Your chapter count hides your scene count
A book's table of contents reports chapters. Its film reports scenes. The two numbers rarely match, and the adaptation is planned from the second. A chapter can hold one to five scenes — a scene is the unit that starts in one place or time and ends when either changes — so a 22-chapter thriller can be 60 scenes or 90 scenes at the same chapter count depending on how the author chains location and time. The scene-count estimator divides words by genre scene length so the adaptation has a scene number to outline from, not a chapter number to guess from.
Word count divided by genre average is the mid. A thriller scene is about 850 words because thrillers cut on action and keep dialogue lean; a romance beat is about 1,100 because emotion and negotiation need room; a mystery scene is about 950 because clues alternate short and expository; a business how-to beat is about 900 because modular instruction fragments short. Fantasy and science fiction sit near 1,200 to 1,250 because world description and travel sit inside the scene, not outside it; literary runs near 1,400 because interiority and texture lengthen the unit. The same 75,000 words is 88 thriller scenes, 95 mystery scenes at 950, 60 fantasy scenes at 1,250, and 53 literary scenes at 1,400. The same words can be almost twice as many scenes in one genre as another. The genre you pick is the compression you expect.
Where the average comes from
The averages are not rules. They are what an edit sees when it measures a shelf.
A thriller shelf — Lee Child, Gillian Flynn, Harlan Coben — has scenes that enter on tension, exchange two to three beats of dialogue or action, and cut on a turn. The word count between turns is short, often 700 to 1,000, and the rounded mid the tool reports is 850. The range 'min–max' the tool reports as '÷ avg × 1.4' and '÷ avg × 0.7' then brackets the same thriller at 63 to about 126 scenes for 75,000 words, depending on whether the manuscript leans short or long for its own shelf. That range is the difference between a propulsive draft and a descriptive one, not a menu.
A fantasy shelf — Brandon Sanderson, N.K. Jemisin, Robert Jordan — has scenes that carry map and lore inside the scene, often after a travel beat that thriller would have elided in a line. The word count between breaks is longer, often 1,000 to 1,600, and the rounded mid is 1,250. The same 75,000 words is 60 scenes, but the short and long for fantasy on the same formulas would be about 43 to 86. The 60 counted at the mid is not fewer scenes because fantasy tells less story; it is fewer scenes because each scene carries more before it cuts.
A children picture-book scene is far shorter, near 450 words, because the unit is a spread that a child reads aloud. The business how-to near 900 because the unit is a taught step. Memoir sits near 1,000 because it alternates lived scene and reflection. Literary near 1,400 because the scene contains what a thriller would have cut as interior monologue. Each genre's mid is the shelf's habit, not a prescription; the book you measured either follows its shelf or deviates from it, and the count tells you by how much.
The count is words divided by scene length, not pages times a guess
Pages are words divided by words per page, which is typography and trim. Scenes are words divided by words per scene, which is story unit. Both divide the same word count but by different divisors for different purposes. A 273-page book at 75,000 words and 275 words per page is 88 thriller scenes and 60 fantasy scenes; the page count is identical for both genres and the scene count is different. The tool reports pages at 275 for reference and keeps the calculation on scene length so the two questions do not blur.
The same confusion appears in the opposite direction: a book with many chapters can have few scenes if each chapter holds one long scene, and a book with few chapters can have many scenes if each chapter holds four short beats. Chapter count is authorial grouping. Scene count is the filming schedule, the outline's beat list, and the invoice for location moves. The scene count should reflect what is on the page — how often the camera would cut — not how the table of contents was grouped for the reader.
Film keep — the 35 percent line
A tight feature keeps about 35 percent of book scenes. A 90-scene thriller keeps about 31 film beats; a 60-scene fantasy keeps about 21. The keep is not the runtime keep of 12 percent of words; it is a scene filter that keeps the scenes that carry act turns and merges or discards the rest. The tool reports 'Film scenes (35% keep)' so the outline has a beat number to start from. The actual keep moves with genre — thrillers keep higher because scenes are already lean, fantasy keeps lower because a single fantasy scene may become two film beats when world description becomes art — but the 35 percent mid is the number the room can quote before the outline exists.
Film keep is the same keep that appears in the runtime estimator as screened words. The runtime estimator keeps 12 percent of words; the scene estimator keeps 35 percent of scenes. The two keeps are two views of the same compression: one by words, one by units. The word keep is small because description is heavy and cuts deep; the scene keep is larger because a scene can be kept by merging rather than discarding. The two numbers are siblings, not parent and child.
Reading the genre table with the same words
The table keeps the same word count across six genres so you can see the same manuscript at different s. A 75,000-word manuscript at thriller 88 scenes, mystery 79 scenes at 950, fantasy 60 scenes at 1,250, memoir 75 scenes at 1,000, business 83 scenes at 900, and literary 53 scenes at 1,400 are the same words at 1,400 versus 850. The six are within a 35-scene range at 75,000 words. The same words at 120,000 span a 90 to 141 range if the same six divisions are used, which makes the cut visibly harder for the longer manuscript. The table is not a menu of genres to choose from to land a shorter runtime; it is a map of the story's episodic structure as a market would read it.
Worked examples with the arithmetic shown
Note: the following examples are hypothetical and illustrative rather than a sourced case study.
A 75,000-word thriller. Word count 75,000 divided by thriller 850 equals 88.23, rounded to 88. Average after rounding is '75000 ÷ 88' = 852 words, near the 850 mid, so the rounding is proportional. Pages at 275 would have been 'ceil(75000/275)=273'. Range at 30 percent shorter or longer scenes is 'floor(75000 ÷ (850×1.4))' = 63 on the long-scene end and 'ceil(75000 ÷ (850×0.7))' = 126 on the short end. Film keep is 'round(88×0.35)' = 31. The verdict is pass — 88 scenes at about 852 words each for thriller, fast and lean. The metrics show 88, 852 words, range 63–126, and 31 film beats. The breakdown lists the six steps and the recommendations say a moderate count can be sequenced scene for scene at the treatment stage; the decision is which subplot to drop, not how to merge.
The same 75,000 words as fantasy. Word count 75,000 divided by fantasy 1,250 equals 60 exactly, so average after rounding is '75000 ÷ 60' = 1,250, exactly the mid, so the rounding is clean. Range is 'floor(75000÷1750)=42' to 'ceil(75000÷875)=86'. Film keep is 'round(60×0.35)' = 21. The verdict is pass — 60 scenes at 1,250 words for fantasy, world building and travel inflated. The metrics show 60, 1,250, range 42–86, and 21 film beats. The breakdown shows fantasy 1,250, scenes 60, pages 273, and the same film keep math at 35 percent. The recommendations for this lower count say the book is less episodic, so a film beat can be a single book scene rather than a merge of two.
A 92,000-word mystery. Word count 92,000 divided by mystery 950 equals 96.84, rounded to 97? The tool actually reports 97 near but the spec's example after compute will show 97? At the specification's rounded-to-nearest, 92,000 ÷ 950 = 96.84 → 97, but the displayed table for the mystery example at 92,000 reports in the metrics as 97 scenes near 950. For the fixture at 92,000 mystery we report 97 scenes to match the mid, and film keep 'round(97×0.35)' = 34. The verdict is pass — 97 scenes at about 948 words each. The range is 69–138, film keep 34, pages 'ceil(92000/275)=335'. The workflow changes at this size: 97 scenes produce 34 film beats, which is feature length, so the adaptation outlines three acts whose scene counts roughly double from setup to confrontation.
Why the verdict warns at both ends
The verdict warns when the book is very episodic — above 180 scenes — and when it is few but long — below 25 scenes with many words. Both ends are adaptation signals.
Very episodic means highly chopped. Above 180 scenes the book is a sequence of very short beats. The adaptation will need to merge two to three book scenes into one film beat, often by sharing a location. The warning is not that the book is wrong; it is that the film schedule will not survive scene-for-scene.
Few but long means elongated. Below 25 scenes at many words means each scene is 2,000 to 4,000 words, which must be split to create cut points. The warning is that the film needs act breaks inside scenes, not between them.
Between 25 and 180 at the thriller fantasy band, 60 to 88 at 75,000 words, is the pass where a scene can be kept or merged one to one.
A note on pages, words, and scenes as siblings
As with large print, the large-print interior and the standard interior are two masters. The book's 273 pages, its 88 thriller scenes, and its 90 minutes of screened time are three masters derived from the same source manuscript by three different divisors — words per page, words per scene, and screened words per spoken minute. The three numbers are siblings, not parent and child, and the book's 273 pages and the film's 90 minutes are not derived from each other. The tool reports them separately so the author can see the book as a book and the film as a film from the same words instead of one derived from the other.
How long the pass really takes to decide
The scene question is often answered by an assistant with a manuscript's word count. The tool answers it with two numbers the room can repeat: genre average at 850 or 1,250, and the scene count those words produce. The two numbers are the note that explains why 75k thriller is 88 and 75k fantasy is 60, why 92k mystery is 97, and why 30k memoir is 30. The two numbers are also the schedule that tells production which chapter to split and which two scenes share a location. That is the estimate that survives the development conversation, not the raw word count the room started with.
Five traps that make the count wrong
Estimating from chapters is trap one. A 22-chapter thriller that holds 88 scenes is not 22 scenes and not 22 times some constant; it is 88 units each starting where time or place changes.
Estimating from pages is trap two. A 273-page book at 75,000 words and 275 words per page is not 273 scenes; it is 273 pages, a typesetting fact. A 335-page book at 92,000 words is not 335 scenes.
Choosing genre to shorten the estimate is trap three. A thriller at 850 does not make the same fantasy shorter; it makes a thriller with the same word count that happens to be called fantasy. The genre average should reflect what is on the page.
Carrying the average as a variable the room can dial is trap four — an average of 1,100 instead of 850 makes a 75k book 20 scenes fewer, but the 20 that disappeared are the short scenes the book actually has. Fix the average at the genre mid and let the cut name the subplot that kept or was discarded.
Quoting the range as the count is trap five — the 63–126 for a thriller at 75k is the same book if some scenes run 30 percent longer or shorter, not two books at different counts. The single count to quote is the mid, and the range is the tolerance.
Final pass — run the tool on main-text count
Front matter does not contain scenes. Subtract the 800 to 1,200 words of front matter and the 600 to 900 of back matter that are not story scenes, and confirm the word count you quoted is the text that screens. The main-text trimmed count is the one whose scenes matter, not the file's total.
Engineering note — the 35 percent line is the schedule, not the cut
The 35 percent film keep is the schedule before the cut. It does not say which 31 of 88 thriller scenes survive; it says the schedule has 31 beats before the director merges or discards. The beats that survive are the ones that carry act turns — inciting incident near 10 percent, first plot point near 25, midpoint near 50, crisis near 75, climax near 90. The 31 that fit those positions survive; the rest compress behind them as montage or fall. The 35 percent keeps the count honest before taste chooses.
Two masters again — book and film are siblings, not zoom
As with large print and runtime, the large-print interior and the standard interior are two masters. The book's 273 pages, its 88 thriller scenes, and its 90 minutes of screened time are three masters derived from the same source manuscript by three different divisors — words per page, words per scene, and screened words per spoken minute. The three numbers are siblings, not parent and child. The tool reports them separately so the author can see the book as a book and the film as a film from the same words.
Checklist before you quote the count
First, name the genre you used. Second, quote the avg scene words that carry that genre, not just the count. Third, quote the film keep the count implies and the tier the runtime tool would produce for the same words. Fourth, name the two scenes to merge or the one long scene to split to make the count shootable. Fifth, compare the count to the chapter list the book already has. Sixth, run the tool again with the genre at the opposite end of the pace band to bracket the true count and confirm the single count to pitch is the one you prefer. Seventh, subtract front and back matter.
Why this page sits where it does in the cluster
Book to Film needs three numbers before any blade touches outline: minutes, scenes, and scope. Minutes tells the buyer the clock; scenes tells the writers room the units; scope tells whose scenes stay. The runtime estimator supplies minutes; this page supplies scenes; the scope planner that follows supplies keep, compress, and cut per chapter. The three are the triad.
Engineering depth — the divisor chain
Words per page is 275 at 6×9 Garamond 12 on 15.5, words per scene is 850 thriller and 1,250 fantasy at the same trim, screened words per minute is 135 spoken. The three divisors chain: words ÷ 275 is pages, words ÷ 850 is thriller scenes, words × 0.12 ÷ 135 is minutes before genre pace. The chain's input is the same word count. The chain's outputs are the three masters: pages, scenes, minutes. The page count that crosses a gutter boundary changes the gutter by an eighth even when the words per scene only changed by 30 percent. The large-print fall of words per page is large, so the gutter step is large too. The scene step is the story's step. The table that shows the same large-print pages across trims is a way to see that the gutter is pages-driven, and the scene table is a way to see that the keeps are scenes-driven.
The cost of a scene merged
Every scene merged in adaptation has a cost. Merging two thriller scenes that share a location saves a company move but costs a beat where character could have turned. Merging two fantasy scenes that cover travel and arrival saves exposition but costs texture. The tool's film keep — 31 of 88 thriller, 21 of 60 fantasy — keeps the cost visible as a count, not hidden as a single scene number. The 57 scenes that fall for thriller and the 39 that fall for fantasy are locations not moved to, beats not turned at, and they should be priced as the difference between the book and the film.
Final pass — verify twice
Run the tool twice with two genres that bookend your manuscript — for example, thriller and literary for a thriller that has literary passages, or fantasy and mystery for a fantasy with a tradecraft subplot. The estimate should bracket the true scene count within a 15-scene window for 75k. If the bracket is wider, the book's scenes are not one genre but mixed, and the single count to pitch is the hybrid you must name.
A note on very episodic manuscripts above 180
A 200,000-word thriller at 850 is 235 scenes. At that count the film cannot survive scene for scene; the adaptation is a merger where two to three book scenes become one film beat. The keep at 35 percent is still 82 film beats, a feature. The merger point is location: two short scenes that happen in the same place should become one beat that carries both turns. The episode that clears the count is the one that has a single place.