There is a flaw (for me, a fatal flaw) in your argument:

So, according to this strategy, voting for Trump is, in absolute terms, very far from ideal, but given the comparative damage Clinton might do, it turns out to be the best choice one can make.

This was precisely the strategy we saw from many who opposed Trump during the primaries but voted for him in the general.

While I don’t find the floating threshold strategy persuasive, there is no compelling reason to think it is a morally compromised strategy.

Let’s use symbols so we don’t restrict ourselves to the perennial Clinton-Trump rehashing.

If Voter X chooses Candidate A over Candidate B because Voter X’s reason for opposing Candidate B is itself morally compromised, then Voter X’s support of Candidate A is morally compromised (whether Candidate A is herself / himself morally compromised).

To fill in the hypothetical blanks with a more specific but also hypothetical scenario:

Suppose the remaining two Democratic hopefuls are Pete Buttigieg and Elizabeth Warren. Suppose my decision-making process is informed exclusively by my homophobia (which in reality I consider morally indefensible as a philosophy — this is hypothetical of course). I may vote for Elizabeth Warren because I perceive that she is not a lesbian and Buttigieg is gay. Elizabeth Warren herself may or may not be homophobic, but my support of her is morally compromised.

Trump may or may not be morally compromised, but that has no bearing on whether or not his supporters are.

Most people vote AGAINST rather than FOR someone.