Most likely to win: Bollier, Gross, Harrison, James

[Meatball Ron]

Harrison > Gross > Bollier > James

[Meatball Ron]

Thought this was a bit of a fun thread to bump after going through my archives -- one of these candidates lost by 2 points, and all the others lost by double-digits (11 to 13 points). Interesting to consider why James wasn't the overwhelming choice.

Because the most active Atlas members are generally dem hacks, thus it's not suprising that James' chances were massively underestimated while the Bollier/Gross/Harrison's ones were massively overexagerated.

I mean, the real answer is the massive uniform polling errors. But also - all four lost - so you can't really prove the counterfactual that James was "more likely to win" than the other three. The fact that he lost by less doesn't really matter at all. He lost.

[Meatball Ron]

Gotta appreciate not only that most people on here not only thought AK, KS, and SC races were close, but also that multiple people put James IN LAST as far as competitiveness.

Michigan looks like it's all but out of reach for Republicans at the presidential level, and I highly doubt James is gonna outrun Trump enough to win.

I'm gonna be bold and say it: I'm starting to feel like South Carolina could be North Carolina 2008. Based on recent polling, would not be surprised if Biden just barely wins it and Harrison outruns him by a few points.

Note SC is now closer for president in the FiveThirtyEight averages than Michigan...

As for the Kansas/Alaska races, I really have no idea about them. Neither narrow Gross/Bollier wins or comfortable Republican wins would shock me.

This, folks, is why you should never trust Nate Silver. I tried to warn people in the runup to the election, but instead I got a lot of scorn for bringing up his record. "It's not going to be 2016" everybody said in unison. They were actually correct about that, as far as polling and "expert" prognostication goes it was far worse.

The issue here is that you were right, but for the wrong reasons (or, more precisely, no reason).

People saying the polls would be different in '20 had two major data-driven regions for believing this: 1) 2016 polls didn't weight for education, while 2020 polls did 2) There were serious signs of trouble for Dems in district-level polls in '16 even while the state-level polls look good; this was not a dynamic in 2020 and the district polls still looked good for Dems

People like you, on the other hand, just repeated that the polls would be wrong again, but couldn't actually provide any reason for that. It's like getting the answer right in math class but not showing your work. Just because you happened to be right doesn't mean you should get any credit.