Act Three Is Earned in Act One

When your ending doesn’t work, the fix is usually not a new plot. It’s an earlier decision. The cleanest demonstration ever filmed is Run Lola Run.

Lola has twenty minutes. Her boyfriend Manni has lost a bag of money, and if she doesn’t replace it, he will rob a store and probably die. The film tells those twenty minutes three times, start to finish. And the divergences between the three runs are almost invisible: a stumble on the apartment stairs, a half-second difference leaving the building, whether Lola brushes past a stranger on the street. Each tiny difference cascades. One run ends with Manni dead. One ends with Lola dead. The third finds the money, saves everyone, and lands the happy ending.

Same setup. One variable changed per telling. It is the cleanest experiment in cause and effect the movies have ever run, and the prose version works exactly the same way.

The bystanders make the point sharper. In each run Lola passes the same woman with a baby carriage, and the film flashes forward through that stranger’s whole life: a lottery win in one timeline, grief in another, quiet grace in the third. Every collision re-prices a life Lola never meant to touch. Your protagonist steers more than she knows.

Three mechanics.

Run the experiment on your manuscript. Write the same setup twice, change exactly one early decision, and see which ending each version earns.

Keep the divergence small enough to believe and big enough to matter. A stumble works. A lottery ticket doesn’t.

Compress the spine. Lola’s clock has no slack to absorb it. Strip your timeline down and every choice shows its consequence.

Try it: take the chapter where you are stuck. Rewrite only its first decision, three ways. Keep the version whose ending arrives soonest.