Because it came up indirectly over at Grim's I looked at the Libertarian vote in the 2016 election. My reasoning is that some, maybe even many Gary Johnson votes would now happily go for Trump instead. If only a few had switched then, NH and Maine would have gone for Trump. If most of the Libertarian votes had gone to Trump he would have carried New Mexico, Colorado, Minnesota, and Nevada. I don't think that high a percentage would be anything like guaranteed, but if a non-race suddenly gets close, different people turn out on election day. Even Virginia would be in play, though Johnson's votes would not have been enough to turn the state. Adding Evan McMullin's votes wouldn't be enough, but the Old Dominion suddenly looks darn close.
I don't know if everyone who voted for Trump still would, nor how many others would now show up/not show up. Conservatives can be irascible and hold grudges, vowing never to vote for a candidate, not because s/he had serious disagreement, but because s/he let them down by not going to the mat over an issue or two that the rest of us would just accept as the unfortunate reality of politics. Libertarians are that, squared. I wish they wouldn't do that - but then they wouldn't be who they are. That tendency to dig in over small things may have been one of the best slowers of cultural change over the years. I think not. But I have no crystal ball, and they are who they are. An unknown percentage of them would vote Trump today. I suspect there are almost no defections going the other way at present. No Trump voters who now wish they had cast their vote for Gary Johnson.
The ex-Democrats and stay-at-homes from the Rust Belt states who voted for Trump are another matter. Republicans keep treating them as solid because the job reports are good, but that's just theory. Some may regret their vote now. He may have picked off all the low-hanging fruit and there aren't many more Michigan votes to be had. If the jobs start trickling away the votes probably will, too. Still, I would bet that right this minute he would have more converts than defections.
There is another factor: People voting like their neighbors. We see elections in terms of states, but when you look at each state by county, it is impressive how strong the populous city versus less-dense countryside holds up blue-red state by state. The (populous) deep blue and (often empty) deep red counties of New York aren't going to change. Yet look at those pink counties, who now realise that their neighbors didn't vote overwhelmingly for Clinton after all. Even the light blue areas. Don't we imagine that people in those places had the impression that they were in 65-35 deep blue pockets? Wasn't that one of those things that everyone just knows? Except it wasn't true. Are those voters encouraged, more willing to speak up now? Will there be more bumper stickers and signs next time? Oregon, too. There could be some cascade preference.
This is not just speculation, it is based on some real numbers. However, it is very preliminary, and a hundred things could change it: market crashes, wars, disasters. Still it is worth thinking in this state-by state way. Nate Silver didn't name his site 538 for nothing.
One more thing: the youth vote is always deeply influenced by what they think the trend is. The usually show a heavy bandwagon effect. Of those who were 14-17 during the last election, do you think they will be more Democratic than the last batch? They might, but for what reason? They might be a Bernie-loving bunch who will follow whatever Pied Piper rises to the top. Hillary did not capture the youth vote as Obama did, and whoever the Democrats put up next time might do better than she did. But their coming-of-age politically is less poisoned by the idea that "Everyone hates Trump." I forecast that they will be less Democratic than the last cohort.