General Discussion

brooklynite

(94,598 posts) Wed Sep 2, 2020, 02:59 PM Sep 2020

Nate Silver (538): How much does Biden need to win by?

Link to tweet

8 replies

= new reply since forum marked as read

Highlight:

Nate Silver (538): How much does Biden need to win by? (Original Post) brooklynite Sep 2020 OP

according to his model qazplm135 Sep 2020 #1

It was right in 2016 Fiendish Thingy Sep 2020 #3

being right once or twice qazplm135 Sep 2020 #4

Enormous Bump Between Lines 3 & 4 ProfessorGAC Sep 2020 #2

Not sure this is "new" news... Wounded Bear Sep 2020 #5

Fucked up. jcgoldie Sep 2020 #6

I'm feeble minded (or at, until I wake up more from my 'power' nap)...what does Nate mean by SWBTATTReg Sep 2020 #7

He is saying that if Biden wins the popular vote by X points, then his chance of winning the whopis01 Sep 2020 #8

qazplm135

(7,447 posts)

1. according to his model

Reply to brooklynite (Original post)

Wed Sep 2, 2020, 03:02 PM

Sep 2020

which could be right, could be wrong.

Fiendish Thingy

(15,624 posts)

3. It was right in 2016

Reply to qazplm135 (Reply #1)

Wed Sep 2, 2020, 03:11 PM

Sep 2020

On the money for popular vote, within MOE for Rust Belt states.

qazplm135

(7,447 posts)

4. being right once or twice

Reply to Fiendish Thingy (Reply #3)

Wed Sep 2, 2020, 03:14 PM

Sep 2020

isn't exactly a huge sample size.

Blind faith in models because they were right once in the past isn't scientific.

ProfessorGAC

(65,076 posts)

2. Enormous Bump Between Lines 3 & 4

Reply to brooklynite (Original post)

Wed Sep 2, 2020, 03:03 PM

Sep 2020

Goes from a coin flip to 3:1 with 1% of popular vote!
The next bump (another 1%) moves it to 9:1.
Interesting.

Wounded Bear

(58,670 posts)

5. Not sure this is "new" news...

Reply to brooklynite (Original post)

Wed Sep 2, 2020, 03:20 PM

Sep 2020

It's always been true that the national popular vote needed to be won by 5% or more to "guarantee" an EC win.

jcgoldie

(11,631 posts)

6. Fucked up.

Reply to brooklynite (Original post)

Wed Sep 2, 2020, 03:27 PM

Sep 2020

I wonder what its like to live in a democracy?

SWBTATTReg

(22,143 posts)

7. I'm feeble minded (or at, until I wake up more from my 'power' nap)...what does Nate mean by

Reply to brooklynite (Original post)

Wed Sep 2, 2020, 03:28 PM

Sep 2020

a Biden EC win, if he wins by 'X points', the % points in the number of popular votes gathered by Biden over trump? In short (this is me, SWBTATTReg), if Biden wins per the latest national poll difference of +7.5 (trump gets 1000 votes, Biden gets 1075 votes, a 7.5% margin in the popular vote, thus per Nate's chart, a over 99% chance of an EC win)?

Am I following this correctly? Sorry, was a little confusing, perhaps someone knows more about Nate's mathematical workings/attempts here...

Nate: Our national polling average is settling in at Biden +7.5 and is likely to be pretty stubborn against future movement as we've added national polls from *18* different polling firms since the RNC ended. Could still use plenty of state polls; Fox News should have some out tonight.

whopis01

(3,514 posts)

8. He is saying that if Biden wins the popular vote by X points, then his chance of winning the

Reply to SWBTATTReg (Reply #7)

Wed Sep 2, 2020, 03:44 PM

Sep 2020

Presidency (electoral college vote) is Y%.

I haven't looked into his math specifically, but I would imagine it is based on something like this:

Assume that the national vote went X for Biden.
Created a weighted scale of what percentage that would mean in each state, based on current polling.
Calculate the probability distribution for the popular vote in each state based on that.
Determine how many electoral votes that would mean.

I am not sure that the current polling directly ties to the vote split in the election. Obviously there is a relationship there - but it isn't necessarily a 1:1 relationship.

Reply to this discussion