Ultimately, whether a question is 'too broad' is always going to be a judgement call. That's why closure is done by 5 votes, which can be countered, and not by a robot or moderator fiat. That said, the rubric that I've historically used is the one I outline in this answer, and it's comments.
Specifically:
As a general rule of thumb, if I know the scope of the answer set, my 'I know it when I see it' guideline for a question that's too broad is, "Would I be willing to take the time to compose an answer that contains all of the items in this list? Would my doing so provide information in a meaningful, readable way that makes the internet better, and provides an answer in a way that another resource either does not or can not?
and
I ask myself "If I knew this answer, would I be willing to write up the entire thing? Can I do so clearly and coherently within the limits of SE's markup?" In the case of Matrix Shards specifically, I'd say that while it's a short enough list all in all - a dozen or so entries IIRC? - I feel like enough of them are sufficiently complicated to get that I don't really like the question. I think it's a valid standard to wonder just how much effort are you asking for from the person answering your question.
Your own standards may vary somewhat, but I think that that standard is a pretty reasonable one, which doesn't require that a voter know the answer to close the question, but does require that they have some understanding of the scope of the answer. (i.e. I don't need to know the names of every President of the US to know that there have been a few dozen of them.)