But do we have a forecast for whether Nate's algorithm will be accurate, based on past accuracy and factors like the economy and fascists pushing a clueless puppet again? Who's watching the watchers?
Nate is an aggregator with a model. Him, 538, and others like them are the ones that are literally watching the watchers. Silver is not a pollster, he’s someone that looks at other pollsters past performance and ranks / calculates how likely they are to be correct in current polls.
He predicted a Hillary blowout in 2016 and was one of the many reasons why people on the left underestimated Trump. I don't want to hear this man's name again.
Yup. 70% Clinton, 30% Trump. With points where it was about 50-50. They even had a 1 in 10 shot of Clinton winning the popular vote and Trump winning the election, which was higher than their chances of a Clinton blowout (6%).
I think Nate made a good point about people not understanding polls. 70% chance to win means Hillary would have won 70/100 elections, not win by 70-30. But many read 70% as some kind of guarantee.
Days before the 2016 election, 538 (which Nate Silver founded and was leading at the time) ran an article titled "Trump Is Just A Normal Polling Error Behind Clinton". Nate Silver and 538 did some of the best forecasting of that election. Don't conflate him with others' screwups.
There was a massive voter suppression campaign in 2016. Specifically there were roll purges in many of the swing states. Trump's team has inserted MAGA cultists at all levels in Georgia. They're trying to do the same throughout the rest of the swing states.
I think the messaging around polling in general is lost on most of the population and lots of people confuse the chance of winning with a prediction of the voting outcomes. This article is approximately 8 years old now and aged like milk, but comparing the odds Trump had of losing a game of russian roulette is very apt. With the benefit of hindsight, more emphasis should have been put into driving comparison home. I think that every poll should include this metric instead of trump's chance of winning.