Short version: the economist behind it appears to have assumed that the distribution of Trump vs. Biden voters who voted in person on the day is identical to the distribution in those who voted absentee. Which of course is insanely stupid - even for a die-hard Trump supporter.
