Page 1
Page 1
Started By
Message

tOfficial Nate Silver fail thread

Posted on 11/9/16 at 11:07 am
Posted by 225bred
COYS
Member since Jun 2011
20386 posts
Posted on 11/9/16 at 11:07 am
Can we get some clarification on just how much devolving Nate's surname has undergone during this process?
Nate tin foil? Plastic?
Posted by RobbBobb
Matt Flynn, BCS MVP
Member since Feb 2007
27938 posts
Posted on 11/9/16 at 11:12 am to
There will be a few luddites that will be on here shortly to tell you that Nate Velcro was within 30% of getting it right, so he was essentially the closest to reality
Posted by buckeye_vol
Member since Jul 2014
35239 posts
Posted on 11/9/16 at 11:17 am to
quote:

There will be a few luddites that will be on here shortly to tell you that Nate Velcro was within 30% of getting it right, so he was essentially the closest to reality
So given the polling information, Trump losing the popular vote, and the margins in swing states, what do you think the probability of this outcome was beforehand?
Posted by RobbBobb
Matt Flynn, BCS MVP
Member since Feb 2007
27938 posts
Posted on 11/9/16 at 5:45 pm to
quote:

what do you think the probability of this outcome was beforehand?

I have been telling you since Day 1 that I had the same info as Nate, and was banking on a Trump win. Trump had the same numbers as Nate, and knew that Penn, Mich, and Wisc was winnable and went there. Clintonistas used Nates 'view' and never went one time to Wisc. Not once

Yet Nate had Trump at 15%, around a week out.

Sorry his algorithm is shite, and biased. Based on voting patterns that he pines for, not what is fairly well known thru research. Basically, hes lazy
Posted by buckeye_vol
Member since Jul 2014
35239 posts
Posted on 11/9/16 at 5:52 pm to
quote:

Trump had the same numbers as Nate, and knew that Penn, Mich, and Wisc was winnable and went there.
And Nate even commended Trump's approach to going to these states and questioned why Hillary was in Arizona. He knew Trump had a good chance, but that doesn't mean he was the favorite to win.
quote:

Yet Nate had Trump at 15%, around a week out.
The model lags the data and the data lag the changes in views. It increased rapidly last week for a reason following the changing in polling data.
quote:

Sorry his algorithm is shite, and biased.
It's flawed (and he has admitted it), but it gave the Trump the best odds of any model despite the data showing a Clinton lead.

What else is he supposed to do? Just ignore the inputs of his model?
Posted by bamarep
Member since Nov 2013
51807 posts
Posted on 11/9/16 at 5:54 pm to
The problem was Nate's method was using polls that were absolute bullshite.

Some of us tried to tell y'all that but you wouldn't listen.
Posted by cwill
Member since Jan 2005
54753 posts
Posted on 11/9/16 at 5:55 pm to
What we do know for sure is that the vast majority of the posters attacking Silver have no understanding of statistics, modeling or forecasting.
Posted by ihometiger
Member since Dec 2013
12475 posts
Posted on 11/9/16 at 5:58 pm to
Posted by buckeye_vol
Member since Jul 2014
35239 posts
Posted on 11/9/16 at 6:01 pm to
quote:

The problem was Nate's method was using polls that were absolute bull shite.
But that's the whole purpose of the model. Now it over to grading the accuracy of the pollsters. So the pollsters who were spot on (Selzer) will be rightfully lauded.
quote:

Some of us tried to tell y'all that but you wouldn't listen.
They were flawed, but they weren't "made up" like you argued.

In fact, with Clinton likely to win the popular vote by over a percent (once the last votes are added in CA and other blue states), the National polls will only be off by a couple points. Interestingly LATIMES will probably have one of the biggest misses.
This post was edited on 11/9/16 at 6:04 pm
Posted by CptBengal
BR Baby
Member since Dec 2007
71661 posts
Posted on 11/9/16 at 6:04 pm to
quote:

o given the polling information


You said that wasn't rigged.

Lolzy.

quote:

buckeye_vol


Beg waiting for you to show up
Posted by bamafan1001
Member since Jun 2011
15783 posts
Posted on 11/9/16 at 6:04 pm to
quote:

What we do know for sure is that the vast majority of the posters attacking Silver have no understanding of statistics, modeling or forecasting.


We know Silver was dead wrong. What good are his statistics and projections if they cant actually predict what will happen?
Posted by buckeye_vol
Member since Jul 2014
35239 posts
Posted on 11/9/16 at 6:11 pm to
quote:

You said that wasn't rigged.

Lolzy.
It wasn't. It was wrong, at the state level, but national polls were fairly close.
quote:

Beg waiting for you to show up
I posted a crow eating thread last night. Congrats MAGAers: Official Crow Eating Thread
Posted by gaetti15
AK
Member since Apr 2013
13367 posts
Posted on 11/9/16 at 6:15 pm to
Andrew Gelman put out an article on his website that goes into good detail discussing what happened.
Posted by TheRodFather
Member since Sep 2014
619 posts
Posted on 11/9/16 at 6:17 pm to
quote:

The problem was Nate's method was using polls that were absolute bullshite. Some of us tried to tell y'all that but you wouldn't listen.


In their defense, a lot of people got egg on their face last cycle when they tried to "unskew" the polls but that time they turned out to be pretty accurate. That lead to the Karl Rove melt and Nate Bronze's reputation.
Posted by RobbBobb
Matt Flynn, BCS MVP
Member since Feb 2007
27938 posts
Posted on 11/9/16 at 6:20 pm to
quote:

They were flawed, but they weren't "made up" like you argued.

How do you type that with a straight face?

A poll is supposed to be a 'snapshot' of what a particular group is feeling at that moment in time. So if its 'flawed' then the results are MADE UP! Because it purposefully does not do, what it is supposed to do: show what the electorate was thinking

And when I say made up, I mean to oversample minorities, white women, college grads is seeking to get a specific result, and not realistically looking for the actual snapshot
Posted by buckeye_vol
Member since Jul 2014
35239 posts
Posted on 11/9/16 at 6:22 pm to
quote:

So if its 'flawed' then the results are MADE UP!
What?
Posted by cwill
Member since Jan 2005
54753 posts
Posted on 11/9/16 at 9:37 pm to
quote:

We know Silver was dead wrong. What good are his statistics and projections if they cant actually predict what will happen?


Exhibit A.
Posted by Vacherie Saint
Member since Aug 2015
39519 posts
Posted on 11/9/16 at 9:41 pm to
538 gives nate silver and F
Posted by D011ahbi11
Member since Jun 2007
13623 posts
Posted on 11/10/16 at 12:53 am to
Nate's biggest fail last night was his hair. Holy shite did he look bad
Posted by LaBornNRaised
Loomis blows
Member since Feb 2011
11004 posts
Posted on 11/10/16 at 12:56 am to
Link?
first pageprev pagePage 1 of 1Next pagelast page
refresh

Back to top
logoFollow TigerDroppings for LSU Football News
Follow us on Twitter, Facebook and Instagram to get the latest updates on LSU Football and Recruiting.

FacebookTwitterInstagram