Estimate a branching story from its scene map and state-dependent variants rather than only its target word count. Separate unique passages from shared scenes, include the context each drafting request needs, and allow for route review and rejected branches. The number of possible reading paths is not the same as the amount of distinct text you need to generate.
Inventory scenes before counting complete paths
Create a map of the planned scenes, choices, and endings. Identify where branches separate and where they return to shared material. This reveals which passages need independent drafts and which can legitimately serve more than one route.
Do not multiply the length of one complete route by the total number of possible routes. That can count shared scenes repeatedly. Equally, do not assume one draft per scene is enough if the scene must change according to earlier decisions. The estimate needs to represent the actual text structure.
Identify state-dependent variants
List the facts that can change a later passage, such as what a character knows, which object the reader acquired, or which promise was made. Decide whether each difference requires alternate wording, a separate scene, or a deliberate redesign that avoids unnecessary variation.
Record only dependencies that the story will actually use. Adding many theoretical variables can make the plan look more sophisticated while creating drafting and review work with little reader benefit. A clear narrative purpose helps you distinguish meaningful branches from complexity introduced by the planning process itself.
Estimate the context required for each request
A useful drafting brief may include the scene's purpose, incoming state, permitted revelations, tone, and intended destinations. That context is part of the generation workload even when the requested passage is short. Keep it bounded to what the scene needs without removing information essential to continuity.
Use a representative set of scenes to observe the request size and revision pattern. Include a shared scene with several incoming routes, a state-sensitive scene, and an ending. An estimate based only on the opening may miss the work concentrated where branches reconnect.
Include alternatives that will not be retained
Track rejected drafts and branches during the pilot. Some may fail because they reveal information too early, create an unsupported object, or lead to an ending that does not match the reader's choices. Those attempts still consume generation and review effort even though they add no words to the final book.
Record the reason for rejection so the next request improves the brief rather than repeating the same problem. If most failures come from unclear state instructions, repair the scene record before increasing the number of attempts. More output is not a substitute for a coherent route specification.
Estimate review by transitions and dependencies
Plan checks for each choice destination and for combinations where earlier decisions affect later validity. Reading several enjoyable routes is useful, but it does not establish that every transition and state condition works. Include the time needed to locate and repair inconsistencies discovered through that broader review.
Separate mechanical checks from narrative judgment. A destination can exist while the arrival makes no sense for the reader's current knowledge. Keep both kinds of review visible in the estimate so a technically connected graph is not mistaken for a fully tested story.
Expand from observed production work
Summarize the pilot by unique scenes, variants, request context, accepted drafts, revision rounds, and review effort. Apply those observations to comparable parts of the remaining map. Mark unusually complex intersections as separate estimates rather than hiding them inside a single average cost per thousand words.
Revisit the estimate when the route map changes. Removing a branch may simplify several later scenes, while adding one choice can create new state combinations that require review. Preserve the link between the current map and the work estimate so the project does not continue using figures based on an abandoned story structure.
Frequently asked questions
Clear answers for this publishing decision.
Does twice as many endings mean twice as much drafting?
Not necessarily. Endings may share earlier scenes or require different amounts of unique material. Estimate the actual scene and state structure instead of using the ending count alone.
Should I draft every possible route separately?
That can duplicate shared passages and create conflicting versions. Identify reusable scenes and genuine variants, then plan their relationships explicitly.
Can I estimate from the final word count?
Use word count as one output measure, but include request context, discarded drafts, state variants, and review work. Two stories of the same length can require very different production effort.
What should trigger another estimate?
Revisit it when choices, shared scenes, state conditions, or endings change materially. Those changes can affect drafting and verification beyond the passage where the edit occurs.
What does this book cost to print and produce, and what should the list price be?
Printing cost follows from the trim, ink, and page count, the list price sets the royalty after that cost, and the production budget should be estimated before generation starts, so compare these numbers before you commit to a format or a plan.
See what a book project costs in credits