"12 months later, Trump would probably still win the 2016 election."

That's by Aaron Blake at WaPo. I don't know how you can compare a person who's actually been carrying out the responsibilities of the presidency with someone you're only asked to imagine as President. It tests something, but not what would happen if we could do the election again, which would require us to compare 2 imaginary presidencies.

But anyway, Imaginary President Hillary Clinton isn't doing too well these days:

The Washington Post-ABC News poll asked respondents how they'd vote in a redo of the 2016 election, and, if anything, Clinton seems to have lost more ground than Trump. Among those who voted, 46 percent say they picked Clinton last year and 43 percent picked Trump — a slightly more favorable sample than the 2016 election, in which Clinton won the popular vote by two percentage points. But in a head-to-head rematch, Clinton's support drops even more than Trump's does, and they wind up in a 40-40 tie. Given that Trump overperformed in key, blue-leaning swing states, that means he'd probably have won again....

[J]ust 72 percent of Democrats said they would vote for Clinton in a rematch — vs. the 84 percent who said they did vote for Clinton last year.

I think this poll makes more sense than the standard day-to-day favorability rating, where Trump seems to be languishing in the 30s, because it's relatively easy to say we're not happy enough with what's going on than it is to point to somebody else who you think would make you happier.

It should be noted that the Democrats in that poll who dropped their support for Clinton did not say they'd vote for Trump. They'd vote for a third party candidate or nobody. And thinking about Clinton is distorted by the knowledge that she, in fact, lost. They might be dropping her because she's a loser.