(Providing an answer, long after the fact)
Answers made in good faith prior to an edit should not be downvoted if made invalid by that edit.
Since this question was asked, we've largely established that it is poor form to change a question so substantially through edits and clarifications that existing answers could be invalidated. In particular, the results of edits made in the circumstances described in this question are the reasons we give for telling question posters NOT to do that. For example, my usual comment to someone seeking to make a significant edit to a question goes something like:
Changing your question after you've received answers is inappropriate, as it invalidates the answers you've received. It can even make those answers wrong, and adversely affect the reputation of those who answered. If you now have a new or additional question, create a new post and ask it there; you can link back to this one if needed for reference.
It is manifestly unfair to leave answers made prior to invalidating edits subject to downvotes by people unfamiliar with the history of the question, if the answer was a fitting answer to the question as it stood at the time the answer was made. For that reason, we strongly discourage such edits.
In a lengthy answer to the question What should I do if I've made a mistake in my question?, I offer a pretty complete explanation to puzzle setters of the impact their post-answers edits can have on the answerers, and how the setter really should bear the burden of mitigating the collateral damage of their edits to their question.
In particular, the following advice to answerers was offered:
For puzzles edited to correct an actual error:
Answerers are often invested in such puzzles, as their errors often turn up when would-be solvers have already worked on the puzzle until they reach an impasse caused by the error. The answerers will hopefully update their solutions, or—if they abandon the puzzle due to the setback caused by the error—ideally should note their solution is wrong/incomplete because the puzzle changed.
For puzzles edited to fix being too broad, unclear, opinion-based, subjective, and/or speculative:
Answerers who posted "answers" (that is, not good-faith attempts to actually solve the puzzle) just to point out the flaw should update or delete their non-solutions, their purpose having been served (and, failing to do so, should not expect immunity from downvotes later). Those who posted good-faith answers to the original puzzle may wish now to update their answer, or if they still feel it's a good answer, may want to add a note to it to indicate why it no longer matches the question as well (or at all).
Having said that, in all cases it is the setter who needs to bear the burden for their puzzle's flaws! Thus, setters should take particular notice of Can something be done about the modification of questions and its accepted answer, which notes that if a substantive change is made to a question, the edit should also include an explicit notice that the question has changed, thus providing an explanation as to why "answers fail to match the question through no fault of their authors."
This answer now stands as a reference for aggrieved answerers to point to, in the event that they have an originally valid, good-faith answer invalidated by subsequent edits and then start getting downvotes.
Hope that helps!
Specific to this question, I'll add a couple final thoughts.
A question with flaws or which is under-specified enough to allow answers beyond the intended one really should be closed, not answered. If the flaw is not glaring, and a perfectly reasonable answer can be made that relies on what eventually turns out to have been a failing in the puzzle's construction, it's understandable that such answers may be submitted before the flaw is noticed and fixed. This was the case with The isolated kingdom, a question referenced by the question here, where a flaw in the puzzle left open an unintended but plausible answer.
However, in this case an answer exploiting that flaw was first made by another answerer, and mirrored by Ian MacDonald's (now self-deleted) answer. As Ian notes in the question here, his answer drew downvotes (and, contemporaneously, he commented that "I'm really disappointed that this community downvotes answers that were made before edits to the question"). But those downvotes actually preceded the edit that invalidated the answer, and were perhaps due to the fact that his answer was actually a duplicate. (Indeed the first answerer's post ended up at +7/-3, but only got one of those downvotes after the puzzle was edited, and in fact got an additional upvote or two after that edit; not all invalidating edits do great harm to the answers they uproot. Still, there was some harm, and we should strive to avoid that.)
The other question and answer referenced in by the question here seems more on point. A comment there has Ian noting "Not my fault the puzzle has multiple answers" under his answer, which appears to have received only one vote, a downvote made after the question received a clarifying edit... exactly what we don't want to see. (Playing devil's advocate a bit, I'll note that Ian's answer does seem to be what triggered the clarifying edit, but I'm not entirely convinced the answer was "correct". Still, I wouldn't have downvoted it.)
