I’ve updated the graph showing the probability of Trump winning in 538’s model.
In 2016 his chances moved between 12% and 48% – a lot of variation.
In 2020 his high point has been 31% (higher than election eve 2016 which of course he won) and the low point 21%.
September has seen his chances slide from 31% to 24%. The Woodward book is probably not helping him as it means the headlines are about his handling of Covid, not the issues he wants to campaign on.