I think part of the expectation is that you should have done some "research" into the question yourself before posting, but it doesn't have to be a serious amount of research, or about an area in which you are an expert. For instance, my most recent question Integrality of iterates of rational functions was something I just wondered about randomly, not related to my own research. So thought about a little myself and tried googling first, and didn't see a clear answer and tried asking here (with the knowledge that I was asking from a very naive perspective, but I had some idea the question that other mathematicians think it an interesting question). It turned out the problem was solved over 20 years ago, and was easy to answer for someone who knew about arithmetic dynamics, but not for me.
The point is we don't want the site watered down with poor questions, and if you have looked into the question for yourself first, and still can't answer it, probably you can at least ask a better question (and, as mentioned in the comments, using your real name and establishing a track record can help get your questions taken more seriously).
Of course, the question should not just show that you've put some effort into it, but for this site it should be of interest to research mathematicians. It's hard to know in advance, if you're not a mathematician or not very familiar with MO, what will be of interest. (In fact, often when we write papers, we don't know whether they'll be of interest or not!) But a decent rule of thumb is that if the question seems to be in the realm of college-level math or lower, then try asking on Mathematics.SE, but if you have no luck there or it seems to require more advanced math, try asking here.