Two years ago, "not a real question" was modified to include "too broad" as a reason. Having spent some time on the review tool, I now believe that closing questions that are too broad as "not a real question" isn't really adequate.
First off, "not a real question" for broad questions is misleading. Often, broad questions are very real questions, like "how do I build a social networking site?" It's just that they're not the format of question that Stack Overflow aims to answer. This is fundamentally different to questions are actually are not questions, like "I made this algorithm."
A real "too broad" category would make it clearer that a question is closed because the user needs to pinpoint something more particular about a problem to get help, and would be good counterbalance to "too localized".
I make the distinction between "too broad" and "not a real question" in that questions that are too broad would require an answer with multiple steps, such that it's clear what's being asked but they're too complex to tackle in a single post, while "not real questions" are simply ambiguous and unclear.
The text could be like:
Questions on this site are expected to pinpoint a particular issue with a certain approach to a problem. This question seeks to find way to approach a problem, or would require an answer with multiple aspects.
This close reason could then also be used for questions that are made of multiple questions.