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.

(NPR)   The night a computer predicted the next president. Otherwise known as Nate Silver v1.0   (npr.org) divider line 29
    More: Interesting, artificial brain, Dr. J, election night, Walter Cronkite, computers, associate dean  
•       •       •

6461 clicks; posted to Geek » on 01 Nov 2012 at 1:10 PM (2 years ago)   |  Favorite    |   share:  Share on Twitter share via Email Share on Facebook   more»



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

Archived thread
 
2012-11-01 11:16:04 AM  
You could be using a Cray supercomputer and it would be useless without an intelligently-designed model. That requires human intervention.

Just ask Dean Chambers this time next week. Or don't, since he'll still be in denial.
 
2012-11-01 12:17:53 PM  
Just like how the computerized voting machines will help "predict" Romney winning on November 6.
 
2012-11-01 01:25:11 PM  
The computer doesn't matter. Nate's calculations could probably all be done in Excel. It is all about the model of assigning accuracy (and type of accuracy) to the various different polling companies and methods. Then when the data comes in it can be corrected and an actually accurate result can be found. I am actually quite curious how well Silver's model holds up in five days.
 
2012-11-01 01:26:07 PM  
Ah yes, this is from back in the days when there was actually a little bit of daylight between when networks called elections. Now it's typically within, like, seconds of each other. And they still dickwave about who was first over the line. (Never mind how that behavior got all of them in trouble in 2000.)
 
2012-11-01 01:28:18 PM  

madgonad: The computer doesn't matter. Nate's calculations could probably all be done in Excel. It is all about the model of assigning accuracy (and type of accuracy) to the various different polling companies and methods. Then when the data comes in it can be corrected and an actually accurate result can be found. I am actually quite curious how well Silver's model holds up in five days.


He's had elections to go on in between 2008 and now. There's been the occasional hiccup, but those have been rare. He hasn't been calling the House; that's where he's been least accurate.
 
2012-11-01 01:35:31 PM  

Gosling: Ah yes, this is from back in the days when there was actually a little bit of daylight between when networks called elections. Now it's typically within, like, seconds of each other. And they still dickwave about who was first over the line. (Never mind how that behavior got all of them in trouble in 2000.)


Now with holograms!
 
2012-11-01 01:36:21 PM  

madgonad: Nate's calculations could probably all be done in Excel.


The hard part would be porting Excel to the UNIVAC.
 
2012-11-01 02:17:50 PM  

Gosling: He's had elections to go on in between 2008 and now. There's been the occasional hiccup, but those have been rare. He hasn't been calling the House; that's where he's been least accurate.


I really wish he's provide some data on the House, but I know that would take a lot more manpower considering your ultimately talking about 435 different races.
 
2012-11-01 02:22:53 PM  
You don't really need a computer. Just poll Ohio and Florida since they decide all presidential elections now.
 
2012-11-01 02:23:31 PM  

madgonad: The computer doesn't matter. Nate's calculations could probably all be done in Excel.


Just don't use a Pentium chip.
 
2012-11-01 02:38:26 PM  
The computer predicted Eisenhower to win by a 26% margin...he actually won by 10.9% . It also got a fair number of states wrong (It predicted Stevenson winning 5 states large enough to total 93 EV...Stevenson actually won 9 much smaller states).
 
2012-11-01 02:47:03 PM  

Gosling: madgonad: The computer doesn't matter. Nate's calculations could probably all be done in Excel. It is all about the model of assigning accuracy (and type of accuracy) to the various different polling companies and methods. Then when the data comes in it can be corrected and an actually accurate result can be found. I am actually quite curious how well Silver's model holds up in five days.

He's had elections to go on in between 2008 and now. There's been the occasional hiccup, but those have been rare. He hasn't been calling the House; that's where he's been least accurate.


Makes sense. House races are generally very poorly polled. Hard to calculate an accurate model when you have very little data to work off of.
 
2012-11-01 03:17:22 PM  

FishyFred: You could be using a Cray supercomputer and it would be useless without an intelligently-designed model. That requires human intervention.

Just ask Dean Chambers this time next week. Or don't, since he'll still be in denial.


i1199.photobucket.com
i49.tinypic.com
 
2012-11-01 03:21:23 PM  

TV's Vinnie: FishyFred: You could be using a Cray supercomputer and it would be useless without an intelligently-designed model. That requires human intervention.

Just ask Dean Chambers this time next week. Or don't, since he'll still be in denial.

[i1199.photobucket.com image 831x143]
[i49.tinypic.com image 360x270]


Pablum pukers?

Is that you Morton Downey, jr
 
2012-11-01 03:24:59 PM  

TV's Vinnie: FishyFred: You could be using a Cray supercomputer and it would be useless without an intelligently-designed model. That requires human intervention.

Just ask Dean Chambers this time next week. Or don't, since he'll still be in denial.

[i1199.photobucket.com image 831x143]
[i49.tinypic.com image 360x270]


There's something deeply ironic about somebody concocting a phrase like "pabulum pukers".
 
2012-11-01 03:54:54 PM  

madgonad: Is that you Morton Downey, jr


I'd take anything he said with a grain of salt
 
2012-11-01 04:04:31 PM  

angry_scientist: madgonad: Is that you Morton Downey, jr

I'd take anything he said with a grain of salt


Or a shovel of dirt. He died eleven years ago.

/did I just prove myself a dumbass?
 
2012-11-01 04:30:27 PM  

madgonad: angry_scientist: madgonad: Is that you Morton Downey, jr

I'd take anything he said with a grain of salt

Or a shovel of dirt. He died eleven years ago.

/did I just prove myself a dumbass?


No, no, it's still work hours
images.mylot.com
 
2012-11-01 04:31:41 PM  
gah it cut off my pic of Morton Salt, oh well
 
2012-11-01 04:36:03 PM  

jbtilley: You don't really need a computer. Just poll Ohio and Florida since they decide all presidential elections now.


There's a lot of truth to that, but this year Romney could win both FL and OH and still lose the election if Obama takes NV, CO, IA, and VA.

In fact, if both of the iconic "swing states" go for the same guy and he ends up losing the election, it will make me wonder what kind of discussion we'll be having about "swing states" 4 years from now. I'm sure there will be tons of articles entitled "Myth of the Swing State" or "Are Swing States a Thing of the Past?" or what have you...
 
2012-11-01 04:56:11 PM  

angry_scientist: gah it cut off my pic of Morton Salt, oh well


I'm seeing it. Refresh your page a few times.
 
2012-11-01 04:59:58 PM  
1.bp.blogspot.com
 
2012-11-01 05:32:08 PM  

MindStalker: Gosling: He's had elections to go on in between 2008 and now. There's been the occasional hiccup, but those have been rare. He hasn't been calling the House; that's where he's been least accurate.

I really wish he's provide some data on the House, but I know that would take a lot more manpower considering your ultimately talking about 435 different races.


The guy over at electionprojection is doing it. He's saying there will basically be no shift in the makeup of Congress.
 
2012-11-01 05:41:33 PM  

odinsposse: MindStalker: Gosling: He's had elections to go on in between 2008 and now. There's been the occasional hiccup, but those have been rare. He hasn't been calling the House; that's where he's been least accurate.

I really wish he's provide some data on the House, but I know that would take a lot more manpower considering your ultimately talking about 435 different races.

The guy over at electionprojection is doing it. He's saying there will basically be no shift in the makeup of Congress.


Man I find that so hard to believe.....
 
2012-11-01 06:11:47 PM  

Blue_Blazer: odinsposse: MindStalker: Gosling: He's had elections to go on in between 2008 and now. There's been the occasional hiccup, but those have been rare. He hasn't been calling the House; that's where he's been least accurate.

I really wish he's provide some data on the House, but I know that would take a lot more manpower considering your ultimately talking about 435 different races.

The guy over at electionprojection is doing it. He's saying there will basically be no shift in the makeup of Congress.

Man I find that so hard to believe.....


His math looks good,
A bunch of redistricting combined with a bunch of retirement.
11 D districts vanished, 8 R

7 D uncontest in NEW districts, 7 R (non are from vanishing districts all who are rerunning from both parties are contested)
 
2012-11-01 07:26:21 PM  
Should be down right amusing to see people scream about how dead on accurate Nate Silver was no matter what happens.

Alternately if he's wrong and Romney wins it will be like the 2004 election again where somehow the polls, predictions, and exit polling are more accurate than the actual vote.

I'll give you a hint: Anyone paying the least bit of attention can accurately predict 90-95% of all races.
 
2012-11-02 01:17:24 AM  

Blue_Blazer: odinsposse: MindStalker: Gosling: He's had elections to go on in between 2008 and now. There's been the occasional hiccup, but those have been rare. He hasn't been calling the House; that's where he's been least accurate.

I really wish he's provide some data on the House, but I know that would take a lot more manpower considering your ultimately talking about 435 different races.

The guy over at electionprojection is doing it. He's saying there will basically be no shift in the makeup of Congress.

Man I find that so hard to believe.....


He has Duckworth winning and calling Bachmann's seat only lean GOP, so he's at least he's being somewhat honest.
 
2012-11-02 03:37:35 AM  

randomjsa: Alternately if he's wrong and Romney wins it will be like the 2004 election again where somehow the polls, predictions, and exit polling are more accurate than the actual vote.


It's well known that the limitations of Zerko-Frankel set theoretical foundations appear in certain edge cases best expressed in nodal singularities isomorphic to certain counties in Ohio..
 
2012-11-02 06:23:12 PM  
Of course they fail to mention the dwarf inside the terminal who flipped a coin and typed out the result on green bar paper that fed up and out through the top of the broken typewriter sitting beside it.
 
Displayed 29 of 29 comments

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