Vox: “President Donald Trump is polling significantly worse than he was at this point in 2016, both nationally and in key states. His chances of victory in the FiveThirtyEight forecast model (11 percent) are lower than they ended up last time (28 percent), as of October 28. But those chances aren’t zero. So what would it take for Trump to win? The most likely victory scenario for the president is a bit of a stretch, but not that complicated.”

“First, he needs about a 3-point polling error or a late switch of votes in his favor in most swing states. Going off FiveThirtyEight’s polling averages, that would be enough to push Arizona, Florida, North Carolina, Georgia, and Iowa — states Biden narrowly leads — into Trump’s column. But that on its own wouldn’t provide Trump enough electoral votes. Trump also needs to win a big state or multiple smaller contests where Biden has a larger lead. His best prospect for that appears to be Pennsylvania, where Biden is up by a little over 5 percentage points in FiveThirtyEight’s average. There’s little indication that this is a particularly likely scenario. Experts who picked up on signs that Trump could win in 2016 are generally not seeing the same signs this time around. But it is a scenario that can’t entirely be ruled out until the votes are counted. Indeed, it’s pretty similar to the analysis of Trump’s path to victory I wrote just before the last presidential election: run the table in the very close states, and then break into the blue wall.”