\
  The most prestigious law school admissions discussion board in the world.
BackRefresh Options Favorite

Is anyone sweating more about the midterms than Nate Silver?

Last time he called Trump at 20% odds of winning. When Trump...
bateful cuckold
  10/22/18
Those two events are not IID so you can't conclude the 3% pr...
Poppy beady-eyed pisswyrm kitty
  10/22/18
Yes they’re not independent but it’s a more forg...
Mind-boggling fiercely-loyal trust fund
  10/22/18
His "model" is little more than an average of poll...
Claret stimulating hospital
  10/22/18
Yes. As you say what his model “adds” are assump...
Mind-boggling fiercely-loyal trust fund
  10/22/18
that doesn't make any sense
Opaque dashing new version library
  10/22/18
Yes it does
Mind-boggling fiercely-loyal trust fund
  10/22/18
it makes no sense to sample selectively when his entire body...
Opaque dashing new version library
  10/22/18
It makes more sense if trump era polling is different from n...
Mind-boggling fiercely-loyal trust fund
  10/22/18
has his method changed?
Opaque dashing new version library
  10/22/18
I’m saying that the same method could be previously ac...
Mind-boggling fiercely-loyal trust fund
  10/22/18
maybe, but that's on pollsters. silver's analysis is only as...
Opaque dashing new version library
  10/22/18
I don’t agree. Silver picks and chooses which polls to...
Mind-boggling fiercely-loyal trust fund
  10/22/18
He says he constantly adjusts his model based on the most re...
Claret stimulating hospital
  10/22/18
Cr
Mind-boggling fiercely-loyal trust fund
  10/22/18
RBG
Greedy Confused Striped Hyena
  10/22/18
credited answer
puce doctorate wagecucks
  10/22/18
Tbf, she is sweating the 2020 general not the midterms. She&...
Mind-boggling fiercely-loyal trust fund
  10/22/18


Poast new message in this thread



Reply Favorite

Date: October 22nd, 2018 12:24 AM
Author: bateful cuckold

Last time he called Trump at 20% odds of winning. When Trump won Silver reasonably said that 20% odds aren’t nothing and should be expected to win sometimes.

Now he has the Dems at 85% chance of retaking the house. The odds of a 20% probability event occurring followed by a 15% probability event are 3%. So if he blows this it looks a lot more like his models are all fraud than hes correctly calling 3% probability events.

(http://www.autoadmit.com/thread.php?thread_id=4112150&forum_id=2#37070364)



Reply Favorite

Date: October 22nd, 2018 12:35 AM
Author: Poppy beady-eyed pisswyrm kitty

Those two events are not IID so you can't conclude the 3% probability you mentioned

(https://en.wikipedia.org/wiki/Independent_and_identically_distributed_random_variables)

(http://www.autoadmit.com/thread.php?thread_id=4112150&forum_id=2#37070430)



Reply Favorite

Date: October 22nd, 2018 12:41 AM
Author: Mind-boggling fiercely-loyal trust fund

Yes they’re not independent but it’s a more forgiving assumption for him that they are because otherwise it just raises the probability that his model is structurally wrong.

(http://www.autoadmit.com/thread.php?thread_id=4112150&forum_id=2#37070453)



Reply Favorite

Date: October 22nd, 2018 12:57 AM
Author: Claret stimulating hospital

His "model" is little more than an average of polls with subjective weights attached and multiple toggle scenarios. It's data visualization fraud

(http://www.autoadmit.com/thread.php?thread_id=4112150&forum_id=2#37070513)



Reply Favorite

Date: October 22nd, 2018 1:00 AM
Author: Mind-boggling fiercely-loyal trust fund

Yes. As you say what his model “adds” are assumptions about which polls are more credited on the basis of those polls assumptions about who are more likely voters, which turned out to be fraud in 2016.

(http://www.autoadmit.com/thread.php?thread_id=4112150&forum_id=2#37070526)



Reply Favorite

Date: October 22nd, 2018 12:35 AM
Author: Opaque dashing new version library

that doesn't make any sense

(http://www.autoadmit.com/thread.php?thread_id=4112150&forum_id=2#37070429)



Reply Favorite

Date: October 22nd, 2018 12:41 AM
Author: Mind-boggling fiercely-loyal trust fund

Yes it does

(http://www.autoadmit.com/thread.php?thread_id=4112150&forum_id=2#37070455)



Reply Favorite

Date: October 22nd, 2018 12:42 AM
Author: Opaque dashing new version library

it makes no sense to sample selectively when his entire body of work is readily available

(http://www.autoadmit.com/thread.php?thread_id=4112150&forum_id=2#37070459)



Reply Favorite

Date: October 22nd, 2018 12:45 AM
Author: Mind-boggling fiercely-loyal trust fund

It makes more sense if trump era polling is different from non trump era polling.

(http://www.autoadmit.com/thread.php?thread_id=4112150&forum_id=2#37070471)



Reply Favorite

Date: October 22nd, 2018 12:46 AM
Author: Opaque dashing new version library

has his method changed?

(http://www.autoadmit.com/thread.php?thread_id=4112150&forum_id=2#37070473)



Reply Favorite

Date: October 22nd, 2018 12:49 AM
Author: Mind-boggling fiercely-loyal trust fund

I’m saying that the same method could be previously accurate and now inaccurate. For example, if assumptions about who are “likely voters” become wrong, as they did in 2016.

(http://www.autoadmit.com/thread.php?thread_id=4112150&forum_id=2#37070486)



Reply Favorite

Date: October 22nd, 2018 12:57 AM
Author: Opaque dashing new version library

maybe, but that's on pollsters. silver's analysis is only as good as the sampling from pollsters.

(http://www.autoadmit.com/thread.php?thread_id=4112150&forum_id=2#37070517)



Reply Favorite

Date: October 22nd, 2018 1:01 AM
Author: Mind-boggling fiercely-loyal trust fund

I don’t agree. Silver picks and chooses which polls to weight on the basis of his view of the pollsters methodology and ability to capture “likely voters.” This is how his model works. He literally rates polls with letter grades on this basis.

(http://www.autoadmit.com/thread.php?thread_id=4112150&forum_id=2#37070534)



Reply Favorite

Date: October 22nd, 2018 12:59 AM
Author: Claret stimulating hospital

He says he constantly adjusts his model based on the most recent available "data." This is how he has been able to stay relevant after the disaster that was 2016. If he gets it wrong again in 2018 his goose is cooked.

(http://www.autoadmit.com/thread.php?thread_id=4112150&forum_id=2#37070522)



Reply Favorite

Date: October 22nd, 2018 1:02 AM
Author: Mind-boggling fiercely-loyal trust fund

Cr

(http://www.autoadmit.com/thread.php?thread_id=4112150&forum_id=2#37070535)



Reply Favorite

Date: October 22nd, 2018 12:42 AM
Author: Greedy Confused Striped Hyena

RBG

(http://www.autoadmit.com/thread.php?thread_id=4112150&forum_id=2#37070458)



Reply Favorite

Date: October 22nd, 2018 12:53 AM
Author: puce doctorate wagecucks

credited answer

(http://www.autoadmit.com/thread.php?thread_id=4112150&forum_id=2#37070501)



Reply Favorite

Date: October 22nd, 2018 12:57 AM
Author: Mind-boggling fiercely-loyal trust fund

Tbf, she is sweating the 2020 general not the midterms. She’s already fucked until then with trump in the wh and gop almost surely holding senate.

(http://www.autoadmit.com/thread.php?thread_id=4112150&forum_id=2#37070516)