skills/ai-film-pipeline-master/references/phase-01-story-generation/symptoms-opening.md

# Opening complaints

Part of the symptom router. Open this file when the complaint is about opening. Each complaint below is followed by the cards that repair it, best first; `LOOKUP.md` says which file holds each card.

---

### "Nobody gets past the first page"

The first ten pages of a screenplay and the first two of a novel are the entire audition. Pick a hook type deliberately rather than hoping the opening is interesting.

1. Lead with a specific person rather than a situation — readers hook on character more often than on event
   `open.character_hook` · `open.voice_hook`
2. Or drop them inside an event already running, but make *who wants what right now* answerable within a paragraph
   `open.situation_hook`
3. Or plant one specific unexplained thing and refuse to explain it
   `open.question_planted_first` · `open.object_opening`
4. Check you are not committing one of the four named failures
   `open.weather_open` · `open.overload_open` · `open.backstory_open` · `open.false_start`

### "I don't know how to start"

1. Decide what the opening promises before deciding what happens in it
   `open.contract_opening`
2. Work backwards from the last image and open on its opposite
   `open.opening_image_answered`
3. For an invented or historical world, demonstrate one rule rather than explaining the system
   `open.rule_of_the_world`
4. For a story about something the characters have normalised, run the routine with one element wrong
   `open.ordinary_day_slightly_wrong`

### "The opening is slow"

1. Cut the first three lines of every scene in the draft, not just the first one
   `open.scene_opening` · `scene.late_entry`
2. Move the title card to land after the first real turn
   `open.title_delayed`
3. Start on a detail and widen, so the first information is deliberately partial
   `open.scale_reveal`
4. If it is slow because the history comes first, the fix is redistribution
   `open.backstory_open` · `scene.absorb_and_delete`

### "The opening is exciting and the rest isn't"

1. You have written a contract you cannot keep — whatever is in the first three minutes, the audience expects for the whole runtime
   `open.contract_opening`
2. If the big threat came before any attachment, shrink it to one household
   `open.unearned_stakes`
3. If the excitement turns out not to have happened, that is the most recognisable amateur signal there is
   `open.false_start`

### "How does this genre open?"

`open.horror_cold_open` · `open.the_meet` · `open.crew_assembly` · `open.procedural_body` · `open.war_arrival` · `open.comedy_premise_statement` · `open.mystery_discovery` · `open.last_ordinary_day`

### "The first line isn't working"

1. Know what a first line is actually for before writing one
   `open.first_line_job`
2. Choose the move: declarative, a question, or starting mid-sentence
   `open.declarative_first_line` · `open.opening_question_line` · `open.mid_sentence_open`
3. Decide whether it promises a voice or a plot — it cannot do both
   `open.voice_vs_plot_first_line`
4. And check it is not overwritten, which is the commonest first-line failure
   `open.overwritten_first_line`

### "The first paragraph is doing the wrong things"

1. There is a working order to the information
   `open.first_paragraph_contents` · `open.information_order`
2. Orientation and intrigue compete; decide which one this piece needs first
   `open.orientation_vs_intrigue`

### "Do I need a prologue?"

1. Usually no. A prologue is earned only under specific conditions
   `open.prologue_earned` · `open.prologue_disguised_chapter`
2. The version that works is in a different time or register, or frames the whole
   `open.prologue_other_register` · `open.framing_prologue`

### "The title isn't right"

1. The strongest titles mean something different by the end
   `open.title_changes_meaning`
2. The logline is a diagnostic before it is a sales tool — a broken one usually means a broken story
   `open.logline_diagnostic` · `open.broken_logline_reading`
3. A title that promises the wrong genre costs more than a dull one
   `open.title_genre_mismatch`

### "The packaging is letting the work down"

1. The thumbnail and the title form one contract, and the piece must pay it
   `open.title_thumbnail_contract` · `open.visual_promise`
2. Underselling is as damaging as overselling
   `open.package_undersell`

### "Opening a pitch, a documentary, an essay or an explainer"

`open.pitch_first_page` · `open.documentary_open` · `open.essay_first_move` · `open.explainer_promise`

### "After the opening, I don't know how to say why this matters"

The lede bought their attention with one person. The paragraph after it converts that attention into a reason, and a reader who has been given a reason will tolerate a much longer middle than one who is still being charmed.

1. Zoom from the one to the many — the lede is a person, this is the class that person belongs to — state the stake, attach the timeliness peg, and carry exactly one number or one dated fact. One. Two and it has become the body
   `open.nut_graf`
2. Position is paragraph three to five after an anecdotal lede. What changes if it comes later is risk rather than style: past about the fifth paragraph the reader is running purely on the anecdote's own charm, so make that bet consciously — and the later it lands, the more it must read as the writer pausing to reflect rather than announcing
3. **The rule that separates it from everything near it: it answers "so what" without answering "what happens" — it sells the piece without spending it.** A bad one is an itinerary in prose: *this story will examine, we will look at, as we shall see*. A good one names what is at risk and for whom, stated as a claim the piece has not yet proved
4. Two tests. Could you paste this paragraph into the email to your editor and have them understand the piece? And delete it — does the piece still have a visible reason to exist?
5. The version beyond print is the line after the hook, ten to twenty-five seconds into a documentary or an explainer, and it is the same problem the newsroom solved in the 1940s. *In this video I'll cover three things* is the summarising failure in narration, and it is where the retention curve breaks. Say the stake instead
6. Do not dramatise it and do not put it in someone's mouth. The direct address is the point, and this is the one paragraph where the instinct to absorb everything into the argument is wrong
   `expo.absorb_into_argument` · `pov.reportage_stance` · `theme.descriptive_not_prescriptive`
7. The neighbours it gets confused with: the lede itself, which only notes that a turn exists; the deliverable named aloud at the very top; the reader's deficit named aloud at the very top; and the scroll-stop, which is not a reason at all. None of them sit after a scene
   `open.essay_first_move` · `open.stated_promise` · `open.explainer_promise` · `open.first_second_hook`
8. Never in fiction. The fiction-side answer to the same need is a planted question
   `open.question_planted_first`
9. Two failures to watch for: one that promises a different article from the one that follows, and one that has quietly become the conclusion and arrived three thousand words early
   `open.wrong_promise` · `theme.explained_theme`
10. And in a single-character profile or a pure narrative feature a full one lands as a spoiler. There the move is to shrink it to one sentence of promise and let it stay ambiguous, not to write it bigger
11. The failure nobody names: a writer who cannot write it usually does not have one, and what is missing is the reporting. The industry response is to write a vaguer paragraph, which produces the piece that sounds authoritative and says nothing. Go back out instead

### "Opening an interactive piece"

1. The tutorial should be a scene, not an instruction sheet
   `open.tutorial_as_scene`
2. The first choice teaches the audience what choices mean here
   `open.first_choice` · `inter.meaningful_choice_test`

### "The opening promises the wrong piece"

1. The named failure — the most damaging one an opening can commit
   `open.wrong_promise` · `open.contract_opening`
2. Front-loading a character who is not the protagonist
   `open.wrong_character_first`
3. A device used once and never again
   `open.abandoned_device` · `open.false_start`

### "Opening a subtitled or translated piece"

`open.subtitled_open` · `rhythm.subtitle_beat`

### "The reader put it down at chapter two"

Chapter breaks are where readers actually stop, not first pages.

1. Re-hook at the section start instead of resuming
   `open.chapter_opening`
2. Open the chapter later in time than expected, or on a line that revises what they believed at the break
   `open.chapter_opening` · `scene.late_entry`

### "The pilot doesn't feel like a series"

1. A pilot has two incompatible jobs — close the episode's question and open the series' question in the same beat
   `open.series_opening`
2. If it only sets up, it is not a pilot; if it only concludes, there is no series
   `open.series_opening`

### "Each episode wastes time reminding people what happened"

1. Reorient through action, not recap — someone doing something that only makes sense in light of last week
   `open.episode_opening`
2. In short-form, compress the reorientation to one accusation
   `dial.exposition_accusation` · `open.open_mid_situation`

### "The narration is doing work the scenes should do"

1. The narrating voice must know something the scene does not, or be wrong about something the scene shows
   `open.voiceover_open`
2. If it only describes what is visible, it is redundancy with music under it — cut it
3. The version with a gap in it: write the narration and the picture as two separate documents, lay them against each other, and cut every line that agrees with what is on screen at that second. Never let the narrator be wry about themselves, because that closes the gap
   `sub.voiceover_counterpoint` · `voice.retrospective_narrator`
4. In documentary this is not a craft move — a narration track contradicting the footage is an editorial accusation
   `expo.doc_narration_contract`

### "The episode needs to open cold"

1. A cold open should be complete in itself, with its own small turn, so the audience feels already paid
   `open.cold_open`
2. In comedy it is conventionally unrelated to the plot, and that unrelatedness is the pleasure
   `open.cold_open`
3. Warning sign: if the cold open is stronger than everything after it, that is a structural problem, not a success

---