I appreciate you supporting your position with reasons, but it’s really not that easy. My argument had four premises and two conclusions. Neither the premises nor the reasoning depended on the idea that polls will accurately predict the winner in 2020. Polls are only one indication of what people will end up doing when the time comes to vote. Even so, the polls in 2016 were not wrong. As I remember it, in the last days before the election, Clinton’s average lead in respectable polls was only a few percentage points. She won the popular vote by more than two percent, and the polls predicted that.

I can summarize my argument for you in even less than six lines. Trump won in 2016 with virtually no margin for error, so he can’t lose any support unless the Democrat loses the same amount. Trump has lost a lot of support, and the Democrats have gained a lot, and so Trump is on pace to lose. The polls don’t need to predict anything for this argument to be sound. They just need to reflect people’s actual feelings about Trump with some level of reliability, and they do that. They did it in 2016, and they did it in 2018.

The point about the economy means very little in 2020. It’s a correlation we’ve observed in the past, but it’s fallible. We’ve never had a president as unpopular or as unfit for office as Trump. It’s exactly the kind of aberration that could break a historical pattern like the one you pointed out.