Exactly, if we're favoured by 71% to win, it's a huge upset if we lose.
2020 US Elections | Biden certified as President | Dems control Congress
- Thread starter Raoul
- Start date
- Status
Exactly, if we're favoured by 71% to win, it's a huge upset if we lose.
Not a big fan of ID politics. But for what it is worth.
Early voting brings one million Black voters to Georgia polls
Early voting brings one million Black voters to Georgia polls
That's not lead, not a poll, its a chance of winning. Trump having 1 in 3 isn't that a remote chance.
Yes, it's an upset. Like Trumps win was. That doesn't make the odds inherently wrong. They are just that, odds. We lose games we should win all the time. (and just lately, win games we shouldn't
)Ok, I think you don’t know how polls, prediction percentages, or perhaps the electoral college works. Look at how many EC votes HRC was predicted to have. Not a landslide. Look at the polls in swing states: look at the expected margin of victories there. in the margin of error. Look at how close the results were in the Midwest
The fallout from a potential Trump victory is too fascinating for my morbid curiosity to resist at this moment. It just came bubbling of out of nowhere this feeling.
In the latest 538 podcast, Silver made a very good point that due to polarization and the inherent R advantage in the Electoral College, a standard polling error on either side could be the difference between a huge upset and a landslide, and people just have a hard time processing that because in electoral politics, 5 or 6 points lead have been considered pretty safe.
This isn't going to end well. Trump and this lot are not going to take losing well. Can anyone honestly see a gentlemanly concession of defeat and a smooth transition of power?
What are the odds on that I wonder?
70-30 is not a huge upset. It just really isn't. Also, as others said, the nationwide polling was also pretty accurate. That doesn't mean anything in the US though, cause the popular vote doesn't elect anyone.
From what I've read, pollsters mostly found that the problem was that they didn't factor level of education into the weighting of the people they polled, and so a particular subset of voters was overlooked. That data gap has been fixed now, so polling should be more accurate - although they won't know until after the election what other significant gaps may exist, of course.
The bigger issue is that a lot of law enforcement in these states are now overt Trump supporters , if he decides to do it he can indeed create something similar to a mini civil war.
The polls didn't predict 29%, to be fair. Models based on polls predicted 29%. Or rather, 538 did, others predicted way lower. There's no reason not to give polls a reasonable amount of trust today, but theres's also no doubt that pollsters did mess up in 2016, mostly in the battleground states. The hope and expectation is that they've accounted for those mistakes this time around (such as weighting of education).
70-30 is not a huge upset in mathematical terms.
That said, it felt like one, since most of the mainstream media but also general commentary including on the caf in the election thread thought it was a slam dunk in the lead up to it. Amongst major "pundits" or aggregators, only 538 was still giving Trump this big a chance. Almost every neutral poll was significantly more in favour of Clinton.
Using the 538 numbers now (correctly) to say Trump always had a chance has some amount of hindsight bias. The likes of Nate Cohn amongst others were all giving Hillary a 85% or greater probability, and we saw the same confidence at Caf and the general left-leaning population.
This was from his last assessment. Mathematically, Cohn realized Trump had a real chance but subjectively, wasn't able to believe it.
That said, it felt like one, since most of the mainstream media but also general commentary including on the caf in the election thread thought it was a slam dunk in the lead up to it. Amongst major "pundits" or aggregators, only 538 was still giving Trump this big a chance. Almost every neutral poll was significantly more in favour of Clinton.
Using the 538 numbers now (correctly) to say Trump always had a chance has some amount of hindsight bias. The likes of Nate Cohn amongst others were all giving Hillary a 85% or greater probability, and we saw the same confidence at Caf and the general left-leaning population.
This was from his last assessment. Mathematically, Cohn realized Trump had a real chance but subjectively, wasn't able to believe it.
feck. Maybe I shouldn't have requested those guitar parts be shipped USPS.
USPS delivered an international "express" package after 5 weeks in August for me.
That last point is key and why I got really pissed off with Cohn for a few years. He was letting his own subjective opinions get in the way of his own data.70-30 is not a huge upset in mathematical terms.
That said, it felt like one, since most of the mainstream media but also general commentary including on the caf in the election thread thought it was a slam dunk in the lead up to it. Amongst major "pundits" or aggregators, only 538 was still giving Trump this big a chance. Almost every neutral poll was significantly more in favour of Clinton.
Using the 538 numbers now (correctly) to say Trump always had a chance has some amount of hindsight bias. The likes of Nate Cohn amongst others were all giving Hillary a 85% or greater probability, and we saw the same confidence at Caf and the general left-leaning population.
This was from his last assessment. Mathematically, Cohn realized Trump had a real chance but subjectively, wasn't able to believe it.
Bugger. Hopefully being next door will contribute to a reasonable delivery time. One has departed from LA and the other is in NYC.
You're right. (Though the reason it was at 29% was largely to account for possible polling errors) - The polls weren't that far wrong though, which is why it was at 29%. (the probability of the polls being fecked.) If he wins this time, then something has gone seriously wrong with them.
I didn't cover 2016, but the post mortem of it would definitely suggest you're correct. It's not the pollsters job to be right; it's the modelers job to analyse the polls to determine how weightworthy they are. (I know Nate Silver disagrees with that statement to a large extent, but it's my opinion. You need to vet your data.)
The systemic problem [of not weighting by education] wasn't THAT large though. Objectively, there was uncertainty in the data. Now I can't say whether I'd have been psychologically sucked in too, but it seems to have happened a lot and scarred people in the years since. They may even have overadjusted. We'll see.
feck. Maybe I shouldn't have requested those guitar parts be shipped USPS.
USPS Priority are taking approximately 4-10 weeks for me.
Had one auction on 28th September which has now arrived safelty.
Another from 4th October is "In Transit to next destination"
And another from 12th August is in the same state.
The first one worried me as it was literally an irreplacable item, but I'm getting used to it. All LA International hub by the way
Then he started speaking in ‘noun, verb, 9/11’ tongues.
He doesn’t have a clue what he’s talking about.
Please let this be the final outcome. Even a couple more red states, just whatever it takes to get him out.
Guessing before it's all over, or thereafter, there will be a few million ballots ruled void someway or another. Dirty tricks and suppression. Could be just enough to win this or that state and sneak the EC, again.
This was weird af. The court has had over a week on this and declined to rule, in which time over 100K more happened. Can't see it being successful, but it's weird.
Perhaps constant pressing from Yorke has damaged some nerve connections.
Don't put this on Yorke, he's gentle.
Who is on bottom that is not gentle?
Surely the high voter turnout in terms of early voting is bad for Trump? His voters will go out and vote for him no matter what but a high turnout this early would lean more towards favouring Biden right?
So that’s bad news?
I think the numbers in parentheses are leads, not changes
- Status