Polling is completely unreliable at this point

They frequently ask people. Always see them in the streets via Jessie Watters. Seems pretty accurate to me.

 

Great post. After reading…..it begs the question….. Why are you talking to your ex? Is she that attractive ? Kidding. Always appreciate good info

 

Maybe you people should follow the best pollsters and not the Intentionally biased ones or those who suck at polling. Most of you ate really slow learners or you're really lazy. Your hyperpartisanship keeps you from listening to people with a clue. I posted videos on the Trump VS Harris polling thread that were a masterclass in understanding polling

If you bet your own money on elections you'd learn to discard your motivated reasoning and follow pollsters that don't suck. If you are a LOSER you follow pollsters who suck. Stop being a lover, you can do it.

My favorite pollsters are Rich Baris @Peoples_Pundit (Big Data poll) and Trafalgar. Rasmussen is okay, Mark Mitchell has been schooled somewhat by Rich Baris. I don't know as much about AtlasIntel snd Insider Advantage but they seem competent.

Political super bettor Robert Barnes moves political prediction markets and he rates Rich Baris #1 for accuracy from 2014-2024.

538 sucks. They dropped Rasmussen because it was too accurate. I've posted the receipts 2 or 3 times.

Nate Silver is a total fraud. I told you what his final prediction for this election would be months ago. He goes back and forth on who will win to generate clicks. Like many pollsters he intentionally overstates how well Democrats are doing to repress Republican fundraising and turnout. His credibility is so low he has to rebuild it by being honest in his final prediction this year, which will show a Trump victory as I told you months ago

People make up their minds on who they're voting for (incumbent or the other party) about a year in advance. A Trump victory was known to be the outcome of this election a year ago based on:

the economy (moving into recession, inflation)
the border in chaos
war abroad

If you want to learn about polling listen to Rich Baris on his Inside the Numbers podcast @Peoples_Pundit or during election seasons listen to Rich on Ehat are the Odds when he is joined by Robert Barnes. This podcast is on Mondays or Tuesdays. Listen and learn or remain ignorant

 

Agreed on OP. It'll be interesting to see if Allan Lichtman's reliance on "keys" predicts the winner yet again. I think the only time his method didn't pan out was Bush/Gore and that outcome wasn't exactly clear.

Historian has correctly predicted 9 of past 10 elections. Here's how.

 

4

Word salad

 

Thanks sfla. What your ex says that there are unknown unknowns in polling error must be true to some extent. I posted this study a while ago, showing that empirical validation of forecasts might take thousands of years of data.

https://www.politico.com/news/magazine/2024/09/03/election-forecasts-data-00176905

The conclusion that we draw from this regarding the utility of polling need not be such stark despair, however. While it is clear that we cannot know with 100% certainty who will win the election, I do think these data can help us to narrow the possibilities. I think we can rule out either candidate winning the popular vote by 10%.

And if we can’t know how the error will break, can we at least guess? I think we can using community level knowledge. While we shouldn’t trust Siena or Tragalgar to single-handedly crack the problem of error, we will likely get a hint by aggregating all the different ways that the various pollsters try. Such an approach has delivered many stunning successes, such as discovering the whereabouts of the lost Scorpion submarine, when an aggregated guess predicted it would be found a mere 220 yards from its actual location. The unknown error on that task was obviously also huge, but somehow the aggregate guess still produced a useful result. Anyway, even if our predictions aren’t foolproof, I do think there is some utility to be found there.

 

The possibilities are already pretty narrow without conducting a single poll, either candidate A or B wins, and margin of victory in PV or EC is not meaningful in determining who governs post-election, since its winner take all. And if no one is winning by 10% you are basically within range of a 50/50 race. I think polling has reached the point to where too much information becomes no information, and an aggregator shifting .2 percent becomes a story (even though its statistically meaningless) because our elections last forever.

 

You are right that in a close election noise can easily swamp the pertinent signal (who will win), but that we can even identify this as a close election is itself a valuable result. And I agree that people get way too caught up in the tiny ups and downs of constant polling. It’s like checking your retirement account everyday while you are in your 30s.