This is a great question and one I've had myself, so it's something I've been paying attention to over the months.
My first (and so far only) question here on WB was reality based (even though the event I am writing about is mythological). It was put on hold because of how I structured and explained it, not for the topic. The hold was then removed when I rewrote it. It was about communications in a huge migratory group.
The entire reason I found WB in the first place was in Googling information about travel times on foot. I found an excellent question and set of answers about traveling on horseback. Absolutely nothing imaginary about it. Although the author was going to use the information in her/his fiction, the question was 100% grounded in (historical) reality and the answers as well.
Since then, I've encountered dozens of well written and accepted questions all about the real world. Mostly historical, but not questions that really belong on a history site. I've also seen a bunch of questions get labeled as "off topic" because they crossed the line into history or politics. Or not... Sometimes the designations are more random. And, yeah, they are, because it's more about who sees the question early on and votes to close. If the question is already well established by the time someone sees it, they're less likely to vote to close it.
I'm newer here but, what I've seen so far tells me that questions about things that did or could happen here on earth are welcome if they're well-written. They do need to be about an issue that builds a world: politics, society, language/communication, technology, medicine, etc. In other words, about the structure of the world.
I've also seen that questions that are obviously science-fiction or fantasy will get a lot more leeway in how well-written they need to be. I've seen some pretty wonky presentations get editing advice and lots of answers.
Like you, I've referred people from Writing to WB. Sometimes it's clear and sometimes there is a fine line. Other times, the person might benefit by formulating 2 questions: one for here and one for there. But, yeah, I haven't seen any hard and fast rule about how to weigh the importance of reality.