538 gives Biden a better chance in WI and MI than NH and NV

[Barack Oganja]

Don't feel like sending screenshots, but right now it gives Biden 94% in WI, 95% in MI, 90% in NV, and 89% in NH. I know those are small differences, and I'm probably splitting hairs here, but what accounts for that difference? I'd say maybe the number of polls, but it's not as if they're going off of one or two polls in NH and NV. They also rely on demographics and past voting history, which should boost those numbers. Are they two low in NH and NV, too high in WI and MI, or have it right in your opinion?

https://projects.fivethirtyeight.com/2020-election-forecast/wisconsin/
https://projects.fivethirtyeight.com/2020-election-forecast/michigan/
https://projects.fivethirtyeight.com/2020-election-forecast/nevada/
https://projects.fivethirtyeight.com/2020-election-forecast/new-hampshire/

[Alben Barkley]

NH is 100% Biden. Every poll has him up double digits, makes nosense.

[forza nocta]

There is no close 2016 state I feel more confident in that Biden will win than New Hampshire.

[Senator Incitatus]

Fewer polls, which favor Biden by more than the fundamentals?

[DaleCooper]

NH is 100% Biden. Every poll has him up double digits, makes nosense.

I think there have actually been a couple legitimate (though still wonky) polls that showed Biden only up by 3 in New Hampshire somewhat recently, whereas there haven't been any polls other than Trafalgar showing Trump anywhere close in Wisconsin.

[charcuterie]

While it's true there are several polls out of NV and NH, the fact is that the 538 model has a very high degree of uncertainty and the reason Michigan and Wisconsin are so high for Biden is that there has been a MOUNTAIN of polls that have Biden up by mid single digits to low double digits. NH and NV polling has been consistent, and the volume isn't bad, but it greatly pales in comparison to what's come out of the midwest.

[Gustaf]

The fundamentals would say that all of those states were basically tied last time (ok NV a bit more D-friendly, but not much). WI and MI have had way more polls than the others. And as I recall NH is one of the most elastic states by their elasticity scores.

So the model is probably more uncertain about them. Note that in terms of margin it predicts a bigger win for Biden in NH than in either MI or WI.

[Barack Oganja]

The fundamentals would say that all of those states were basically tied last time (ok NV a bit more D-friendly, but not much). WI and MI have had way more polls than the others. And as I recall NH is one of the most elastic states by their elasticity scores.

So the model is probably more uncertain about them. Note that in terms of margin it predicts a bigger win for Biden in NH than in either MI or WI.

That's a good point, didn't notice that

[MT Treasurer]

Yes, they did that in 2016 as well. They were wrong back then and they will be wrong again.

[ProgressiveModerate]

I mean, there’s not an order states are bound to flip in like Atlas likes to think but that Biden + 17 is really kicking in