Didn't know that this thread was already underway (but probably should've assumed so). Well, in any case, 538 has started their forecast for the presidential election. For methodology: they use a Monte Carlo method, which means they take the current situation and simulate probable changes in voter behaviour or just possible votes in likely outcomes and put that together to determine the winner. That they do 40.000 times and take the percentage of who wins how often as result. 2016 they had Trump at the end at 28% (which means substantial chances for a Trump win, while a Clinton win was more likely). This time they predict 27% for Trump currently, 73% for Biden. Which still means Trump has a realistic chance.
This includes also the projection for the states if you are interested, here for example for Florida.
Surprisingly for me Biden has an outsider chance in Texas. A 30% chance for an political earthquake is somewhat interesting.
Expect the model to move as election comes closer and new polls and other data rolls in.
