John Heilemann: “All the polling is telling the same story. Never before has an election been so heavily polled or suffused with data—and never has that data been so intensely scrutinized, analyzed, and algorithmically massaged. Is it possible that all this polling will turn out to be wrong?”
“Yes! In 2016 and 2020, the polling industrial complex screwed the poodle en masse, and not by a little (with 2020 producing an even bigger miss than the first time the data wizards had to contend with the Trump phenomenon).
So despite the best efforts of the industry to cure what has ailed it in the Trump era, the polls could be systematically underestimating Trump’s support again; they could also be understating Harris’s.”“When Nate Silver recently ran like a jillion simulations through his forecasting model, what the model spit back was that the most likely outcome in November wasn’t what the polling averages are telling us today… but that Harris wins all seven swing states—with Trump winning all seven the next-most likely scenario, by a hair.”
“But as we sit here today trying to decide if Democrats are right to panic, there is literally nothing in the mass of available public polling to suggest that the Blue Wall is crumbling, or that the Sunbelt swing states are out of reach, or that the race writ large is slipping away from Harris.“