- the ones that are too far-fetched or unrealistic
- the ones you could never prove
- the ones that are the most likely
- the ones that might have left other evidence
Start with the situations that are most likely and might have left behind other evidence. Work to prove or disprove those conjectures. You may still not have the answer, but at least you will have documentable research.
Brainstorming is good. Sometimes we need to think through totally crazy possibilities to get at other more realistic ones we might have overlooked.
And occasionally the crazy situation is what happened. But usually the reality is less dramatic--but no less interesting.