Quote (Thor123422 @ 24 Oct 2020 07:55)
If the polls aren't MASSIVELY incorrect even relative to 2016, then Trump is sunk.
Trump needs the voting dynamics in the country to be so radically shifted relative to even 2 years ago that all models are worthless.
Now.... the thing that scares me.... is this could very well be the case.
I've already made this same argument to thundercock yesterday: what you say is only true if we go by the polling averages. But since the spread of the polls is unusually high this year, this approach or way of thinking about the race is very risky.
Normally, polls vary a bit based on minor methodological differences between pollsters and on sampling error, and these effects can effectively be smoothed away by aggregating the polls. But I'm very sceptical of the soundness of this approach in a world where we're, on the same day, getting quality polls showing Biden up between 10 points and 4 points nationally, or up between 3 and 13 points in Pennsylvania. This kind of split between the polls has been a consistent feature of this race, and the gap is just too big and persistent to be explained by outlier samples or minor methodological differences. This kind of split is only possible when pollsters use a significantly different methodology, or see vastly different electorates.
Either the polls showing a landslide, or the polls showing a mid-single digit race have to be way off. And if it's the latter type which is showing the true picture, then Trump's behind but still very competitive in the decisive battleground states.
It basically comes down to this: do we think that a Quinnipiac poll showing Biden up 11 in Floirda and 13 in Pennsylvania is garbage? If the answer is yes, then including it in the polling average will inflate this average, leading to all assessments based on this average to be biased in Biden's favor too. This of course cuts both ways, but I think we've seen far more polls unrealistically inflated toward Biden this cycle than toward Trump.
My personal assumption is that Biden will win the popular vote by 6-7%, and the tipping point state in the EC by around 3%. A 10% Biden win wouldnt totally surprise me though, and neither woudl a narrow Trump win.
This post was edited by Black XistenZ on Oct 24 2020 12:33am