Design the Long-Term Tension and Scene Conflict Separately
This page focuses on adversarial conflict in which both sides learn and act. A natural environment can keep applying pressure, and an inner contradiction can constrain a decision; neither needs to masquerade as a person taking tactical action.
Long-term tension answers, “Why will these sides keep colliding?” Scene conflict answers, “What concrete outcome are they fighting over today?” For example, one side wants to preserve an artifact in its original state while the other wants to make it safe for use. That is the long-term tension. Whether to remove its original casing today is a scene conflict that can reach a result.
A protagonist’s goal and obstacle alone do not yet show how an opponent acts. Give each side four things: the outcome it wants, the leverage it holds, the line it cannot publicly abandon, and what it learned from the previous round.
Advance Through Countermoves, Not One-Sided Difficulty
| Round | Move and change |
|---|---|
| Probe | Move and reply: Pursue the result with a low-cost method; reply: Reveal a constraint or condition; Change: Information and positions |
| Raise | Move and reply: Commit more important leverage; reply: Change the rules, alliance, or interpretation of evidence; Change: Resources and available options |
| Lock-in | Move and reply: Force the other side to choose publicly; reply: Accept a loss or expose the other side in return; Change: Relationships and escape routes |
| Decision | Move and reply: Act among the remaining options; reply: Give a final response; Change: An irreversible outcome |
A story does not need exactly four rounds. The table only checks whether both sides learn and whether each exchange leaves something changed. A conflict that can finish in two rounds should not be stretched.
This editorial worksheet is not the skill’s fixed four-round output. Limited agreement can change review access without an irreversible loss in every round.
Original Example: The Broadcast-Tower Restoration Bid
An old-city broadcasting association must choose a team to restore a mountaintop transmitter that has been out of service for years. The old unit is in the workshop for initial offline checks, with the tuner still inside it. Engineer Liang Zhou wants to keep the original tuner and prove that the old equipment can still work. Safety supervisor He Min insists on removing it first so that an uninspected component cannot enter the test broadcast. Reinstalling the assembled unit at the tower now conflicts with first removing the tuner for further inspection; their long-term aim is the same restored service. Here Liang owns the engineering proposal, He Min approves installation and testing, the association schedules events, and the sponsor controls funding rather than test permission.
Probe: Liang Zhou submits a bench test showing a stable short signal during the unit’s initial check. He Min accepts that result. She points out that the test did not reproduce the damp mountaintop environment and approves only further offline inspection. Resolution: Liang Zhou wins permission to test, not permission to install.
Raise: Liang Zhou finds an old maintenance log and proposes sealing the unit according to the original structure. He Min discovers that the log’s final pages are missing and instead requires the team to design a replacement option first. Resolution: the historical value remains, but the original unit is no longer the sole option.
Changed conditions: The sponsor wants a commemorative broadcast and makes further funding conditional on reviving the original unit. Liang cannot use funding demands to bypass He Min. Liang discloses the missing log pages to the association and proposes an offline bench comparison of original and replacement units, without installation. He Min approves supervised, low-power offline testing; the association keeps the anniversary plan provisional, with equipment selection following checks. The original earns a comparison opportunity, not permission to join the mountaintop transmitter. The sponsor’s response to a possible delay remains unknown.
The conflict before the climax never turns into mere shouting. Every round changes authority, evidence, and options, and each side adapts to the last result. A later scene must still resolve whether the final test succeeds; this article does not invent an executed result.
Dramatize one round: earning a comparison opportunity
This expands the changed-conditions round from the transmitter example. The complete candidate establishes limited test permission and scheduling only. Availability of replacement components and test success remain for later scenes.
Liang Zhou handed He Min the sponsor’s demand. “They want the anniversary broadcast. Funding still depends on reviving the original unit.”
She left his installation request untouched and opened the maintenance log at its missing final pages.
“Money doesn’t replace those pages,” she said. “Do you still want to put it straight back on the mountain?”
Liang withdrew the request. “No installation. I want to compare the original with the replacement. The original still matters. That doesn’t let it skip inspection.”
“Where will you compare them?”
“On an offline bench. We prepare the replacement option and required components, then compare.”
He Min wrote supervised, low-power and offline on the new application, crossing out installation. “This scope is acceptable. Until the mountain conditions have been checked, a bench result cannot authorize installation there.”
Liang read the conditions and left the application for her signature. He did not slip the sponsor’s demand beneath it.
The association’s scheduler had been listening. Liang turned to him. “The log has gaps. We need the comparison. We can’t choose the anniversary equipment today.”
The scheduler kept the anniversary on the calendar and marked the equipment undecided. “Keep the date provisionally; select equipment after checks. You still have to discuss any delay with the sponsor.”
“I’ll tell them what we can do now,” Liang said.
He put the signed test application into his folder separately from the installation request. The original had earned a comparison opportunity. The mountain-installation line still had no signature.
Liang has not won his initial installation demand. He changes his proposal and earns continued evaluation. He Min accepts the new scope while retaining her limits. The association preserves a provisional date without accepting a delay for the sponsor. Permission is neither an executed test nor restored broadcasting.
The round can end here, with specific progress and a remaining question. It need not immediately add an accident or larger loss. A later test scene must establish its materials and authorized scope before characters choose from its results.
How to Escalate Without Repeating Yourself
Escalation can change four dimensions:
- Leverage: Move from a verbal opinion to reputation, resources, or a promise.
- Public scope: Move from private negotiation to a choice that forces the team to take sides.
- Reversibility: Move from a retractable proposal to an action that permanently changes a relationship or object.
- Time and exits: Let the last round’s result invalidate an old compromise.
Change one dimension when escalation is needed, without enlarging loss every round. A harsher line answered with identical evidence may be repetitive. Tone can still express a relationship, but it does not itself change actual permission.
Common Conflict Failures and Fixes
| Failure | Fix |
|---|---|
| The opponent only says no | Give the opponent an independent goal, leverage, and acceptable conditions |
| The protagonist repeats the same approach every round | Let the character adjust using the previous result; an old method can remain useful under newly established conditions |
| A new accident suddenly forces escalation | Derive escalation from established leverage, promises, and relationships first |
| The argument is fierce but the scene resets | Change authority, resources, information, or a relationship |
| One side wins and the loser seems to learn nothing | Show how the losing side adjusts its evidence, alliances, or rules |
FAQ
Do different character goals automatically create conflict?
For adversarial conflict, identify where the desired outcomes cannot both be met under current conditions and how each side will act on the result. Different opinions alone may remain a discussion. For internal or environmental conflict, instead examine how pressure makes a choice costly; there is no need to force in a second actor.
Must conflict become increasingly dangerous?
No. A confrontation can shift toward reputation, relationships, resources or authority, or conclude through negotiation. Identify this round’s specific progress, response or new condition. Adjust when an old solution stops working; more danger or greater loss is not required in every conflict scene.
Do it with the skill
In Claude Code, say: “Use /story-long-write only to design the conflict ladder for this confrontation. For both sides, list incompatible outcomes, leverage, red lines, and what they can learn. Then map each move, countermove, and the leverage or exit it changes. Do not write prose.” In Codex, use $story-long-write.
Read the method: Contradiction, scene conflict and drives · Stage payoffs and optional action costs · Author choices and established facts