(09-03-2024, 05:39 PM)OG-JAGFAN Wrote:

(09-03-2024, 04:38 PM)Lucky2Last Wrote: I'm going to need a link to that.

---

I don't see anywhere on his site or twitter that reflects those numbers.

Polls mean squat.  I learned that the hard way when I was foreign to politics and watched the Hillary/Trump Election.  I was watching all these "expert" and "professional" pollsters all swearing that Hillary was easily going to take the win. They all had this smug arrogance about them with all the polls showing HIllary would win.  They learned a hard lesson not to trust the polls that night.

Most of the polls do not tell you "who" they are polling.  Prime example,  IF an "expert" interviewed 100,000 people and had data that Harris was going to easily win, that would be on the front page of the NY times and all over ABC/CNN/MSNBC and every other news outlet .  The part that same pollster may have left out is he did the poll in California, or the North east, or somewhere crazy liberal.  My point being is outside of clear labeled national polls you cannot trust them.

Nate Silver's model attempts to aggregate multiple polls and figure out the most likely result in each state.  There are definitely polls like you describe where they fail to get a "representative sample" as they say, but Nate Silver claims he can aggregate multiple polls to truly get a good sample.

But he has never claimed to predict a winner.  He's like a weatherman, he's only giving probabilities.  In 2016, his final model gave Hillary a 66% chance of winning.  People got really mad at him and called him a fool when Trump won instead, but all he said was, the election would be like shuffling a black card with two red cards.  Draw a random card.  If it's black, Trump won. Right now he's saying the deck has five black cards and four red.  Trump is more likely to win.

---