Reduce the training goal to one observable bottleneck
“Become more skilled” or “improve the craft” cannot easily become a scene. Select a task with an observable result: complete a shadow-puppet transition within six lines without exposing an empty stage, repair a folding fan with three scraps of old material, or have a trio change meter on the same phrase.
Use this training-goal card:
| Field | What this scene establishes |
|---|---|
| Future use | The real task in which this ability will matter |
| Current bottleneck | Execution, judgment, resources, or coordination |
| Success signal | A result someone present can observe |
| Failure signal | The deviation that shows the problem remains |
| Adjustable variable | Posture, order, quantity, tempo, or division of work |
| Scene boundary | Ability, rank, or result not yet earned |
A precise goal makes feedback useful to the next attempt. A training chapter may carry low pressure and no reversal, but it should still change ability, resources, information, or a relationship.
Use a training-round table for attempt and adjustment
Every training round should give the character information usable in the next one. Failure can be local: a scene change may produce the new weather yet expose an empty stage for a moment, showing that the materials work while the handoff timing does not. A mentor's or partner's feedback should point to an observed result instead of merely saying “try harder.”
| Round | Action, feedback, judgment, and next change |
|---|---|
| Baseline | Action: Changes three scenery cards in the old order; feedback: rain appears, but the stage goes blank during the swap; judgment: cards work, but the handoff is late; next change: move the eave card on the spoken cue. |
| Retest | Action: Pushes the eave card at the cue; feedback: no blank stage, but the rain card covers the figure's shadow; judgment: timing improves, but the rain card sits too low; next change: keep the timing and change only card height. |
| Check | Action: Keeps the cue and raises the rain card; feedback: the change is continuous and the figure remains visible; judgment: this round's goal is met; next step: test the audience view in full rehearsal. |
A scene need not contain three rounds. One failure followed by one evidence-based adjustment can deliver a complete change. When the training experience is itself a major pleasure, expand more rounds, but give each a different result.
Complete original scene: Change a paper-theatre set in six lines
The following is an author-designed fictional paper-shadow rehearsal. Its shadow box, scenery cards, and rehearsal rules belong only to this example; it is not a real performance manual. Apprentice Nian Qi must change “sunny wharf” into “rainy eave at night” within six spoken lines. Actor Mei Su moves the character silhouette and speaks the lines. In the previous rehearsal, Nian asked Mei to slow down without warning, and Mei felt blamed for the failed change. Today they agree on a fixed cue; neither the card sequence nor the delivery speed will be improvised to rescue the other.
The instructor pulled the viewing cloth across the far side of the box. “One task today. By the sixth line, the rain set is established and the figure has never disappeared.”
On Mei's first line, Nian removed the sunny card. Halfway through the third, she reached for the eave and rain cards together. They caught in the slot. Light passed through the empty paper frame, turning the viewing cloth blank for an instant. The rain finally appeared, but the figure seemed to fall into another play.
Mei finished all six lines before speaking. “I can't see your hands from here. Did the blank happen before or after ‘the boat has docked’?”
From the viewing side, the instructor answered, “Half a line after. The sunny card was gone before the eave arrived.”
Nian arranged the three cards on the table in the order she had used. None was wrong; she had waited until the third line ended before starting. “Next time, I push the eave at ‘the boat has docked.’ The rain enters when your figure reaches the doorway.”
“Original pace?” Mei asked.
“Original pace. Same cue.”
On the second attempt, the eave entered with the cue and the cloth never went blank. Nian waited for the figure to reach the doorway, then pushed in the rain. The rain lines sat too low, covering the figure's head and shoulders.
Mei held the figure in place. “The timing connects. The rain card is covering me now.”
Nian did not reorder everything. She put a small paper mark on the rain handle one finger above the previous position, keeping the eave and cue timing unchanged.
On the third attempt, the eave entered at “the boat has docked.” The figure reached the doorway; rain appeared behind its shoulders without covering its face. At the sixth line, the sunny card was stored while rain and figure remained on the cloth.
The instructor returned from the audience side. “This round is complete. Full rehearsal still has the complete lamp positions and audience view. Keep today's order.”
Nian wrote the three-card order inside the box lid and copied the cue for Mei. “Last time I asked you to slow down because I hadn't separated the change.”
Mei took the paper and underlined “the boat has docked.” “If the cue changes, we change it together. Don't make me guess during the run.”
They walked through the third sequence once more without lighting the box or adding lines. Nian called each action; Mei answered at the fixed cue. They confirmed division of work and order. The public performance had not happened.
The scene delivers three kinds of progress: the blank stage is located, the rain-card height is adjusted separately, and the partners establish how cues will change. Feedback has changed both technique and working relationship. Public performance remains untested.
Repair a training log by restoring the feedback chain
| Common problem | Direct repair |
|---|---|
| “After seven days, she finally succeeded” | Select one decisive failure and show the judgment and next adjustment it creates |
| Every round adds only fatigue or pain | Locate the error in position, timing, quantity, judgment, or coordination |
| The mentor supplies the answer immediately | Let the character observe the result first, then use the mentor to identify an omission or propose a testable change |
| One adjustment produces total mastery | Deliver a bounded success and retain conditions not yet tested in the real task |
| Nothing changes after training | Record the new method, spent resource, changed division of work, or next check |
To compress several days, keep three points: baseline failure, decisive adjustment, and verification result. The days between may be summarized, but each retained point should add different information. If training offers no new judgment, relationship experience, or progress, merge it into a more consequential scene.
Copyable training-scene brief
Future use:
One bottleneck for this round:
Observable success and failure signals:
Available attempts, time, or material:
Baseline attempt and actual feedback:
The character's limited judgment:
One variable changed next:
Retest result:
Progress actually earned in this scene:
Ability, permission, or real-world condition not yet tested:
Return the training-round table before drafting a complete candidate. Every result must change the next attempt. Do not replace process with elapsed days or invent a mentor secret, awarded rank, or new crisis.
FAQ
Must a training scene contain many failures?
No. One failure that locates a problem and one evidence-based retest can complete the change. The number of attempts follows the book's pace and intended experience, not a quota.
Can training be summarized as a montage?
Yes. Preserve the baseline, decisive adjustment, and verification result while summarizing repetitive rounds. Expand one key round when the summary no longer shows why the method changed.
Do it with the skill
Use $story-long-write to design this training scene only. List the future use, one bottleneck, success and failure signals, and available resources. Then plan a baseline attempt, actual feedback, one adjustment, and a retest. Preserve unearned permission and untested real conditions. Review the training-round table before drafting prose.
Read the method: Bottlenecks, trials, and progress in cultivation scenes · Training goals, attempts, and visible feedback · Function and intensity of a trial-and-error chapter