In the latest version I found of Nate Silver’s model (not 538), he has Ohio coming in at 52.4% for Trump and 43.6% for Harris, an 8.8% spread. I did not dig deeper to find the dates or particular polls from Ohio he’s basing that on.
However, based on these numbers, he is likely modelling that Trump wins Ohio in 90%+ of outcomes to Harris’s <5% of outcomes.
This is the same way he spoke to his model in previous elections. It wasn’t that Hillary was expected to win 80-90% of the popular vote or electoral college just weeks before the 2016 election, it was that his model had her winning that percentage of the outcomes when he ran the model.
The numbers I gave are the model outputs for the state as of yesterday off his subscriber based model talk page.
So no.
Of course these are the likelihood of a win and not polling differences. That’s why I said model output, not a poll aggregate.
An 8 point spread in a state for polling averages is incredibly large. For reference Ohio is as deeply spread red in polling averages as Nee Jersey is blue. You think New Jersey votes red this year in any reasonable reality? No.
For an even more crazy but accurate comparison: Alaska has the same mid point statistical odds of going Texas Ohio, but its error bars are more than double Ohio. Meaning? There is an incredibly slim but massively more possible chance Alaska goes blue than Ohio.
In the latest version I found of Nate Silver’s model (not 538), he has Ohio coming in at 52.4% for Trump and 43.6% for Harris, an 8.8% spread. I did not dig deeper to find the dates or particular polls from Ohio he’s basing that on.
However, based on these numbers, he is likely modelling that Trump wins Ohio in 90%+ of outcomes to Harris’s <5% of outcomes.
This is the same way he spoke to his model in previous elections. It wasn’t that Hillary was expected to win 80-90% of the popular vote or electoral college just weeks before the 2016 election, it was that his model had her winning that percentage of the outcomes when he ran the model.
The numbers I gave are the model outputs for the state as of yesterday off his subscriber based model talk page.
So no.
Of course these are the likelihood of a win and not polling differences. That’s why I said model output, not a poll aggregate.
An 8 point spread in a state for polling averages is incredibly large. For reference Ohio is as deeply spread red in polling averages as Nee Jersey is blue. You think New Jersey votes red this year in any reasonable reality? No.
For an even more crazy but accurate comparison: Alaska has the same mid point statistical odds of going Texas Ohio, but its error bars are more than double Ohio. Meaning? There is an incredibly slim but massively more possible chance Alaska goes blue than Ohio.