The problem of coincidence in stories is that they, by nature, fundamentally violate setup/payoff and human conceptions of cause and effect. They are also incredibly easy for a bad writer to abuse.

The solution that EFAP (and myself) have come to is that coincidences are okay if:

• They begin a story.
• They don’t resolve a situation.
• They create situations.

I realized recently that these are not quite satisfying, so I decided to come up with a different means of determining the good/bad-ness of a coincidence.

These are related to the Butterfly Effect and inference.

Basically, a coincidence is fine if the order of events leading to the coincidence can be rebuilt with the benefit of hindsight.

This doesn’t necessarily require that the writer explicitly explains the order of events, but it does require that the writer keep in mind these things.

For example, a good coincidence would be one where two characters meet by chance in the same location because both characters, independently, needed to go to that location. An example of such ‘in the wild’ may be the remnants of the Fellowship of the Ring meeting Gandalf the White.

A bad coincidence is one in which one (or both) characters have no particular reason to be meeting each-other.

Regarding the beginning of stories, read Beginnings.

Freak coincidences are generally not okay either.