Yes, we should close opinion-based questions
You are saying that this site is about interesting discussions. But it's not. Every site on the StackExchange network is a Q&A that is aimed at helping people that have a problem finding a solution to said problem. The site is not meant to generate long discussions where people voice their opinions and develop ideas in an iterative and open-ended way.
But we are different
Now we come to a big problem: WorldBuilding is different from a lot of other sites on the network. On StackOverflow for example most questions are real-world problems that people have right now and where they need help as fast as possible because time wasted trying out things without any result is exactly that - time wasted. Which leads to money that is wasted, because for most of the active users there programming is their job.
On WorldBuilding things look a bit different. Chances are you found this site via a different site on the StackExchange network through the Hot Network Questions and thought that a title sounds interesting or funny. But we are still aimed at people needing help. People who are creating a world for their book, or their RPG campaign, or their game, or because they like to imagine things in their free time without any clear "goal", or any of dozens of other reasons. But they all have something in common: if you want to have an internally consistent world you will most likely at some point have problems. And this is when WorldBuilding.SE comes into play.
We are aimed at providing clear help. You have a problem? We have an answer!
The goal is not to brainstorm with you. You are the one creating the world - not us. You have a problem - not us. And it's not possible to help you get a clear answer if your question is aimed at starting a discussion. That doesn't work and as Molot referenced, there are a lot of forums out there that are better suited for open-ended collaborative discussions.
We will always be a bit opinion-based, but questions should still be written in a way that you can use somewhat objective methods to determine which answer is the "best" answer in the proposed scenario.
Drawing the line
Now we come to the real problem. Where to draw the line? What is "too opinion-based"?
Many people believe there is a razor thin line and that some questions cross this nearly invisible line and some stay on the "leave it open"-side. Looking at the Hot Network Questions from WorldBuilding I regularly find myself wondering how those are not closed immediately. There are so many better examples of good questions and answers on the site. Therefore I propose a different point of view on the "drawing a line" debate:
It's not a razor thin line - it's a big blurry region!
On WorldBuilding questions regularly are a bit opinion-based. And the opinions of users about what counts as "opinion-based" vary greatly. Even for a single user it can vary greatly depending on the topic asked. Maybe people are more lenient when a question is asking about planet-scale stuff than when it's asking about regional stuff. Maybe they are more lenient when it comes to weapon-design or creature-design than when it comes to tectonics. Maybe they are fans of exploring time-travelling ideas and problems or they favor any of the hundreds of other things that people ask about here.
This leads to a situation where your big blurry region does not align perfectly with my big blurry region. And there is no way to say which big blurry region is the "best" one. That happens on such a Meta-Point-of-View.
Sometimes certain words may trigger people to Vote-to-Close a question because they feel it "crossed their line". It's just that the question is on the "opinion-based" end of their big blurry region and there are certain words that they have seen too often in not acceptable questions. For example I know that I have to be extra careful when reading questions that use the words "How", "would", "this", "affect" and "society" in one sentence together with a question mark at the end. A question like this is very often opinion-based and too broad. Using those words in certain constellations might push a question that is on the "opinion-based" end of my "still acceptable" big blurry line to the "not acceptable" side. Realizing that I have this bias was not easy. And even if you had the same big blurry region as I have, but not this bias towards certain combinations of words, you might think this question is acceptable. Then we take another couple dozen people and see how their personal big blurry regions align with ours. And suddenly a question crossed the "razor thin line" of what is acceptable and what is not acceptable.
Comparing two questions with each other you might feel like both are the same and both are on your "acceptable" side, but maybe they are both on the big blurry region and a few people had a slightly different big blurry region leading to the question being closed.
Putting a question on hold is not the end
You can always start a Meta discussion to get a certain question reopened. And you also just cast a Vote-to-Reopen if you have the corresponding privilege. Everyone can only vote to close or reopen on any given question once and if you find enough people with a similar big blurry region you might get a question reopened that was deemed as off-topic by a different group before. If a question is closed and reopened multiple times this question is a perfect example for these big blurry regions. If you take enough people you will have used the whole possible spectrum, so the region becomes even more blurry. Sometimes it may depend on who is currently online and who is voting and in which mood they are and a whole ton of other factors that may lead to a question not getting closed when it should be closed and suddenly becoming an HNQ.
If you want open-ended opinion-based discussions you should check out the chat. We often discuss whatever comes to mind there and would be glad to have new interesting and funny conversations. The way things are handled in the chat just wouldn't work on the Main Site. StackExchange is not good at this kind of discussions, but the chat is a good place to talk, to discuss things and to get immediate feedback. And your stuff won't get closed or downvoted there just because you are brainstorming.
Old questions are not a good indicator for what is currently on-topic
I want to say a few words about your examples. First of all you have to be careful when reading old questions. Many of those are not what would now be considered on-topic. Often they are not closed because they were on-topic at the time and nobody bothered reactivating them once they were not on-topic anymore.
The same applies to Hot Network Questions. A lot of those are getting closed after a few days. But thousands of people have seen them until that happens and they have a lot of votes and often a lot of answers. Be careful when looking at the HNQ's of any site on the StackExchange network. Those rarely reflect what the site would normally deem a "good" question. They attract the attention, votes and answers of people who normally are not on the site and probably don't know specific details about what should be considered a good question on the site.
In my opinion both of your examples are perfect examples of what I just described. They are nearly one and two years old and looking at their views/votes/answers I assume they have been on the HNQ at some point. And I would say that both of them are opinion-based and should be closed by current days guidelines.
One of our moderators mentioned in the comments:
Santa/Satan was asked at Christmas as part of a topic challenge, that probable bought it some leeway at the time as well.
The challenge in question can be found here: Fortnightly Topic Challenge #23: Santa. This shows that other circumstances can contribute to a question being well-received, although it might not have been a great example for the site. In this case especially the timing and maybe the Meta Effect, though that is not such a big problem on WorldBuilding most of the time, as the traffic tends to not be that high.
Why don't you flag old questions in that case?
I wouldn't flag such old questions because they show what was on-topic once and it's not worth it to start discussions about such old questions that are not active anymore. That's one of the cases where my big blurry region wanders to the very lenient side and I go with the "live and let live"-approach that some people wish for the site in general.
It's different with active questions. They should represent what the goal of the community is. And if the current guidelines say that we want concrete answers to concrete questions then questions that invite people for brainstorming sessions are off-topic and should be flagges accordingly.
Optimizing for pearls, not sand
There has been a related answer to a similar problem on RPG.SE: This site is too heavily moderated, which mentions the official statement Optimizing for pearls, not sand.
That’s why we’re determined to keep question quality high, even at the cost of refusing a little sand. It’s true that you can’t have Q&A; without questions, but having the wrong sorts of questions is far more dangerous
and
in the long run we’d much rather err on the side of having interesting and on-topic questions
I agree with this sentiment. We should aim at having high-quality questions with high-quality answers. It's a personal take on why I am on this site. I don't want to read tons of half-baked brainstorming sessions. I want to read good questions with great answers that make me think "Wow, that's brilliant!".
Of course this is just one of the reasons I am on this site and a bit exaggerated. I want to help people who are having problems, but I am more of a reader than a writer, so for the current discussion the above is my personal sentiment.
This doesn't mean that we should disallow all sand and only look at the pearls. There are many mediocre questions that are worth reading and answering. But we should try to minimise the questions that we deem off-topic and allowing people to participate in open-ended discussions will only lead to everything being on-topic and thereby burrying the pearls we have under tons of sand, which makes it difficult to find the pearls, because everybody on the internet has an opinion and will voice it, especially when asked for it.
Don't take it too personal
Sometimes questions are closed where you are one-hundred percent sure that they should not be closed. And sometimes there are HNQ's where you are one-hundred percent sure that they should be closed.
And often you are right with that.
It takes some time for the "correct" judgment to be found and sometimes things slip through, reviewers have a bad day or some new people got some privileges and want to see how closing a question works without looking through too many Meta discussions. This happens.
Of course you shouldn't take these things personal anyway, especially if it's not even about your own questions. But I am pretty sure that everyone who has been active on the site for more than, let's say, a month or two has seen questions that he thought were great getting closed and feeling a bit disappointed, which is why I say "Don't take it too personal".
Summary
- SE is not made for discussions
- WorldBuilding is inherently opinion-based
- There is no razor thin line, but instead a big blurry region and everybody thinks the region should be slightly different in size and shape than someone else proposed
- Sometimes certain topics or words can trigger some people to think a question "crossed the line"
- Just wait and see what the majority thinks
- If waiting didn't resolve anything ask on Meta, come to the chat and Vote-to-Reopen
- Old questions can be bad indicators of what is on-topic and what is off-topic
- Try to aim for the high quality content - quality before quantity
Life is not black and white - there are a lot of different shades of grey that you have to live with