The 2002 midterm elections were a pretty good predictor of George W. Bush's strength heading into re-election. Bucking historical trends, Republicans won 50% of the popular vote nationwide, up 2.4% from 2000. In 2018, Republicans won 44.8% of the popular vote, down 4.3% from 2016

The 2010 midterms, in which Democrats were decimated, might seem to refute this idea: considering Obama won re-election in 2012. Dems' share of the popular vote decreased by a whopping 8.3% from 2008. GOP seized control of the House and gained the most seats since 1938. But...

Obama won the national popular vote in 2008 by a wide enough margin, 7.2%, that he could afford to lose a significant portion of it and still win re-election. And that's what happened. His margin decreased by 3.2% in 2012, to 3.9%, but he still won overall

As we all know, Trump lost the popular vote in 2016 by 2.1%. If the 2018 midterms have any predictive power, as the 2002 and 2010 midterms did, then we would expect his share of the popular vote to further decrease in 2020

Yeah, I'm aware we have an Electoral College system. You don't need to clarify that. This is just one metric for gauging an incumbent president's re-election chances. Stuff could change. But there have been plenty of signs for years that Trump is a relatively weak incumbent

Steve Bannon has said, correctly, that Trump had to draw an "inside straight" in order to win the 2016 election. He did so. Not sure what kind of hand you'd have to draw in order to overcome the structural disadvantages currently disfavoring Trump, but it'd have to be a doozy

As usual, nobody interacts with tweets that are seen as "favorable" to Biden or pro-Biden, even if it's purely descriptive rather than normative. Sad!