage in his duties as a mod.
 
2012-11-01 12:22:21 PM  
"And I stand by the view that a mechanical reading of polling averages is an inadequate basis to project an event unprecedented in American history: the re-election of a sitting president without a clear-cut victory in the national popular vote."

Guess it is time to bust out that XKCD comic again?

This is pure wishful thinking on the author's part. There simply isn't a large sample size of close elections (much less close elections where the incumbent is a black man and was the sitting president during the worst economic collapse since the depression) to make us think this is or is not a unfathomable event.

Gore was the VP of a popular president and won the national vote and lose the electoral college. That was unprecidented too.

So what?

Virtually every election has something unprecedented occur.

That XKCD comic is frankly too cute for me, but there is almost always something historic going on. Romney would be the first MORMON for christ sake. Obama was the first black man. etc. etc.
 
2012-11-01 12:22:25 PM  

Rev. Skarekroe: The REAL reason the projection is wrong is that it's a proven fact that white people won't vote for a black person. They might say they will, but when they get in the privacy of the voting booth, they just can't do it.
It's a fact. You can look it up.


Wow... took a while, but here we go? 4 years ago wasn't proof enough for you?

I actually think that inner city issues may play a role though. Nature may have disenfranchised a lot of Philly voters...meaning, Romney will get an unexpected edge in Penn
 
2012-11-01 12:22:48 PM  

Spaz-master: Quite. But polls aren't what matters.
Only Tuesday does


So polls don't matter now? That's not what you said less than 12 hours ago.

Gallop has Romney up for ten straight days above 50%. If Obama wins, will be the first to ever do it under those numbers..
If Silver's is right... yeah.. he can have a church in the jungle


Link
 
2012-11-01 12:23:01 PM  

Elandriel: Wow, that mod MoeLane at that site is amazingly stupid. Someone says "Also professor Wang (snerk) has 93% chance, interesting. We will see what happens on election night", mod presumes guy is speaking on behalf of site, guy explains himself, mod says YOU DIDN'T DO WHAT I TOLD YOU, GUESS WHAT HAPPENS NOW? BANT.


They're pretty much all like that over there. They've flat out stated that liberal views won't be tolerated because they've proven them all wrong so they are justified in removing them, and if you want to have debate you should go elsewhere. They admit it's an echo chamber but it's OK because they are right.
 
2012-11-01 12:23:07 PM  

Carth: Spaz-master:

I will be here through the night. And will you stay if they are announced in a surprise manner?
I will be here the entire night. Will you?

You think they'll have a winner election night? I'm betting we don't know Ohio's results until Wednesday morning maybe even Wednesday night.


Why do you think that? Obama's several points up in Ohio, and leading in enough other states that he doesn't actually need Ohio. The election will be resolved late Tuesday night, easily.
 
2012-11-01 12:23:27 PM  

Carth: Spaz-master:

I will be here through the night. And will you stay if they are announced in a surprise manner?
I will be here the entire night. Will you?

You think they'll have a winner election night? I'm betting we don't know Ohio's results until Wednesday morning maybe even Wednesday night.


That's a possibility. It's also possible with Fl, Va, Penn and NH
 
2012-11-01 12:23:50 PM  
Decent enough article, but it doesn't really back up its argument very well. It was a lot of "This is what Nate does, but it doesn't feel right to me," rather than "This is what Nate does, and here's evidence to the contrary." I nearly stopped reading when he praised the common sense approach over accounting for multiple variables across a historical spectrum of data.
 
2012-11-01 12:23:50 PM  

vygramul: jsteiner78: I think the best point made in this article, that was shockingly coherent for RedState, is the fact that models can go rapidly off the rails when a key underlying assumption fails to hold true. Mortgage backed securities being a very salient example.

My main takeaway is that I still think 538 maybe the best model out there given what we think we know, but if Romney were to win decisively, it's dumb to conclude that their must of been widespread fraud because 'THE MODELS SAID OBAMA WOULD WIN!!!!111!!!' -

its infinitely more likely that there was an assumption in Nate's models that was flawed than a grand conspiracy of a rigged election

Another mistake: the model didn't say that Obama will win. It said Obama is more likely to win. It's like the weatherman saying there's a 30% chance of rain today. That's not a license to leave your umbrella at home.


Unlike the U of CO "model" that used data in it's assumptions that it claimed it was 'predicting'....math majors may commence laughter...
 
Displayed 50 of 159 comments

First | « | 1 | 2 | 3 | 4 | » | Last | Show all

View Voting Results: Smartest and Funniest


This thread is archived, and closed to new comments.

Continue Farking
Submit a Link »
Advertisement
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