Scrutable

People are going to be really scrutinising the data from postal votes over the next few days.

First Trans Senator: Sarah McBride from Delaware.

Not that surprising, but it's a significant thing.

Some good news.

Strong majority for legalising weed in New Jersey, at 67%. NJ the twelfth state to grow up.
https://www.nytimes.com/interactive/202 ... juana.html

There's a lot still to come. There will be surprises for Biden, and Ohio is looking positive on that front.

Damn. Lindsey Graham wins South Carolina.

Senate:

Lindsey Graham wins easily in South Carolina.

Literally a few dozen votes between Cunningham and Tillis in North Carolina.

Iowa: polls just closed. Now a must win

Maine: also a must win.

Nevada looking good for Biden, though.

Let's hope what happens in Las Vegas doesn't stay in Las Vegas tonight.

Conspiracy theory: 538 is a grift.

Almost all the people who are going to wonk out over polling stats like that will be Democrat supporters. A site with tiny incremental adjustments to a satisfying narrative would massively engage and addict that audience, providing a great opportunity to earn money/promote narratives.

Remember that nobody external has verified their models.

Bird on a Fire wrote: Wed Nov 04, 2020 3:12 am Conspiracy theory: 538 is a grift.

Almost all the people who are going to wonk out over polling stats like that will be Democrat supporters. A site with tiny incremental adjustments to a satisfying narrative would massively engage and addict that audience, providing a great opportunity to earn money/promote narratives.

Remember that nobody external has verified their models.

It wouldn't have to an intentional grift for the conflict of interest to bias the results.

Anyone know the Arizona timings?

Reminder: Arizona (plus holding Nevada) means Biden can win without waiting for Pennsylvania court cases (assuming he wins WI and MI).

People are saying Arizona looks good. Was +2.6 to Biden in the 538 average so a bit of polling miss to play with. Nevada was +6.1 per 538.

secret squirrel wrote: Wed Nov 04, 2020 3:16 am

Bird on a Fire wrote: Wed Nov 04, 2020 3:12 am Conspiracy theory: 538 is a grift.

Almost all the people who are going to wonk out over polling stats like that will be Democrat supporters. A site with tiny incremental adjustments to a satisfying narrative would massively engage and addict that audience, providing a great opportunity to earn money/promote narratives.

Remember that nobody external has verified their models.

It wouldn't have to an intentional grift for the conflict of interest to bias the results.

538 is only as good as the polls feeding into it. It just averages other peoples polls. If everyone f.cks up polling of Hispanics what is 538 supposed to do?

Silver will be boasting about his 10% chance when others had 5%, mind, and he'll be thoroughly infuriating about it.

So, Biden leading 55/45 in early votes in Phoenix, AZ (urban, 40% Hispanic vs 30% average for AZ). Early votes add up to the total number of votes cast in 2016.

Seeing as early voting has been such a partisan issue, I was worried that we'd see basically all* Ds voting early and basically all R voting for covid. That would make it really hard to follow the votes as they come in without prior knowledge of how many of each kind of vote had been cast.

Looks like that's not going to be an issue, if we have 45% R support in early votes in Phoenix. Still not clear what the final totals for the county will be, but at least we probably don't have to worry about early counts being massively biased.

*say, two thirds

Looks like Trump will win Ohio. The big cities still have quite a few votes to count, but it doesn’t look like it will make it up. It will be a big swing to Biden regardless.

They're already reporting 70% of the vote in Arizona but f.ck knows. Biden doing well in Maricopa County.

lpm wrote: Wed Nov 04, 2020 3:19 am

secret squirrel wrote: Wed Nov 04, 2020 3:16 am

Bird on a Fire wrote: Wed Nov 04, 2020 3:12 am Conspiracy theory: 538 is a grift.

Almost all the people who are going to wonk out over polling stats like that will be Democrat supporters. A site with tiny incremental adjustments to a satisfying narrative would massively engage and addict that audience, providing a great opportunity to earn money/promote narratives.

Remember that nobody external has verified their models.

It wouldn't have to an intentional grift for the conflict of interest to bias the results.

538 is only as good as the polls feeding into it. It just averages other peoples polls. If everyone f.cks up polling of Hispanics what is 538 supposed to do?

Well, it averages, weights and correlates them in ways that aren't transparent, and attempts by independent statisticians to reverse-engineer them have produced results they can't understand.

I'd go with cockup rather than conspiracy, though. I'm just having fun.

The whole polling industry is going to be under scrutiny. Big misses in 2016, 2018 Senate (but not House) and 2020.

Anyone seen anything on NE-2 and ME-2?

Will be deciders in "win Arizona, don't win Pennsylvania" scenarios.

Might go and see if I can sleep for a bit.

This is the most disappointed I've been about early election results since December 2019.

NC looks like it’s gone to Trump

Georgia's still not entirely lost. A lot of votes to count. It's been swinging back to Biden a bit (from losing by 3 to losing by 2).

Switched from CNN to BBC just now when they decided to do a history lesson on Coca-Cola. Had to turn up the volume. Americans really are f.cking loud.

Georgia's also missing 80,000 very blue votes due to a polling station error, apparently.

Iowa early results are coming in a bit better than 2016.

Clinton lost by 9.4 so it's like Ohio - won't win, but good performance should translate to WI, MI, PA.

Iowa Senate race looks lost. Heading for 51-49 to Republicans, assuming AZ and Maine join Colorado (Alabama is already lost). Maybe Montana could rescue it?

lpm wrote: Wed Nov 04, 2020 3:24 am The whole polling industry is going to be under scrutiny. Big misses in 2016, 2018 Senate (but not House) and 2020.

Yes for sure. It'll be interesting to see what the issues are.

I'm suspecting an over-reliance on race and underweighting of the importance of class, and lack of consideration of how the two interact. (I expect most pollsters are probably middle class white guys)

They also need to get around the fact that nobody sane would ever give up their time to answer survey questions these days without some seriously strong motivation.

Also the fact that a lot of polling seems to be partisan doesn't help.

Of course, a lot of this could all be actual fraud.