Scott Fines
scottfines.bsky.social
Scott Fines
@scottfines.bsky.social
Math and Engineering is my jam.
AFAICT, this election looks very much like 2020, which…is good for Dems? I dunno, seems like we should just be patient and wait for final tallies before we freak out.
November 6, 2024 at 3:04 AM
I know no actual numbers, but this would be quite easy if turnout is down in those areas(or flat). Smaller denominator and all that
Early numbers:

Trump is winning a bigger share of the vote than he did four years ago in places that have almost finished counting. In the 68 counties in five states where nearly all ballots have been counted, Trump has won 66% of the vote this year compared to 64% in 2020.
November 6, 2024 at 1:16 AM
Reposted by Scott Fines
ann selzer, you fucking madwoman
November 5, 2024 at 11:48 PM
This is totally random, but two weeks ago I watched a 12u soccer tournament be decided by penalties and it was the best possible ending and all sports should do it
Just a reminder that if the election is tied they go straight to penalties instead of extra time
November 5, 2024 at 11:37 PM
Reposted by Scott Fines
NEW: I ran 80,000 simulations of the election. My wife left me for a guy who doesn’t do that.
November 5, 2024 at 3:56 PM
Reposted by Scott Fines
No one expected overruling Chevron would drive turnout like this.
November 5, 2024 at 7:21 PM
So one of the things that it’s fun to know from your wife the elections expert is that a bunch of states expanded their early voting laws in 2020 for COVID and also that having longer voting windows increases turnout. Hence, 2020 and (starting to look like) 2024z
November 5, 2024 at 7:21 PM
I have a sinking suspicion that one of the reasons this doesn't happen (aside from the integrity issues) is that most of the politics news-desks around the country don't understand polling well enough to realize how much Cohn has been bullshitting this cycle.
November 5, 2024 at 3:50 PM
Yesterday everyone on this site was all "WE'RE GONNA WIN! GRIND THEIR BONES INTO ASH". This morning it's all "oh man, we are so screwed!"

And this is a perfect example of sampling error--all the hopeful people have logged off to protect their sanity, and all that is left are the doomers.
November 5, 2024 at 3:44 PM
I briefly followed some people on early voting results, and then witnessed them freaking out about Guam, and realized what a horrible mistake I had made.
November 5, 2024 at 3:41 PM
My hot take, although I’m not super confident in it, is that Harris wins Florida tomorrow in an absolute squeaker. Recount territory close, but it’ll stand in the end.
lmao wait they're bragging about being +4 in Florida?????

holy shit they may actually be fucked tommorow

(we will have to see but...)
November 5, 2024 at 2:33 AM
Reposted by Scott Fines
Ah
November 5, 2024 at 1:25 AM
The election is nigh.

What I should be doing: Working on my power-plant query engine.

What I am actually doing: laughing at insane polls and outlining the math on how I would write a polling prediction engine.

Goddamnit
November 4, 2024 at 11:01 PM
While you're at it, expand the House of Representatives so that we don't live in a world where 500k people get one representative.
DC should be admitted as 27 different gerrymandered states and the resulting overwhelming Democratic majority should be used to amend the Constitution to neuter the Senate
depends on how many new states we add. Here in San Diego County we're standing by and ready for admission as the 52nd/53rd state
November 4, 2024 at 10:21 PM
Reposted by Scott Fines
A vision of the 2024 US presidential election, from 1925
November 4, 2024 at 9:22 PM
lol. Lmao
Dartmouth just released a poll that has Kamala winning NH by 28 points. What.
November 4, 2024 at 9:54 PM
I do think that on the whole we(BlueSky) are putting a little too much on the DMR poll. Selzer has an excellent record, and her poll is just self-evidently of higher quality than most of the trash out there. But it is still a single poll, and outliers do happen. She's very good, but not infallible.
November 4, 2024 at 8:11 PM
Reposted by Scott Fines
2016: selzer's method works, everyone else models away the signal and gets surprised

2020: selzer's method works, everyone else models away the signal and gets surprised

2024: selzer's method works, everyone else models away the signal and gets surprised (TBD)

...why didn't anyone try her method?
November 4, 2024 at 7:53 PM
Lots of people adding context, making it more complicated, but the basic story is this:

Some dude did some shit online.

The fact that we're all being like "OOOH right wing conspiracy!" is just evidence that the RW has lost its goddamn mind, and not a sign that we need to dig deeper on the details
Can someone explain the squirrel thing to me
I’m sorry for bringing over a screenshot but this is the bubble they’re living in right now

this is their closing message

the pulse of right wing America, everyone
November 4, 2024 at 5:19 PM
I don’t want to dog, but my wife runs elections here in town, and fights tooth and nail to get lines shorter. Lines have been long this year. It’s not always nefarious, sometimes it’s because there aren’t buildings big enough to handle the demand.
November 4, 2024 at 2:26 PM
Reposted by Scott Fines
It’s the abortion bans, stupid
November 3, 2024 at 7:04 PM
Pollsters seem to have the implicit theory that voters change their opinions frequently with little or no justification.

Gerrymandering, by contrast, operates on the exact opposite theory: voters almost never change their opinions.

Only one of these theories is true. I’ll let you guess which.
November 3, 2024 at 3:31 PM