One of the best respected analysts in the business, Nate Silver, currently has Trump at a 53% probability of winning (https://projects.fivethirtyeight.com/2024-election-forecast/), in a system designed to magnify the effect of small changes near the halfway point (historically, getting about 65% of the popular vote would net you a landslide worth approximately every seat in legislature - https://www.pewresearch.org/short-reads/2020/12/11/bidens-vi...). Silver's model can pick Trump as a slight favourite despite being behind in the popular vote, again because of the nature of the system. In particular, each state gets an extra constant two votes on top of the ones apportioned by population, favouring rural areas which currently prefer the Republican party.
The same analyst, BTW, gave Trump a ~30% chance of winning, shortly before the election, in 2016 (https://projects.fivethirtyeight.com/2016-election-forecast/). This was notably higher than the probability ascribed by pretty much anyone else with any authority or respect in the matter.
You mean G. Elliot Morris; Nate Silver has nothing to do with fivethirtyeight any more (since he and Disney broke up, he now has his own private subscription publication—Silver Bulletin—which also does forecasting and which has much better odds for Trump from what I understand, but I can't give you details because its paywalled and I’m not paying him; aside from that gig, he’s also employed by Polymarket, now.)
My understanding is that Silver's model has been more favorable to Trump than 538, by virtue of 538 having Kamala up just a little while Silver had her up less or Trump up a little. In the past couple days 538 shifted a little toward Trump; I'm not sure Silver's model did the same. His updates have been saying "small moves, still basically tied" for a while, but like you I'm outside the paywall.
One of the best respected analysts in the business, Nate Silver, currently has Trump at a 53% probability of winning (https://projects.fivethirtyeight.com/2024-election-forecast/), in a system designed to magnify the effect of small changes near the halfway point (historically, getting about 65% of the popular vote would net you a landslide worth approximately every seat in legislature - https://www.pewresearch.org/short-reads/2020/12/11/bidens-vi...). Silver's model can pick Trump as a slight favourite despite being behind in the popular vote, again because of the nature of the system. In particular, each state gets an extra constant two votes on top of the ones apportioned by population, favouring rural areas which currently prefer the Republican party.
The same analyst, BTW, gave Trump a ~30% chance of winning, shortly before the election, in 2016 (https://projects.fivethirtyeight.com/2016-election-forecast/). This was notably higher than the probability ascribed by pretty much anyone else with any authority or respect in the matter.