If The 538 Presidential Poll....

@SW-User By Kackler picking Minnesota's Walz over Pennsylvania's Shapiro, she severely lost the Keystone State.

@SW-User You had a question about Nate Silver's projections in 2016. I was watching the election carefully, and the last week before Election Day, Clinton's chances were decreasing. Silver's last projection prior to the election had Clinton with a 64% chance of winning. It had been as high as the low 80s earlier in the campaign.

@windinhishair Still, his predictions have been getting less and less watertight. His probabilities are essentially arbitrary, and give him a get-out clause if he is wrong. Professor Lichtmann, by contrast, makes clear, unambiguous authentic predictions based on historical precedents.

He’s been right every time but one since 1984. He’s gone for Harris.

@SW-User I agree that he has become much less relevant, and for a reason. Silver's predictions this time around don't seem to be tethered to reality. I'm not really a big fan of Lichtman, either, despite his record. Remember that his predictions are for who will win the popular vote, not the Electoral College. So he correctly predicted Gore would win in 2000, and he did win the popular vote. And he incorrectly predicted Trump would win the popular vote in 2016, which did not occur. In this election, he is predicting Harris, and she will almost certainly win the popular vote. I think she will win the election too, but that is less certain given the tilt of the Electoral College to Trump.

@SW-User Did you do him yet?🤣

@SW-User Lichtman uses fundamentals that have nothing to do with polling. He was wrong in 2016 only because we're stuck with the Electoral College. The same thing could happen this year. Silver uses aggregates of polls and some fundamentals to come up with what basically amounts to a weather report.

Electoral-vote.com only uses state-level polls, averaging the ones for the past week. They consider any difference of 4% or less to be within the margin of error. As of today, they have Harris at 257, Trump at 230, with 51 tied. They've been fairly accurate. Their thinking is that errors in any one poll will cancel out if several are used. Generally, swing states are polled often, especially as the election approaches, so they have a lot to work with. They do have wonky results sometimes, like right now they have Indiana as "barely Republican" based on a poll in April that had Trump only 4 points ahead of Biden. If anyone bothers to poll Indiana again, presumably Trump and Harris will be further apart and the state will show as the solid Republican one it should be.

@LeopoldBloom You turned me on to electoral-vote.com a few years ago, and I read it every day without fail. V and Z are great!