As far as I know, early results are not being announced.
As I understand it, analysts are looking at polls, and asking people who mailed ballots in “have you ever voted before?”
The logic is basically that in America we have very few cases in which the government can stop people from reporting things. So if the data is available it will be reported.
Now the fact that we still have states that register people by party is a bit odd, and I wish more states would do away with it. Some of the numbers you are seeing is a legacy of that system.
Someone upthread mentioned, and I believe, that results of the tally are not released before polls close. I mentioned that the votes here in Washington are counted, but they are not tabulated before the polls close.
I suspect that the ‘results’ are the results of the polling of people who have voted.
No, it’s either: (a) results by party based on voter registration information (FL for example) or (b) predictions based on the county or precinct in which the votes have been returned (TargetSmart data, which is of dubious value in my opinion).
Please feel free to share! So many people seem to see it as a different metaphor: which candidate do you want to marry? Which candidate do you want to be loyal to? Which candidate do you want to reward?
It’s hardly about the candidates at all, for me. The trolley metaphor is my attempt to remove personalities from the equation, to clarify that we should vote based on likely results of the vote, not on the moral righteousness of the people involved.
Imgur isn’t working for me right now. I’ll have to try uploading from the Mac when I get home. In the meantime, I was able to tweet it:
I think that’s the first time I’ve been Tweeted!
Of course, a massive turnout and/or a Biden landslide will be regarded by all good Trumpistas as Proof Positive of Massive Democrat Voter Fraud™. Because there’s no way “Sleepy / Senile / Corrupt / Pedophile / {Insert Pejorative Here} Joe” could defeat Dear Leader in a fair contest.
I only have three followers, at least one of whom is a Doper. So it might not spread very far. 
Do remember that a lot of those “people who didn’t vote in 2016” are people who could not vote, because they were under 18.
Given the enormous early voting turnout favoring Democrats, I think the best thing the DNC can invest in at this point are daylong, all-you-can-drink boat parades and Trace Adkins concerts in swing states on November 3.
IT seems to be accelerating. There have been 69.3 million early votes so far and Texas has already hit 87 percent of their 2016 total votes. I am honestly surprised that California is only at 51 percent, we have had our mail in ballots for 3 weeks now.
Actual results are not being reported. What’s reported is partisan breakdown of the vote, if it exists, or a model that predicts the partisan breakdown based on other factors, if it doesn’t. (That’s what TargetSmart does.)
In neither case is that going to exactly match the election outcome, especially for the modeled cases.
IIRC Hillary Clinton dominated early voting and Trump dominated voting on election day. I suspect that 2020 will show Biden winning the early vote handily and Trump catching up on election day. The question is whether Trump can win decisively enough on election day to overtake Biden in key states. I hope voters don’t hear headlines and assume that Biden’s already got this. I don’t think they would but elections can have crazy outcomes.
This assumes third party voters have a "second choice "party and it aligns the way you think it does. My sister is an independent who usually votes Republican. She voted for Nader.
I started a reply 12 hours ago but real life intervened. The thread has moved forward a bunch since then. Which is actually helpful; saves me a bunch of typing.
In general I agree with your math. Assuming the voter is absolutely ambivalent as between Trump and Biden for the next 4 years, an abstention or a 3rd party vote is equally pro- or con- either candidate. Your math there is unassailable.
But if, as is far more likely in the real world, the voter actually has a preference for one ticket over the other, but neither of the big two is the voter’s favorite, then, and only then, an abstention or a 3rd party vote amounts to increasing the odds your less-favored candidate wins.
This post says it very well:
Another way to look at it is to pretend the USA had single transferable voting. So your Libertarian or Whitist or Green or Communist true believer could vote e.g. [L,R] or [W, L, R] or [C, G, D] or whatever and both make their statement about their actual favorite(s) and express their input to the 2-track trolley problem.
And if they were truly absolutely ambivalent about which of the big two parties rules for the next 4 years, they could vote just [L] or [W, L] or [C] or [G, C] and leave it at that. Letting everybody else push on the two-track trolley. But if they aren’t absolutely positively ambivalent then they by expressing their 2nd, 3rd, or 12th choice as R or D they thereby increase the odds they’ll be more satisfied after the dust settles rather than less. And here’s the punch line: if doing something (including that final preference on their ballot) increases the odds, then doing nothing decreases the odds compared to the other path: doing that something.
Moving from the generic to the concrete …
Since you were that voter in the other thread who “couldn’t see” voting Trump or Biden but actually preferred Harris, I concluded (perhaps mistakenly) that you had a preference, slight though it may be, for the D ticket of Biden / Harris over the R ticket of Trump / Pence. And if my reading was accurate, then by choosing to abstain over choosing your more favored choice, you chose to increase the odds of getting your less favorable choice.
If indeed I read too much into your preference then I’m sorry to have mischaracterized you. Though in my post above that you responded to example I wasn’t speaking to you, just speaking about you as a random but concrete example out in voterland.
Hope that makes some sense.
I don’t understand what that means, when the postal ballot figures so far of Dem v Rep were shown on screen, what was it that was being reported?
What the ballot actually said is unknown at this point. Ballots aren’t yet being tallied.
It’s also based on polling, exit and otherwise.
While exit-polling is usually accurate, I always think of my brother, who was exit-polled once and deliberately gave all wrong answers.
As metioned - polling, predictions based on modeling from the known breakdown of prior votes or registration in the localities, and in states with official partisan registration of the voter which is a public record, it could even stand for the last known registration of the individual postal vote applicants. But not a report of who voted for who.