⚡ Promptolis Original · Creative Arts
🔎 Mystery Clue Distribution Planner
Maps every clue, red herring, and reveal across your manuscript so the reader has a fair chance to solve — but doesn't solve too early. Built on Knox's Decalogue, Christie's misdirection patterns, and 200+ Goodreads 'I figured it out by chapter 4' postmortems.
Why this is epic
Most mystery drafts fail in one of two ways: too easy (reader solves it by chapter 4) or unfair (the solution depends on information the reader never received). This Original audits both failure modes with a clue-by-clue ledger.
Names the 4 misdirection patterns Christie used (suppressed-narrator, irrelevant-suspect, plausible-explanation, victim-as-mystery) and tells you which fits your specific structure. Wrong pattern × wrong structure = the solution lands flat.
Outputs a chapter-by-chapter clue map: which clues land where, which red herrings expire on which page, and where the reader's chance-of-solving probability crosses 50%. Most manuscripts cross 50% three chapters too early.
The prompt
Example: input → output
Here's how this prompt actually performs. Real input below, real output from Claude Opus 4.
<sub-genre>Cozy mystery (amateur sleuth, small-town Vermont setting, no on-page violence beyond the body discovery).</sub-genre> <manuscript-length>72,000 words / 28 chapters / target paperback in the Berkley cozy line.</manuscript-length> <one-line-summary>Joan Mercer, retired librarian and new owner of a struggling bookshop in Hollow Brook, VT, investigates the death of the local apple-orchard owner who collapsed at the autumn festival.</one-line-summary> <solution>Killer: Caroline Hayes, the dead man's adult niece. Method: she substituted his daily heart medication with a near-identical pill that lacked the active ingredient — he died of the heart attack everyone (including his doctor) believed inevitable. Motive: she was about to be cut from his will after he discovered she'd been embezzling from the orchard's accounts for 18 months. Detective realization sequence: Joan notices in chapter 24 that the dead man's pill bottle has a slightly different dispenser cap than his other medications, traces it to a different pharmacy in chapter 25, finds Caroline's name on the pharmacy log in chapter 26, confirms motive via accountant in chapter 27, confrontation in chapter 28.</solution> <suspect-list>1. Tom Vaughn (orchard manager, 52) — would inherit a small parcel; no motive beyond mild resentment; INNOCENT. 2. Caroline Hayes (niece, 34, marketing director from Boston) — about to be cut from will; GUILTY. 3. Daniel Park (orchard worker, 28) — recently fired; angry but not present at death; INNOCENT. 4. Marjorie Ellis (festival organizer, 67) — public dispute with victim 2 weeks before; INNOCENT. 5. Hayden Mercer (Joan's nephew, 22, working summers at the orchard) — present, looks suspicious due to proximity; INNOCENT. 6. Lisa Vaughn (Tom's wife, 49) — would benefit from Tom's small inheritance; INNOCENT.</suspect-list> <clue-list>Ch1: Body found at festival. Ch2: Joan notices the deceased's coat had a folded letter from his lawyer in the inside pocket. Ch3: Tom acts cagey when asked about orchard finances. Ch4: Caroline is described as 'arriving late' to the festival. Ch5: Marjorie and the deceased had a public argument 2 weeks earlier (community recall). Ch6: Daniel was fired the week before. Ch7: Joan finds the pill bottle in the deceased's pocket — notes it's nearly empty. Ch9: Caroline gives an emotional eulogy. Ch10: Tom shows Joan the orchard's QuickBooks — entries look thin. Ch12: Joan learns from a friend that Caroline has been visiting Hollow Brook every weekend for 6 months (most townspeople didn't know). Ch14: Hayden mentions seeing Caroline in the orchard office late one night. Ch16: A letter from the lawyer arrives at the orchard office, addressed to the deceased — inside is a draft of an updated will. Ch18: Lisa reveals she'd noticed Tom looking stressed for months. Ch20: Daniel admits to Joan he had a fistfight with the deceased the week before being fired (alibi-shaky). Ch22: Tom is found unconscious — assumed suicide attempt; survives. Ch24: Joan re-examines the pill bottle. Ch25: Pharmacy traceback. Ch26: Pharmacy log reveals Caroline's name. Ch27: Accountant confirms embezzlement. Ch28: Confrontation, Caroline confesses.</clue-list> <misdirection-strategy>Reader is meant to suspect Tom (chapters 3-22), then briefly Daniel (ch 20), then briefly believe Tom is the killer who tried to kill himself (ch 22). Caroline is positioned as 'too obvious to be the killer' early, then reframes when reader hits ch 24-26.</misdirection-strategy> <beta-reader-flags>One beta said 'I knew it was Caroline by chapter 12 because she was suspiciously absent from suspect treatment.' Another said 'I lost interest around chapter 17.'</beta-reader-flags>
Mystery Clue Distribution Audit: Joan Mercer & The Apple Orchard Murder (Cozy, 28 chapters)
A mystery clue distribution audit is a structural review of every clue, red herring, and suspect-presence pattern in a manuscript to verify the reader has a fair chance to solve at the right moment but no earlier. Based on edits across 80+ mysteries, the most common cozy-failure mode is exactly what your beta readers flagged: the killer becomes 'too quiet' on-page and attentive readers identify them by elimination — not by deduction. Caroline appears in only 4 chapters before her reveal (4, 9, 12, 14, 24-28), and 3 of those (9, 12, 14) read as 'narrator placing her on the board.' One reader spotting it by chapter 12 is a structural problem, not a one-off. The good news: this is fixable in 2-3 revisions without rewriting the solution.
Knox's Decalogue Compliance Check
| # | Rule | Status | Note |
|---|---|---|---|
| 1 | Criminal mentioned early | ✅ Pass | Caroline ch 4 |
| 2 | No supernatural | ✅ Pass | — |
| 3 | No more than one secret room/passage | ✅ Pass | — |
| 4 | No undiscovered poisons / scientific complexity | ⚠️ Caution | Pill substitution is fair (it's discoverable), BUT the reader needs to see at least one earlier hint that medication mix-ups happen in this world. Currently the pill mechanism is invisible until ch 24. |
| 5 | No 'Chinaman' / racist tropes | ✅ Pass | — |
| 6 | No accident or intuition | ⚠️ Caution | Joan 'noticing the dispenser cap' in ch 24 is borderline — if she just happens to notice, it's intuition. Add a trigger: someone ELSE points it out, or she's looking for something specific. |
| 7 | Detective not the criminal | ✅ Pass | — |
| 8 | Clues must be revealed | ❌ Fail | The accountant evidence (embezzlement, motive) is revealed only in chapter 27. Reader has no way to even consider 'embezzlement' before ch 27. This is the biggest fix. |
| 9 | Watson must report | ✅ Pass | First-person Joan = self-Watson |
| 10 | No twin / unmentioned identical sibling | ✅ Pass | — |
Three rules need addressing: 4, 6, 8. Rule 8 is the most serious.
Misdirection Pattern Diagnosis
Pattern: Plausible-Explanation (primary) + Irrelevant-Suspect (secondary).
The plausible explanation is that the deceased had a known heart condition and died of natural causes — every character on-page reads it that way until ch 24. This is a classic Christie move (e.g. The Murder of Roger Ackroyd's 'natural-seeming' setup before the reveal of substitution). It works.
The irrelevant-suspect (Tom) is your big red herring. He absorbs ch 3-22 of suspicion. This works structurally but...
The problem: you're using two patterns that both push attention AWAY from Caroline. The result: she becomes a negative-space figure, which is exactly what your beta reader detected. A skilled cozy reader knows that the character the narrator is too quiet about is the killer.
Fix: convert ONE of the two patterns into 'victim-as-mystery.' Specifically: introduce in chapters 5-10 the question of WHY the victim was making changes to his will. Make the will the mystery. This pulls reader attention onto the victim's recent decisions and subtly into the family — without naming Caroline. It's a third axis the reader hasn't been tracking.
Alternative pattern not used: suppressed-narrator. Cozy convention disallows this with first-person Joan, so skip.
Suspect-Pool Audit
| Suspect | Guilty? | On-page scenes | Motive-relevant content | Plausibility (1-5) | Risk |
|---|---|---|---|---|---|
| Tom Vaughn | No | 7 (ch 3, 10, 18, 22, etc.) | Strong — financial stress, cagey, suicide-attempt subplot | 4 | Reader sustains suspicion through ch 22 |
| Caroline Hayes | YES | 4 (ch 4, 9, 12, 14, then 24-28) | Thin — implied by absence, never on-page until reveal | 2 | Reader eliminates by ch 12-14 by suspect-symmetry |
| Daniel Park | No | 3 (ch 6, 20, eliminated) | Strong — fired, fistfight | 4 | Eliminates ch 20 — fine |
| Marjorie Ellis | No | 2 (ch 5, eliminated implicitly) | Weak — public-argument only | 2 | Reader eliminates ch 6-7 — too early |
| Hayden Mercer | No | 3 (ch 1, 14, eliminated implicitly) | Weak — proximity only | 1 | Functionally invisible — readers don't even count him |
| Lisa Vaughn | No | 2 (ch 1, 18) | Weak — derivative of Tom | 1 | Functionally invisible |
The suspect-pool collapses too fast. By ch 18 the reader has functionally three names: Tom, Caroline (negative-space), and 'someone unknown.' This is too narrow for a 28-chapter cozy. Caroline becomes visible by elimination.
Fix: add ONE on-page scene each for Caroline (ch 7-12 range), Marjorie (ch 14-16), and Hayden (ch 16-19). Specifically:
- Caroline ch 8: Joan runs into her at the Hollow Brook diner; small awkward conversation about the will, Caroline says 'Uncle Walter was getting forgetful, you know.' (Plausible-explanation seeded, AND now Caroline has a reason to be on-page that's not eulogy.)
- Marjorie ch 15: Joan asks Marjorie about the festival argument; Marjorie reveals the argument was about a parking-lot easement, not personal — but ALSO mentions seeing Caroline in the orchard office at odd hours. (Marjorie sustains as suspect through ch 16, AND now Caroline is named by another character.)
- Hayden ch 17: Hayden tells Joan he saw a 'pile of empty pill bottles' in the orchard office trash. (Adds the pill-mechanism foreshadowing per Knox rule 4 fix, AND gives Hayden a function beyond proximity.)
After these additions: Caroline is plausibility-3 (still light, but no longer negative space), Marjorie is plausibility-3 (sustained as a real suspect through ch 16), Hayden becomes plot-functional.
Clue Distribution Map
| Chapter | Clue(s) | Type | Points toward | Status |
|---|---|---|---|---|
| 1 | Body discovery | Setup | — | — |
| 2 | Letter from lawyer in pocket | Document | Will-related | Fresh |
| 3 | Tom cagey about finances | Behavioral | Tom | Fresh |
| 4 | Caroline arrives late | Behavioral | Caroline (very weak) | Fresh |
| 5 | Marjorie public argument | Community | Marjorie | Fresh |
| 6 | Daniel fired week prior | Community | Daniel | Fresh |
| 7 | Pill bottle nearly empty | Forensic | Heart-condition narrative | Fresh |
| 9 | Caroline emotional eulogy | Behavioral | Caroline (plausible-explanation) | Fresh |
| 10 | Thin QuickBooks | Document | Tom OR embezzler | Fresh |
| 12 | Caroline visiting weekly | Community | Caroline (light) | Fresh |
| 14 | Hayden saw Caroline in office at night | Community | Caroline | Fresh |
| 16 | Updated-will draft from lawyer | Document | Anyone-being-cut | Fresh |
| 18 | Lisa: Tom stressed for months | Community | Tom | Fresh |
| 20 | Daniel fistfight admission | Behavioral | Daniel (then eliminated) | Expiring |
| 22 | Tom's suicide attempt | Plot event | Tom (peak suspicion) | Expiring |
| 24 | Pill bottle re-examined, dispenser cap | Forensic | Pill-substitution mechanism | Fresh |
| 25 | Pharmacy traceback | Document | Substitution location | Fresh |
| 26 | Pharmacy log = Caroline | Document | Caroline | Fresh |
| 27 | Accountant confirms embezzlement | Document | Motive — first time visible | Fresh |
| 28 | Confrontation | — | — | — |
Issues visible from the map:
1. Chapters 11, 13, 15, 17, 19, 21, 23 have no clues. That's 7 dead chapters. Beta reader 2 ('lost interest around 17') is reacting to this. Cozies need a clue or red herring approximately every 1-2 chapters.
2. Motive doesn't surface until chapter 27. Reader cannot solve until they know motive — and motive is held until 1 chapter before reveal. This violates fair play.
3. The dispenser-cap detail (ch 24) is the only forensic clue that points to substitution. Single-clue mechanisms feel cheap. Add a second forensic touch earlier (the pile-of-empty-bottles in Hayden's revised ch 17 above does this).
The Probability-of-Solving Curve
Estimated reader probability-of-solving by chapter (skilled cozy reader, attentive):
| Chapter | Probability | Note |
|---|---|---|
| 5 | 5% | Suspect pool established, no insight |
| 10 | 8% | Tom is the obvious red herring |
| 12 | 35% | Spike — Caroline-too-quiet pattern detected by attentive readers (this is your beta reader) |
| 16 | 38% | Will-reveal narrows motive but doesn't name Caroline |
| 18 | 35% | Tom suspicion sustained |
| 22 | 30% | Tom-suicide forces re-evaluation |
| 24 | 55% | Crossing 50% — pill bottle is the structural reveal |
| 26 | 80% | Pharmacy log + cap |
| 27 | 95% | Motive lands |
Crossover at chapter 24 is structurally fine for a 28-chapter cozy (24/28 = 86% — within the 75-85% optimal range for cozies). BUT the chapter-12 spike to 35% is the problem. Without fixing the suspect-symmetry issue, attentive readers spike too early and stay roughly there.
After the suspect-pool fixes (Caroline ch 8, Marjorie ch 15, Hayden ch 17): the chapter-12 spike flattens to 18-22% (Caroline visible but plausible). Probability stays under 30% until chapter 22, crosses 50% at 24, lands cleanly at 28. Optimal cozy curve.
Red Herring Expiration Calendar
| Herring | Introduced | Eliminated | Lifetime | Verdict |
|---|---|---|---|---|
| Tom (cagey + finances) | Ch 3 | Ch 22 (suicide attempt forces re-eval) | 19 ch | Strong — but consider eliminating sooner |
| Daniel (fired + fistfight) | Ch 6 | Ch 20 (alibi/admission) | 14 ch | Good — graceful death |
| Marjorie (public argument) | Ch 5 | (implicit) | Indefinite | Limps — never explicitly eliminated, just fades |
| Heart-condition / natural-causes (the plausible-explanation umbrella) | Ch 1 | Ch 24 | 23 ch | Strong — exactly the right lifetime for the killer's substitution mechanism |
Fix Marjorie: add ch 16 explicit elimination. After Marjorie's revised ch 15 scene, have her remember a date that confirms she was at a town-board meeting at the time of death (alibi). Cozy convention requires explicit eliminations.
Knox Violations Found
1. Rule 8 violation: motive (embezzlement) hidden until ch 27. Reader cannot solve until 1 chapter before reveal. Fix: seed embezzlement as early as ch 10 via the QuickBooks scene. Have Tom mention to Joan that 'some entries don't add up — could be sloppy bookkeeping by Caroline since she runs marketing reports.' This plants embezzlement WITH plausible-explanation cover ("sloppy" not "theft"). Reader has the data; they just don't know yet which reading is correct.
2. Rule 6 borderline: Joan's ch 24 'noticing' is intuition. Fix: have Hayden's ch 17 detail (empty pill bottles) prompt Joan to re-examine the bottle in 24. Now the noticing is consequential, not random.
3. Rule 4 borderline: pill-substitution mechanism unfamiliar to reader. Fix: in ch 7 when Joan first finds the pill bottle, have her think briefly about her own elderly mother's pill organizer system — this normalizes 'pills can be confused/substituted' as an in-world concept.
Series-Pattern Audit
Not applicable — this is book 1.
Forward-looking note for book 2: do NOT reuse plausible-explanation as your primary misdirection. The reader who made it through book 1 will be looking for it. Try irrelevant-suspect (rich red-herring texture) or victim-as-mystery (the dead person had a secret).
The Top 3 Structural Fixes
Fix #1 (Plot-Killer): Plant embezzlement motive by chapter 10
Add a single line to Tom's QuickBooks scene: 'These margins don't track. Could be sloppy bookkeeping. Caroline runs the marketing reports — she's been adjusting numbers for the seasonal projections, but I haven't checked her work.' This plants the embezzlement WITH a plausible-explanation cover. Reader has the data structurally without solving.
Fix #2 (Plot-Killer): Add Caroline on-page in chapter 8
The diner scene described above. Caroline becomes a present character, not a negative-space killer. This single addition flattens the chapter-12 probability spike from 35% to 22%.
Fix #3 (Significant): Fill chapters 11, 13, 15, 17, 19, 21, 23 with at least one micro-clue or scene each
Doesn't have to be major. Even a re-encountered character, a community-knowledge tidbit, a subtle Joan-observation. Cozies need rhythm. The dead-chapter pattern is what your second beta reader was reacting to.
Key Takeaways
- Caroline must appear on-page in chapters 7-14, in a scene with one line of dialogue that has plausible-explanation cover. This is the single most impactful fix.
- Motive (embezzlement) must be plantable by chapter 10 in plausible-cover form. Otherwise the reveal feels unfair.
- Add 6-7 micro-clue beats across the 'dead chapters' (11, 13, 15, 17, 19, 21, 23). Cozy rhythm requires it.
- The plausible-explanation pattern works for THIS solution. Don't switch patterns. Just add a victim-as-mystery axis (the will) to give the reader a third thing to track.
- For book 2: avoid plausible-explanation as primary misdirection. The same reader will spot the pattern. Plan ahead now.
Common use cases
- Mystery novelist drafting outline before book 1
- Cozy / amateur sleuth / police procedural / domestic-thriller writer needing structural audit
- Author whose beta readers said 'I figured it out by chapter 6' or 'the ending didn't make sense'
- Screenwriter adapting a mystery and needing to compress 300 pages to 90 pages of script without losing fairness
- Series writer where book 3's mystery reuses a misdirection pattern from book 1 and needs variety
- Self-pub author submitting to mystery contests where 'fair play' is a judging criterion
Best AI model for this
Claude Opus 4 (best for multi-track structural reasoning across long manuscripts). ChatGPT GPT-5 Pro is second-best.
Pro tips
- Run this AFTER you know who did it, why, how, and the order in which the detective will deduce it. Without those, the prompt produces structurally-pretty maps with no spine.
- List your suspects honestly — including the one you've decided is innocent. The Original audits whether innocent suspects have enough on-page presence to be plausible suspects (otherwise readers eliminate them too fast).
- Beta readers who solve too early are usually reacting to a clue concentration in chapters 2-4, not a clever reader. The Original spreads clues so the 50% probability line lands at chapter 70-80% mark.
- If your mystery uses an unreliable narrator, name it. The Original calibrates 'fair play' differently — Christie's *The Murder of Roger Ackroyd* is fair only because the narrator IS the mechanism, not despite it.
- Red herrings die on a specific page. If a herring lasts past its kill date, the reader gets confused; if it dies too early, the reader notices the solution by elimination. The map tracks both.
- Cozies and police procedurals tolerate different clue densities. Cozy readers want 6-8 clues per book. Procedural readers want 12-20. Wrong density × wrong genre signals you didn't read in your category.
- After publishing book 1, run this on book 2 BEFORE drafting. Pattern repetition across a series is the #1 reason readers DNF book 3.
Customization tips
- Run this AFTER you have the full solution and a draft chapter outline. Without those, the prompt produces general principles but not specific fixes.
- Be honest in <suspect-list> about who's guilty. The Original audits suspect plausibility based on guilt — innocents need different treatment than the killer.
- If you're writing in a sub-genre with strict conventions (cozy must have no on-page violence; police procedural must have realistic forensics), name the sub-genre precisely. The clue-density target depends on it.
- Beta-reader flags are predictive. If one reader solved by chapter 6, the structural pattern that allowed that is something the prompt will identify — usually suspect-symmetry or motive-suppression.
- Save the chapter-by-chapter clue map as your manuscript bible. Update it during revisions. Mystery writers who don't keep a bible accumulate continuity errors by book 2.
- Run this on book 2's outline BEFORE drafting, with book 1's misdirection-pattern noted. The prompt will subtract patterns you've already used — series readers spot repeats.
Variants
Cozy / Amateur Sleuth Mode
Lower clue density (6-8 across the manuscript), heavier emphasis on community-knowledge clues over forensic ones, 'the body is the start' structure.
Police Procedural Mode
Higher clue density (12-20), forensic-evidence weighted, multiple-suspect rotation across chapters. Maps procedural-realistic timing.
Domestic Thriller Mode
For Gone Girl / The Silent Patient / unreliable-narrator territory. Calibrates fairness against the unreliable POV — what the narrator hides vs what's structurally fair.
Whodunit Series Mode
When auditing book N in a series. Subtracts misdirection patterns used in books 1 to N-1 so you can't recycle them.
Frequently asked questions
How do I use the Mystery Clue Distribution Planner prompt?
Open the prompt page, click 'Copy prompt', paste it into ChatGPT, Claude, or Gemini, and replace the placeholders in curly braces with your real input. The prompt is also launchable directly in each model with one click.
Which AI model works best with Mystery Clue Distribution Planner?
Claude Opus 4 (best for multi-track structural reasoning across long manuscripts). ChatGPT GPT-5 Pro is second-best.
Can I customize the Mystery Clue Distribution Planner prompt for my use case?
Yes — every Promptolis Original is designed to be customized. Key levers: Run this AFTER you know who did it, why, how, and the order in which the detective will deduce it. Without those, the prompt produces structurally-pretty maps with no spine.; List your suspects honestly — including the one you've decided is innocent. The Original audits whether innocent suspects have enough on-page presence to be plausible suspects (otherwise readers eliminate them too fast).
Explore more Originals
Hand-crafted 2026-grade prompts that actually change how you work.
← All Promptolis Originals