What Makes a Chapter Ending Actually Work

There is a difference between a chapter that ends and a chapter that stops. Both put a break on the page. Only one makes the next chapter feel necessary.

Four kinds, in ascending order of how long they keep working.

The withheld fact. A character learns something and the chapter closes before the reader does. It works exactly once per reader per story. Used twice, the reader stops trusting the narration; used three times, they start skimming to the reveal. It is the cheapest ending available and it is priced accordingly.

The interruption. Someone walks in, the phone rings, the door opens. This is a scene-management tool wearing an ending's coat. It creates momentum without creating a question, which means the momentum drains between sessions.

The changed position. A character ends the chapter wanting something different from what they wanted at the start. No information is withheld. The pull comes from wanting to see what the new want costs. This survives repetition because it is not a trick — it is the plot.

The stated stake. A character says out loud what they are willing to lose. The reader now knows the shape of the next chapter and reads on to find out whether the character meant it. Underused, and the most durable of the four.

The test is simple: if you summarized the chapter's ending in one sentence to someone who had not read it, would they ask a question? A withheld fact produces "what was it?" — curiosity, which decays. A changed position produces "what will they do?" — investment, which does not.