I’m pretty pessimistic, I was talking to the owner of the little cafe next to my office and she said there were a ton of Trump supporters with their trucks parked in front of the polling place, she felt uncomfortable and turned around. The polls don’t take into account the type of garden-variety voter suppression that is basically endemic in places like Florida and Georgia and Texas. I say this as one of the people who was overly optimistic when Florida passed the rights restoration amendment two years ago. I went from thinking it would
Make all the difference to now thinking it won’t matter. Until I see us turn the state blue, I’m no longer going to be optimistic about Florida.
1 Like
Fuck no. Not only that, but I have already committed here in the Beer thread that I won’t be touching any Cantillon on Election Night this year.
3 Likes
Re: All the early voting Doom and Gloom in this thread:
https://mobile.twitter.com/NateSilver538/status/1321447653230653443
https://mobile.twitter.com/NateSilver538/status/1321447903378984960
https://mobile.twitter.com/NateSilver538/status/1321448266333081601
https://mobile.twitter.com/NateSilver538/status/1321448815325491203
https://mobile.twitter.com/NateSilver538/status/1321449548456333313
9 Likes
Honestly in the middle of a pandemic with disastrous job numbers this seems more likely than the counter argument.
2 Likes
I’ve had the same thought, but I do wonder about applying a 90/10 in poker to a 90/10 here. I mean, we talk about “running the election 40,000 times,” but I dunno. A poker hand is one event that we can perfectly model the numbers for. The election is millions of events (individual voters voting) and I’m just not sure the analogy holds up that well. The poker analogy would imply that a given person (or enough people) is going to vote for candidate A 90% of the time and candidate B 10% of the time, right? I don’t think that’s what happens? Hopefully this isn’t a donk math take.
2 Likes
It was explained by someone else itt that the uncertainty implied by election probabilities relates to polling errors, not voters switching.
So it’s not like we run this election 100 times and a couple of times it’s a Trump massacre.
2 Likes
That makes sense.
I guess I feel like the polls say what they say, and everything outside the polls (economy in the tank, society shut down, he’s always been underwater in approval, now an incumbent rather than a challenger running against deeply unpopular HRC) suggests he will lose. To me, in order for him to win, we’d need not only the polls to be wrong in his favor, but also all that stuff to not matter as much as some nebulous factor breaking in his favor (e.g. there being millions of “hidden” Trump voters or him bringing lots of new people into the coalition despite never doing anything to expand it). It just seems very unlikely.
Yeah. There was me trying to liken it to sports betting duh.
ETA plus widespread cheating ofc
And not polling errors as in counting statistics uncertainty. Polling errors in constructing the accurate sample of the electorate, which isn’t really random, just unknown.
1 Like
The gap in my county between Dem and Rep early in person voting went from 2k to 4K overnight.
Let’s be real for a second… in real life one of these candidates is at 100% and the other is at 0%. We just don’t actually know which one. It’s really binary, either the pollsters are totally wrong about everything and Trump is going to squeak this out through cheating and low info white turnout, or he’s more fucked than anyone since Mondale.
The reason we don’t know is what you outlined. Math is only ever as good as its inputs. It’s why when my actuary brother starts down the wrong track on something I can almost always track where he went wrong to a bad input. You fix the input and he leaves the conclusion at roughly the same speed a calculator would.
1 Like
https://mobile.twitter.com/NateSilver538/status/1321447653230653443
The independent thing is huge. I was also thinking about how the Republican registration advantage in places like Florida would affect this. When someone gets an independent to register as Republican (or Democrat) it’s not like they’re gaining a vote. They’re identifying someone that was likely to choose their candidate and increasing their likelihood of voting a slight amount. Clearly valuable.
BUT, in cases like the pandemic where Democrats may have been less willing/able to get the same level of sign ups as Republicans it means the remaining Independents are likely to be more Democratic leaning than we’d otherwise expect.
Or maybe this is all hogwash. 
2 Likes
Trafalger is completely bogus. Just made up. Rasmussen is hugely biased.
1 Like
10 Likes
3 Likes
Not according to one of the foremost polling experts in the world who is doing an AMA on 2p2
15 Likes
Everyone needs to go over there and post this gif
4 Likes