The Methodology

StoryEngine's guidance is grounded in a structural study of 344 produced works across film, television, stage, and musical theatre. This page explains exactly what was studied, how, what each figure means, and where the limits are. Our claims should be auditable — so here they are.

1. The Corpus

The study spans roughly 90 years of produced storytelling (1931–2024), deliberately mixing critical recognition with popular endurance so the patterns hold for both award bodies and audiences.

Film — 121 screenplays

• 31 Academy Award winners for Best Picture / Best Original or Adapted Screenplay (2005–2024 core, with earlier anchors).
• 58 IMDb Top 100 titles — audience-endurance selections spanning 1942–2024.
• 32 Cinema Archives "Best 1000 Films" selections — auteur and canon titles, 1944–2015 (Kubrick, the Coens, Malick, Hitchcock, P.T. Anderson and others).

Television — 112 series

Series spanning 1951–2024, from I Love Lucy to Succession, The Bear, and Fleabag — drama, comedy, prestige cable, network, and streaming, selected for critical and cultural significance.

Musical theatre — 75 musicals

The AFI Top 25 Greatest Film Musicals (1933–2002) combined with 50 landmark stage musicals, covering book structure, song placement, and the integration of number and scene.

Stage — 36 major playwrights

From Ibsen, Chekhov, and Shakespeare through O'Neill, Williams, Miller, and contemporary voices (Wilson, Letts, Nottage, Kushner, McDonagh, Baker, Kane), organized by structural school (family play, two-hander, social/political, memory play, epic ensemble).

2. How Each Work Was Analyzed

Every work was read or screened and coded against a consistent set of structural markers. For each title we recorded, where applicable:

• Wound placement — where (and whether) the protagonist's core psychological wound is established.
• Antagonist type — human, internal, systemic, or environmental.
• Subtext — whether dialogue routinely operates on a second level (meaning diverging from words).
• Resolution type — triumphant, bittersweet, tragic, open/ambiguous, or nihilistic.
• Coda — whether a quiet scene follows the apparent climax, and what it does.
• Plus format-specific markers (series question and character trajectory for TV; song-type and act structure for musicals; act/scene architecture for stage).

Markers were tallied across each medium's set, and the percentages you see on the site are simply the share of analyzed works exhibiting that marker. The full analysis is documented in StoryEngine's internal research files (the source documents behind the tool's generation rules).

3. What the Metrics Mean

Term	Working definition used in coding
Wound	The protagonist's specific unhealed psychological need — not backstory, but the thing the story exists to confront.
Internal / systemic antagonist	The primary opposing force is the protagonist's own flaw, a system/institution, or the environment — rather than a single human villain.
Subtext	Dialogue in which the literal words and the intended meaning diverge; what is withheld carries the scene.
Coda	A short, quiet scene after the climax whose function is thematic statement rather than plot resolution.
Series question (TV)	A central dramatic question that cannot be answered without ending the series.

4. Headline Findings

• 97% of analyzed films establish the protagonist's wound in the first image or first spoken line.
• 94% feature an antagonist that is internal, systemic, or environmental rather than a traditional human villain.
• 100% of the scripts analyzed use dialogue that operates on subtext.
• 59% include a Coda — a quiet scene after the apparent climax.
• Resolutions skew bittersweet (≈44%) and tragic (≈20%) over triumphant.
• 97% of analyzed TV series are built on a single unanswerable "series question"; 89% of TV protagonists cycle or regress rather than arc cleanly upward.

5. Honest Limitations

• Curatorial bias. The corpus is selected for acclaim and endurance, so the findings describe successful work, not all work.
• Qualitative coding. Markers like "wound" and "subtext" require judgment; they were coded consistently but by human reading, not an automated rubric.
• Genre sample sizes vary. Genre-specific percentages (e.g., on the Crime or Horror pages) are drawn from smaller subsets — directional signals, not large-sample statistics. Where a figure rests on a small set, treat it as a strong tendency, not a proof.
• Absolutes are scoped. When you see "100%," it means "every work in the set we analyzed," not "every story ever written."

The point of the research isn't to reduce storytelling to a formula. It's to give you a fast, evidence-grounded read on where your draft sits relative to what has reliably worked — so you make your choices on purpose.

6. Want the Source?

The 8 Universal Constants — the most load-bearing findings, with examples — are available as a free PDF: download it here. Questions about the methodology? hello@storyenginehq.com.