As I said yesterday, watch for the early closing states of Florida and N. Carolina. If these states are called for Biden, it’s game over.
538 has a page where you can force their model to call certain states as winners for a candidate, and then see what the model says about likely victory. (14 states) If Florida and N. Carolina go Biden, then the chances of an overall Biden victory are 99% +
This. Trump is the symptom. Electing Biden would be great–akin to breaking a dangerously high fever with aspirin. But it won’t mean we’re cured.
But using that model all it takes is for PA to go to Trump and everything else falls apart for Biden. And PA is going to be too close for my comfort based on current polls.
However this could also cause unnecessary stress in the opposite direction. Worth remembering that Biden could lose both Florida and North Carolina and still win the election with the Hillary states + the close rust belt states she lost.
Biden has multiple outs, Trump has few.
It’s very unlikely we’ll know PA on election eve, or even the next morning. But we’ll probably know FL, NC, and maybe even GA. If 1 of those goes for Biden, he’s probably won, regardless of PA. If 2 or 3, then he’s definitely won, and it won’t even be close.
The Nytimes recently updated their poll page. Biden is currently +6 in PA. If you adjust by the 2016 error rate he’s +2. I’d prefer +15 but I feel cautiously good about +2.
Whoa, I just noticed they updated the ECV count. If you adjust all polls by the 2016 error rate Biden wins 335 votes. That’s up from 280 yesterday. That’s a good trend.
NC will be close and SCOTUS has confirmed it can receive election-day postmarked ballots up to Nov 12, so I wouldn’t bet on knowing that one early.
True. However, if Biden gets sworn in I’ll quit worrying about that specific election.
For those putting faith in 538’s prediction.
Right! They were one of the few to give Trump a significant chance of winning (almost 1 in 3). They’ve consistently been the best data-based election predictors around.
Yes…that was pretty spot on. Trump had almost 30% chance to win and he did.
A statement of odds isn’t technically a prediction, is it?
As others have indicated, 538 gave Hillary Clinton about a 70% chance of winning, and gave Trump a nearly 30% chance to win. That was certainly not a slam-dunk prediction that Clinton would win.
Do you know what else has just about that same chance of occurring? Playing Russian roulette with two rounds in the chambers of a six-shooter revolver. Would you play those odds? Would you think you had a slam-dunk chance of winning, or would you be nervous? (I would be nervous, and I was nervous in 2016.)
For comparison, Biden’s chance of winning the 2020 election is closer to playing Russian roulette with only one round in the chamber. I’m still nervous.
I don’t know their track record. I trust 538 because their track record back to '08 is very good.
Do you actually think anyone here is unaware of that?
As has been said, Biden’s chances are much better than that, polling has improved, and even assuming the polls are as wrong as they were in 2016 still gives Biden the win.
Even so, I’m not going to relax until after the election, just because of that experience.
I 100% believe he is very likely to win a legit election. My concern is:
Trump does better than expected + Voter suppression and shenanigans = Trump victory.
I don’t disagree with that.
From what I can tell the Economist model uses a very similar technique. It just appears they incorporate a bit less uncertainty between now and election day and they seem to exclude some partisan polls that 538 includes.
You can actually just look at the PA forecast. Economist has it at 93%. 538 has it at 86%. That seems to be largely due to the Economist not using that Insider Advantage poll that was T+3, which makes Biden’s margin 6.4 rather than 5.2.