Blue Hopium ONLY info

  • Thread starter Thread starter nycfan
  • Start date Start date
  • Replies: 306
  • Views: 9K
  • Politics 
My contribution to this thread is to say that I really don't think it's needed. The stuff on the regular thread isn't bad. It's national polls, which don't matter, and a smattering of state polls that are mixed in effect. Other than AZ, which has looked tough all year long, nothing bad is happening in any of the other states.
 
Geisterfahrer said:
What does any of this mean? How are they gaming the system? Didn’t Nate just say his gut was Trump?
Click to expand...
He answers later - subtle model shifts OR giving personal takes that slightly differ to give their personal selves an "out" if she wins....

So they can do it all over again if we have more elections.
 
Here's another consideration. We talk about correlations a lot, because they are important. But in a tied race, correlations don't matter much.

Think about it. Let's say NC and GA are both decided by half a point. That's less than the correlation coefficient between the two states. That is, you can predict NC within a point or two if you know the GA vote. But a point or two isn't enough.

So if we are looking at razor thin margins, the states become more like independent coin toss trials. And if I'm right that Kamala is in great shape in MI, good shape in PA and bad shape in AZ, then the Pubs have to win 2 of 3 coin flips to win.
 
aGDevil2k said:
He answers later - subtle model shifts OR giving personal takes that slightly differ to give their personal selves an "out" if she wins....

So they can do it all over again if we have more elections.
Click to expand...
OK, to be clear: the article does not contain any statements from Nate Silver at all.

It's reporting only that the Silver model has shifted a bit in the last week. Models should shift as new data comes in.

What Nate said in the Times article remains valid, which was basically "trust the model over your gut." That doesn't mean trust the model over everything, but if you're second guessing the model because you don't like the conclusions, that's really not a great way to go about things.

Nate gets way too much flack. It's not as if he's shy about what he's doing. He makes models to give statistical forecasts of events. If you don't want a statistical forecast, then don't look at him -- but why lash out at him. It's no different than what stock traders do every day. If they think a stock has a 60% of outperforming the market, they will buy it (assuming they have capital free). In many cases, those trades will suck but they are outweighed by more good trades.

I also have never seen Nate Silver claim to have gotten any election "right." What he has said was that he wasn't wrong, and that's true. When he predicted Biden to have an 88% chance to win, and it was way closer than the predicted margins . . . well, that's why you have a model. In 2016, it was a 72% HRC probability. Well, the difference between an 88% chance and a 72% chance is that if things go horribly wrong, the 88% gives you more margin for error. And that's exactly how it turned out. HRC barely lost, Biden barely won.

That doesn't mean Nate was "right" but he wasn't wrong in either election.
 
FLORIDA: Poll indicates that Kamala Harris leads amongst voters who have already voted. (This despite 500,000 more Republicans having voted so far.)

Florida St. Pete poll:

– Trump 50-45

– Scott 49-46

Donald Trump leads Kamala Harris in Florida, but not among those whose ballots are already cast

NOTE: If Harris really is ahead by 1% amongst those who have already voted, despite Republicans having an 11-percent Ballot Edge in the Early Vote, this is earthquake-level news for Florida

How could this be? There are two key factors:

– Women voters outnumber men 54–46 according to TargetSmart

– The 4.8 million Early Voters include over 1 million Independents.

Florida 2024 Election Turnout Dashboard

TargetSmart's Early Vote Dashboard
 
I'll be honest—I don't like polls. I really don't. But I think that the crosstabs in the "trusted" polls are showing a very interesting story, among the folks that ALREADY voted:

ABC/Ipsos: Harris 62-33

CNN: Harris 61-36

NYT/Siena: Harris 58-40

HarrisX: Harris 61-32

USAToday/Suffolk: Harris 63-34

Flipping over to TargetSmart (Tom Bonier), if I'm reading this right (and if I'm wrong, please let me know so I can delete this post), the modeled partisanship of the early vote is (this is Tom & team guessing who is voting) is about 47% D, 44% R, and 9% I. If that's true, that means that the shift to Harris is big—a significant realignment.
 
nycfan said:




Just because Republicans are turning out in higher numbers in early voting doesn’t guaranty they are all voting for Trump. There will be some GOP anti-Trump cross-over.
Click to expand...

My parents are part of the crossover crowd. My dad can’t stand Trump and my mom has become pretty progressive but just hasn’t bothered changing her affiliation.
 
CFordUNC said:
He’s pointing out that as of now turnout in the MAGA counties (the ones that went heavily for Trump in the GOP primary) in Wisconsin is lagging that of the “more Haley” voting (in the GOP primary) counties.
Click to expand...
Meaning that some voters in the Haley counties aren't just waiting to vote on election day but might be sitting it out altogether.
 
CFordUNC said:
FLORIDA: Poll indicates that Kamala Harris leads amongst voters who have already voted. (This despite 500,000 more Republicans having voted so far.)

Florida St. Pete poll:

– Trump 50-45

– Scott 49-46

Donald Trump leads Kamala Harris in Florida, but not among those whose ballots are already cast

NOTE: If Harris really is ahead by 1% amongst those who have already voted, despite Republicans having an 11-percent Ballot Edge in the Early Vote, this is earthquake-level news for Florida

How could this be? There are two key factors:

– Women voters outnumber men 54–46 according to TargetSmart

– The 4.8 million Early Voters include over 1 million Independents.

Florida 2024 Election Turnout Dashboard

TargetSmart's Early Vote Dashboard
Click to expand...
Fascinating, if accurate, this seems big. Keep bringing it!
 
CFordUNC said:
I'll be honest—I don't like polls. I really don't. But I think that the crosstabs in the "trusted" polls are showing a very interesting story, among the folks that ALREADY voted:

ABC/Ipsos: Harris 62-33

CNN: Harris 61-36

NYT/Siena: Harris 58-40

HarrisX: Harris 61-32

USAToday/Suffolk: Harris 63-34

Flipping over to TargetSmart (Tom Bonier), if I'm reading this right (and if I'm wrong, please let me know so I can delete this post), the modeled partisanship of the early vote is (this is Tom & team guessing who is voting) is about 47% D, 44% R, and 9% I. If that's true, that means that the shift to Harris is big—a significant realignment.
Click to expand...

Capture.PNG
 
CFordUNC said:
I'll be honest—I don't like polls. I really don't. But I think that the crosstabs in the "trusted" polls are showing a very interesting story, among the folks that ALREADY voted:

ABC/Ipsos: Harris 62-33

CNN: Harris 61-36

NYT/Siena: Harris 58-40

HarrisX: Harris 61-32

USAToday/Suffolk: Harris 63-34

Flipping over to TargetSmart (Tom Bonier), if I'm reading this right (and if I'm wrong, please let me know so I can delete this post), the modeled partisanship of the early vote is (this is Tom & team guessing who is voting) is about 47% D, 44% R, and 9% I. If that's true, that means that the shift to Harris is big—a significant realignment.
Click to expand...
I honestly don't know how to read that, but I doubt you are reading it correctly. What you are seeing is not plausible.

I don't have time to look through all the polls, but what I just saw from the ABC/Ipsos poll was that 8-10% of their respondents said they have already voted. The poll was in the field from Oct 18-22. So it is working off much, much less data than Bonier. That makes the results not really comparable.

It does seem likely that Kamala is slightly outperforming Bonier's estimates but outside of that I'm not sure we can glean anything.
 
CFordUNC said:
NOTE: If Harris really is ahead by 1% amongst those who have already voted, despite Republicans having an 11-percent Ballot Edge in the Early Vote, this is earthquake-level news for Florida
Click to expand...
This isn't really accurate either. There is still a sample selection bias issue here. It's promising but you can't glean too much from it and it's definitely not earthquake-level news.

Here's the thing: if you're seeing something earthquake-level, then it's almost certainly wrong because there are thousands of trained data analysts following this and if your analysis was correct, it would be everywhere.

In particular, you'd expect independents to turn out for Kamala in greater numbers early. They will be more likely to be highly educated and not retired, meaning that they have busy schedules and are likely trying to fit voting into that. Independents who back Trump might include a lot of retired folks who don't need to vote early because they have nothing to do on election day.

You're also sort of comparing apples to oranges here because all of these polls lag in time. You're looking at a current snapshot and a poll that was in the field last week at the latest.
 
You must log in or register to reply here.
Back
Top