"List question" is a very broad and unspecific term. There a lot of things that can be qualified as "list questions". But only some of them are unacceptable and don't entirely work. Even your examples in #2 I wouldn't consider the problem to necessarily be that they are a list, but perhaps one of scope. As for unacceptability, I refer to this post of mine that bwarner linked already as a good starting read.
But the problem you're trying to determine here seems to be more about when there are multiple possible answers to a question, and not that it is an endless repository of options. Let's lay down some ground points then.
- It is perfectly acceptable for there to be more than one correct answer to a question. This is why we even have a voting system at all, and why we allow multiple answers. This does not mean that "Of these three cards, which is the best for a drain deck?" should have three answers, though.
- If the answer a question needs is a list, then the list should fit in one answer. An example I use from Arqade is "Where do I find all of the subtanks in Megaman ZX?". There are exactly four of these - a proper answer will list all four. This should not get four answers, one for each location.
- The word "best" in a title is not an instant condemnation sign. Implicitly, any question asking for the solution to a problem is seeking the best solution to that problem. A question whose core is still fully legitimate, just the presence of the word "best" is irksome, that is still a legitimate question. This does not mean that this question should get a separate answer for every possible best option, though.
There's a resonating point that I made in all of these. Which is that a lot of questions will benefit a lot more from a comprehensive answer that covers multiple points, than one that just covers a single one. The conclusion can still be a single one.
Let us look at your two examples.
What is the best spell for a multiplayer drain deck, Drain Life, Exsanguinate, or Bond of Agony
A great answer to this question will cover all three cards, explain the advantages to each one, and either come to a conclusion of a single card being more universally useful or a conclusion on which scenarios will make which cards shine. A good answer will highlight one card and explain why it is more useful than the other two. A meh answer will just name a card and notate why it is good. A bad answer will just name one card.
Is there opinion? Yes, but the more opinionated an answer is, the less helpful an answer is. The more fundamental and fact-backed an answer is, the more useful it is to other people and thus the more rated it will be. What makes this still a helpful question is that there still is the room for fact-backed answers. At least, in theory. I don't know these cards so I have no idea, but my assumption is that the root question is as such.
What is the best spell for a multiplayer drain deck?
As we start cutting things down, what's happening is that the scope is losing helpfulness. When you reach this point, an asker should be thinking, "Is the answer too situational for this question to be helpful to others?" I don't know drain decks well, but for example, "What is the best set of Slivers for a Sliver deck?" is too broad. Sliver deck is a strategy, certainly, but it also makes use of other strategies. Do you go Shadow/Poison and go for direct player kill? Do you go stat-stacking and overwhelm by Power/Toughness monstrosities? The reason Slivers work is because you have a lot of options that all work with a core theme.
I hope this illustrates my point well enough. It's been too long since I played a deck that I know an example that isn't too broad. But the basic point is, you don't need to specify "What is the best out of these options" and can instead opt for "What are the best options for this strategy?" if the latter question is narrow enough in scope that the answers will be genuinely helpful. This can be the case even if there is more than one opinion on the answer - the ultimate core of the question is "How can I accomplish this strategy?", which you'll note the lack of "best" but it will still get the same answers.
What is the best black multiplayer spell?
This is when it obviously gets to be too broad. I don't think I need to cover this in more depth beyond what I said in the previous point. At this point, though, the scope is so large that it illustrates the problem most people think of with lists. That is, because there isn't a narrow guideline, all people have to go on in answering this is their opinions. Factual stricture like saying "Well, for this strategy, the best option is..." starts to get away from the question that was actually asked. Which, while the answer is useful, it doesn't handle the question itself. That makes the question bad.
Takeaway. If it invites a list, that doesn't make it bad. If the list isn't useful because it's too broad, that is a problem. If the list isn't useful because the scope makes a non-opinionated answer difficult to spring up properly, that is a problem. But if there's multiple options and perspectives to solve a problem, it can still be a valid question. I recommend this post I wrote as just some further reading on the subject of this specifically.