It's easy. Take the parameters of the model (which appear to be Primary results) and plug in the known historical values. In 25 of 27 cases, the model gave the actual experienced result.
It's more than that. He uses economic indicators, no war, bonus to incumbent, etc. Etc. He said before coronavirus that it might be too late, even if the economy had some unforeseen catastrophe, people already veiw trump as a good economic leader. Let's hope he was and is right. I am far more worried about fraud.
I'm guessing not Presidential Elections or he would be pretty freaking old.
It's easy. Take the parameters of the model (which appear to be Primary results) and plug in the known historical values. In 25 of 27 cases, the model gave the actual experienced result.
It's more than that. He uses economic indicators, no war, bonus to incumbent, etc. Etc. He said before coronavirus that it might be too late, even if the economy had some unforeseen catastrophe, people already veiw trump as a good economic leader. Let's hope he was and is right. I am far more worried about fraud.
Agreed on the parameters. I think they are all pretty much available (or estimable with some degree of confidence) historically.