I'm sure everyone is sick of this post already, but here's a (somewhat) new angle as to why these questions shouldn't be duplicates. (This was already discussed somewhat in Chat starting here, but I think it should be brought here so it can be included in the voting.)
Today I edited my answer to the question about the number of students in Hogwarts to add in some evidence and counter-evidence. As a potential explanation for a certain discrepancy, I suggested that Ravenclaw may have had many more students than the other houses (in Harry's year). I pointed out that this could account for why Gryffindor has classes with Slytherin and Hufflepuff but not with Ravenclaw.
I then checked to see if there was any discussion on this site about Harry never having classes with Ravenclaw, and I found this question which asks if Harry ever had classes with Ravenclaw.
According to the stricter standard for duplicates, the old question should be closed as a duplicate of the question I answered.1 This is because even though "How many students attended Hogwarts?" and "Did Harry ever have class with Ravenclaw?" are two different questions, the latter is answered by one of the answers to the former, and thus the information already exists and is at risk of being repeated. To quote from TheLethalCarrot's answer (name adjusted to reflect this case):
In this case @Alex's answer can be copied over to the dupe
without changing it and it would answer the question fine, sure it
would have some extra rambling but it would answer the question.
Just a couple of hours later I was reading this question about whether Dumbledore set up the Potters to be killed. In one of the answers I saw a quote which said that the secret cannot be forced out of the secret-keeper. I suspected that there was bound to be a question here about that very issue, and I found this question. Therefore, according to the strict duplicate policy the latter question should be closed because it is already answered in an answer to the former question. Once again we can say:
In this case @erip's answer can be copied over to the dupe without
changing it and it would answer the question fine, sure it would have
some extra rambling but it would answer the question.
I subsequently found another example while reading this question about how Arthur Weasley knew to send his Patronus to Grimmauld Place. One of the answers stated that a patronus has the ability to find someone, and tried to prove this from the fact that Snape must have been able to locate Harry in the Forest of Dean by following his patronus. There is another question that asks how Snape was able to find Harry in the Forest of Dean. Even though these are two different questions, the latter is answered by the answer to the former. Once again we can apply TheLethalCarrot's criterion (and this time we don't even have to change the name because it's the same user):
In this case @Slytherincess' answer can be copied over to the dupe
without changing it and it would answer the question fine, sure it
would have some extra rambling but it would answer the question.
I found these three situations with virtually no effort, and without spending any time "searching them out". I imagine that there are probably dozens of other such situations.
This would mean that anytime someone incorporates information from an answer into an answer to another question they will be rendering the first question a duplicate.
Are we really going to close all these questions just because an answer to a different question contains information that could be an answer to the current question as well?
In fact, it is even possible that someone could deliberately take content from an answer to one question and use it as part of an answer to a different question, rendering the first question a duplicate. (I don't know why someone would want to do that, but it's possible, and it would be hard to prove nefarious intent.)
In sum, my point is that under the strict interpretation of "duplicate", almost any question can be made into a duplicate of any other. There are probably already dozens of such situations. This, I would argue, is indicative of a major flaw in the strict definition of "duplicate".
1 The question has since been closed as a duplicate of a different post, but that is not relevant to my argument.