{
 "currency": "usd",
 "baseUrl": "https://hollywoodmetrics.ai",
 "storePath": "/store",
 "volumes": [
  {
   "folder": "001_nobody-knows-anything",
   "n": 1,
   "slug": "nobody-knows-anything-we-checked",
   "pn": 1000001,
   "title": "Nobody Knows Anything. We Checked.",
   "displayTitle": "Nobody Knows Anything. We Checked.",
   "subtitle": "What 74,571 films and 1,571 screenplays reveal about how movies actually work",
   "author": "Hollywood Metrics",
   "imprint": "Hollywood Metrics",
   "isbn": {
    "ebook": "979-8-90623-998-3",
    "paperback": "979-8-90623-997-6"
   },
   "priceCents": {
    "paperback": 1999,
    "ebook": 999
   },
   "ladder": [
    {
     "minQty": 1,
     "label": "One copy",
     "unitCents": 1999,
     "offPct": 0,
     "savingCents": 0
    },
    {
     "minQty": 2,
     "label": "Two or more",
     "unitCents": 1839,
     "offPct": 8,
     "savingCents": 160
    },
    {
     "minQty": 5,
     "label": "Five or more",
     "unitCents": 1639,
     "offPct": 18,
     "savingCents": 360
    },
    {
     "minQty": 10,
     "label": "Ten or more, for a room or a class",
     "unitCents": 1439,
     "offPct": 28,
     "savingCents": 560
    }
   ],
   "pages": 366,
   "words": 86503,
   "chapters": [
    {
     "file": "00-part-1.md",
     "kind": "part",
     "number": "1",
     "part": null,
     "title": "The Outside of a Film",
     "subtitle": "A century of records, and how much of what everyone repeats survives contact with them",
     "kicker": "",
     "words": 0
    },
    {
     "file": "01-zero-point-zero-six-eight.md",
     "kind": "chapter",
     "number": "1",
     "part": "1",
     "title": "Zero Point Zero Six Eight",
     "subtitle": "The number this book was built to find, and what it turned out to be",
     "kicker": "Chapter One",
     "words": 2587
    },
    {
     "file": "02-five-hundred-votes.md",
     "kind": "chapter",
     "number": "2",
     "part": "1",
     "title": "Five Hundred Votes",
     "subtitle": "What is actually in the catalogue, and what the catalogue is not",
     "kicker": "Chapter Two",
     "words": 2641
    },
    {
     "file": "03-one-film-in-six-and-a-half.md",
     "kind": "chapter",
     "number": "3",
     "part": "1",
     "title": "One Film in Six and a Half",
     "subtitle": "The denominator under every claim anyone has ever made about film money",
     "kicker": "Chapter Three",
     "words": 2105
    },
    {
     "file": "04-the-barbell-and-the-silence.md",
     "kind": "chapter",
     "number": "4",
     "part": "1",
     "title": "The Barbell and the Silence",
     "subtitle": "The best-known finding in film economics, and the correction that turns it upside down",
     "kicker": "Chapter Four",
     "words": 2450
    },
    {
     "file": "05-money-buys-reach-not-respect.md",
     "kind": "chapter",
     "number": "5",
     "part": "1",
     "title": "Money Buys Reach, Not Respect",
     "subtitle": "What a production budget purchases, and the one thing it reliably does not",
     "kicker": "Chapter Five",
     "words": 1837
    },
    {
     "file": "06-horror-pays.md",
     "kind": "chapter",
     "number": "6",
     "part": "1",
     "title": "Horror Pays and Prestige Doesn't",
     "subtitle": "Genre as a financing instrument rather than a description of content",
     "kicker": "Chapter Six",
     "words": 2434
    },
    {
     "file": "07-ninety-seven-minutes.md",
     "kind": "chapter",
     "number": "7",
     "part": "1",
     "title": "Ninety-Seven Minutes for Sixty Years",
     "subtitle": "A claim everyone repeats, false for the entire period they were repeating it",
     "kicker": "Chapter Seven",
     "words": 2002
    },
    {
     "file": "08-they-dont-make-them.md",
     "kind": "chapter",
     "number": "8",
     "part": "1",
     "title": "They Don't Make Them Like They Used To",
     "subtitle": "The most seductive artifact in the catalogue, and why the correction makes it worse",
     "kicker": "Chapter Eight",
     "words": 1987
    },
    {
     "file": "09-someone-is-wrong.md",
     "kind": "chapter",
     "number": "9",
     "part": "1",
     "title": "Someone in This Relationship Is Wrong",
     "subtitle": "Two crowds, and the decade their disagreement changed sign",
     "kicker": "Chapter Nine",
     "words": 1999
    },
    {
     "file": "10-the-only-thing-that-pays.md",
     "kind": "chapter",
     "number": "10",
     "part": "1",
     "title": "The Only Thing That Reliably Pays",
     "subtitle": "The belief this catalogue most decisively overturns",
     "kicker": "Chapter Ten",
     "words": 2423
    },
    {
     "file": "11-a-dollar-is-not-a-unit.md",
     "kind": "chapter",
     "number": "11",
     "part": "1",
     "title": "A Dollar Is Not a Unit",
     "subtitle": "Not one line of code in this system deflates a dollar, and nor does anyone else's",
     "kicker": "Chapter Eleven",
     "words": 2038
    },
    {
     "file": "12-part-2.md",
     "kind": "part",
     "number": "2",
     "part": null,
     "title": "The Instruments",
     "subtitle": "What a score in this book is made of, and what it measures when nobody is checking",
     "kicker": "",
     "words": 0
    },
    {
     "file": "13-the-recipe.md",
     "kind": "chapter",
     "number": "12",
     "part": "2",
     "title": "The Recipe",
     "subtitle": "The master score, taken apart, exactly",
     "kicker": "Chapter Twelve",
     "words": 2130
    },
    {
     "file": "14-a-rank-of-who-turned-up.md",
     "kind": "chapter",
     "number": "13",
     "part": "2",
     "title": "A Rank of Who Turned Up",
     "subtitle": "The popularity component is not a rating, it is a headcount",
     "kicker": "Chapter Thirteen",
     "words": 1573
    },
    {
     "file": "15-two-ingredients.md",
     "kind": "chapter",
     "number": "14",
     "part": "2",
     "title": "Two Ingredients",
     "subtitle": "What most films in this catalogue are actually scored on",
     "kicker": "Chapter Fourteen",
     "words": 1859
    },
    {
     "file": "16-twenty-columns-five-things.md",
     "kind": "chapter",
     "number": "15",
     "part": "2",
     "title": "Twenty Columns, Five Things",
     "subtitle": "How many independent measurements twenty features actually contain",
     "kicker": "Chapter Fifteen",
     "words": 1820
    },
    {
     "file": "17-part-3.md",
     "kind": "part",
     "number": "3",
     "part": null,
     "title": "The Inside of a Screenplay",
     "subtitle": "Fifteen hundred screenplays, twenty measurements, and the answer this project was built to find",
     "kicker": "",
     "words": 0
    },
    {
     "file": "18-fifteen-hundred-screenplays.md",
     "kind": "chapter",
     "number": "16",
     "part": "3",
     "title": "Fifteen Hundred and Seventy-One Screenplays",
     "subtitle": "A reasonable thing to attempt, and how it was attempted",
     "kicker": "Chapter Sixteen",
     "words": 2049
    },
    {
     "file": "19-nobody-read-these.md",
     "kind": "chapter",
     "number": "17",
     "part": "3",
     "title": "Nobody Read These",
     "subtitle": "The grades were never craft judgements, and what that does to the question",
     "kicker": "Chapter Seventeen",
     "words": 1975
    },
    {
     "file": "20-twenty-measurements.md",
     "kind": "chapter",
     "number": "18",
     "part": "3",
     "title": "Twenty Measurements and No Signal",
     "subtitle": "The central result, stated plainly and then attacked four ways",
     "kicker": "Chapter Eighteen",
     "words": 2126
    },
    {
     "file": "21-seventy-four-percent.md",
     "kind": "chapter",
     "number": "19",
     "part": "3",
     "title": "Seventy-Four Percent Had No Dialogue",
     "subtitle": "The instrument was broken, and here is exactly how",
     "kicker": "Chapter Nineteen",
     "words": 1995
    },
    {
     "file": "22-three-findings-dismantled.md",
     "kind": "chapter",
     "number": "20",
     "part": "3",
     "title": "Three Findings, Dismantled",
     "subtitle": "Everything this project published as a positive result",
     "kicker": "Chapter Twenty",
     "words": 2359
    },
    {
     "file": "23-we-fixed-it.md",
     "kind": "chapter",
     "number": "21",
     "part": "3",
     "title": "We Fixed It and Got the Same Answer",
     "subtitle": "The chapter that turns a broken study into a finding",
     "kicker": "Chapter Twenty-One",
     "words": 2145
    },
    {
     "file": "24-budget-beats-everything.md",
     "kind": "chapter",
     "number": "22",
     "part": "3",
     "title": "Budget Beats Everything You Can Read",
     "subtitle": "What the text adds over what the greenlight memo already knows",
     "kicker": "Chapter Twenty-Two",
     "words": 1603
    },
    {
     "file": "25-everything-else.md",
     "kind": "chapter",
     "number": "23",
     "part": "3",
     "title": "Everything Else Anybody Knows",
     "subtitle": "The director, the writer, the studio and the distributor, measured the same way",
     "kicker": "Chapter Twenty-Three",
     "words": 2756
    },
    {
     "file": "26-half-life-of-an-honest-number.md",
     "kind": "chapter",
     "number": "24",
     "part": "3",
     "title": "The Half-Life of an Honest Number",
     "subtitle": "What an organisation does when it finds its own bug",
     "kicker": "Chapter Twenty-Four",
     "words": 1721
    },
    {
     "file": "27-part-4.md",
     "kind": "part",
     "number": "4",
     "part": null,
     "title": "The Cases",
     "subtitle": "Thirteen films taken apart one at a time, because an aggregate tells you what is true and never what it means",
     "kicker": "",
     "words": 0
    },
    {
     "file": "28-how-to-read-a-case.md",
     "kind": "chapter",
     "number": "25",
     "part": "4",
     "title": "How to Read a Case",
     "subtitle": "The method, stated once, so the eight chapters after it can get on with it",
     "kicker": "Chapter Twenty-Five",
     "words": 838
    },
    {
     "file": "29-breaks-both-instruments.md",
     "kind": "chapter",
     "number": "26",
     "part": "4",
     "title": "The Film That Breaks Both Instruments",
     "subtitle": "The Shawshank Redemption, measured twice, wrong twice, in opposite directions",
     "kicker": "Chapter Twenty-Six",
     "words": 1228
    },
    {
     "file": "30-the-streaming-zero.md",
     "kind": "chapter",
     "number": "27",
     "part": "4",
     "title": "The Streaming Zero",
     "subtitle": "The largest failure in the catalogue is not a failure",
     "kicker": "Chapter Twenty-Seven",
     "words": 1072
    },
    {
     "file": "31-eighty-seven-times-the-line.md",
     "kind": "chapter",
     "number": "28",
     "part": "4",
     "title": "Eighty-Seven Times the Line",
     "subtitle": "The other tail, which is forty years older and exactly as misleading",
     "kicker": "Chapter Twenty-Eight",
     "words": 992
    },
    {
     "file": "32-the-cheapest-things.md",
     "kind": "chapter",
     "number": "29",
     "part": "4",
     "title": "The Cheapest Things That Ever Worked",
     "subtitle": "One finding that is completely real, and one number that should never have been printed",
     "kicker": "Chapter Twenty-Nine",
     "words": 1080
    },
    {
     "file": "33-three-wins-and-no-money.md",
     "kind": "chapter",
     "number": "30",
     "part": "4",
     "title": "Three Wins and No Money",
     "subtitle": "CODA, and what an exception to a strong finding is usually made of",
     "kicker": "Chapter Thirty",
     "words": 1056
    },
    {
     "file": "34-two-crowds-two-films.md",
     "kind": "chapter",
     "number": "31",
     "part": "4",
     "title": "Two Crowds, Three Films",
     "subtitle": "Where the two audiences disagree most, and what a famous flop's score is made of",
     "kicker": "Chapter Thirty-One",
     "words": 1033
    },
    {
     "file": "35-the-film-nobody-wrote-about.md",
     "kind": "chapter",
     "number": "32",
     "part": "4",
     "title": "The Film Nobody Wrote About",
     "subtitle": "The middle of the catalogue, which is what the catalogue is mostly made of",
     "kicker": "Chapter Thirty-Two",
     "words": 923
    },
    {
     "file": "36-two-hundred-and-thirty-two-pages.md",
     "kind": "chapter",
     "number": "33",
     "part": "4",
     "title": "Two Hundred and Thirty-Two Pages",
     "subtitle": "A second bug, and the one case that is still wrong after the fix",
     "kicker": "Chapter Thirty-Three",
     "words": 1027
    },
    {
     "file": "37-part-5.md",
     "kind": "part",
     "number": "5",
     "part": null,
     "title": "The Advice",
     "subtitle": "A second corpus, of nine hundred books about how to do it",
     "kicker": "",
     "words": 0
    },
    {
     "file": "38-nine-hundred-guides.md",
     "kind": "chapter",
     "number": "34",
     "part": "5",
     "title": "Nine Hundred and Twenty-Six Style Guides",
     "subtitle": "The other corpus, and why it is worth measuring",
     "kicker": "Chapter Thirty-Four",
     "words": 2396
    },
    {
     "file": "39-where-the-advice-disagrees.md",
     "kind": "chapter",
     "number": "35",
     "part": "5",
     "title": "Where the Advice Disagrees With Itself",
     "subtitle": "Act one ends on page thirty. Act one ends on page twenty-five.",
     "kicker": "Chapter Thirty-Five",
     "words": 717
    },
    {
     "file": "40-what-it-talks-about.md",
     "kind": "chapter",
     "number": "36",
     "part": "5",
     "title": "What It Talks About, and What It Does Not",
     "subtitle": "Twelve topics, four crafts, and some conspicuous gaps",
     "kicker": "Chapter Thirty-Six",
     "words": 1813
    },
    {
     "file": "41-the-tells.md",
     "kind": "chapter",
     "number": "37",
     "part": "5",
     "title": "The Tells",
     "subtitle": "What the text says about how the text was made",
     "kicker": "Chapter Thirty-Seven",
     "words": 757
    },
    {
     "file": "42-what-could-be-reprinted.md",
     "kind": "chapter",
     "number": "38",
     "part": "5",
     "title": "What Could Actually Be Reprinted",
     "subtitle": "Two hundred words of somebody else's writing, and where they are",
     "kicker": "Chapter Thirty-Eight",
     "words": 1036
    },
    {
     "file": "43-part-6.md",
     "kind": "part",
     "number": "6",
     "part": null,
     "title": "The Record",
     "subtitle": "The tables, each printed with the filter that produced it",
     "kicker": "",
     "words": 0
    },
    {
     "file": "44-how-to-read-the-record.md",
     "kind": "chapter",
     "number": "39",
     "part": "6",
     "title": "How to Read the Record",
     "subtitle": "The cleaning rules, stated once, so every table can be checked",
     "kicker": "Chapter Thirty-Nine",
     "words": 763
    },
    {
     "file": "45-the-decade-atlas.md",
     "kind": "chapter",
     "number": "40",
     "part": "6",
     "title": "The Decade Atlas",
     "subtitle": "Eleven decades on one template, so the differences are differences",
     "kicker": "Chapter Forty",
     "words": 5967
    },
    {
     "file": "46-the-genre-atlas.md",
     "kind": "chapter",
     "number": "41",
     "part": "6",
     "title": "The Genre Atlas",
     "subtitle": "Every genre counted twice, because there are two defensible ways",
     "kicker": "Chapter Forty-One",
     "words": 1939
    },
    {
     "file": "47-the-lists.md",
     "kind": "chapter",
     "number": "42",
     "part": "6",
     "title": "The Lists",
     "subtitle": "Four rankings, each with its gate printed",
     "kicker": "Chapter Forty-Two",
     "words": 5488
    },
    {
     "file": "48-appendix-a-method.md",
     "kind": "appendix",
     "number": null,
     "part": null,
     "title": "Method",
     "subtitle": "Where every number came from, and what it is allowed to say",
     "kicker": "Appendix A",
     "words": 3433
    },
    {
     "file": "49-appendix-b-features.md",
     "kind": "appendix",
     "number": null,
     "part": null,
     "title": "The Twenty Features",
     "subtitle": "What was measured on each screenplay, exactly",
     "kicker": "Appendix B",
     "words": 1667
    },
    {
     "file": "50-appendix-c-sources.md",
     "kind": "appendix",
     "number": null,
     "part": null,
     "title": "Sources, Licences and What Is Not Reproduced",
     "subtitle": "Where the data came from, and the line this book does not cross",
     "kicker": "Appendix C",
     "words": 664
    }
   ],
   "independentPass": false
  }
 ],
 "dropped": [],
 "notes": [
  "PUBLISHING.md quotes $24.99 as the retail list price; the registry holds $19.99 for direct sale. Both may be right - direct is cheaper because no retailer takes a cut - but the difference is deliberate and nothing here resolves it."
 ],
 "shipping": {
  "zones": [
   {
    "id": "us",
    "name": "United States",
    "quoteCountry": "US",
    "firstCents": 699,
    "extraCents": 125,
    "daysMin": 11,
    "daysMax": 15,
    "freeOverCents": 7500
   },
   {
    "id": "uk",
    "name": "United Kingdom and Ireland",
    "quoteCountry": "GB",
    "firstCents": 649,
    "extraCents": 120,
    "daysMin": 7,
    "daysMax": 16,
    "freeOverCents": 8500
   },
   {
    "id": "eu",
    "name": "Europe",
    "quoteCountry": "DE",
    "firstCents": 845,
    "extraCents": 140,
    "daysMin": 9,
    "daysMax": 17,
    "freeOverCents": 9500
   },
   {
    "id": "anz",
    "name": "Australia and New Zealand",
    "quoteCountry": "AU",
    "firstCents": 979,
    "extraCents": 160,
    "daysMin": 11,
    "daysMax": 18,
    "freeOverCents": 11000
   },
   {
    "id": "row",
    "name": "Rest of the world",
    "quoteCountry": null,
    "firstCents": 1895,
    "extraCents": 320,
    "daysMin": 14,
    "daysMax": 25,
    "freeOverCents": null
   }
  ],
  "countries": [
   {
    "code": "US",
    "name": "United States",
    "zone": "us"
   },
   {
    "code": "CA",
    "name": "Canada",
    "zone": "us"
   },
   {
    "code": "GB",
    "name": "United Kingdom",
    "zone": "uk"
   },
   {
    "code": "IE",
    "name": "Ireland",
    "zone": "uk"
   },
   {
    "code": "DE",
    "name": "Germany",
    "zone": "eu"
   },
   {
    "code": "FR",
    "name": "France",
    "zone": "eu"
   },
   {
    "code": "NL",
    "name": "Netherlands",
    "zone": "eu"
   },
   {
    "code": "ES",
    "name": "Spain",
    "zone": "eu"
   },
   {
    "code": "IT",
    "name": "Italy",
    "zone": "eu"
   },
   {
    "code": "BE",
    "name": "Belgium",
    "zone": "eu"
   },
   {
    "code": "AT",
    "name": "Austria",
    "zone": "eu"
   },
   {
    "code": "SE",
    "name": "Sweden",
    "zone": "eu"
   },
   {
    "code": "DK",
    "name": "Denmark",
    "zone": "eu"
   },
   {
    "code": "FI",
    "name": "Finland",
    "zone": "eu"
   },
   {
    "code": "PL",
    "name": "Poland",
    "zone": "eu"
   },
   {
    "code": "PT",
    "name": "Portugal",
    "zone": "eu"
   },
   {
    "code": "CH",
    "name": "Switzerland",
    "zone": "eu"
   },
   {
    "code": "NO",
    "name": "Norway",
    "zone": "eu"
   },
   {
    "code": "AU",
    "name": "Australia",
    "zone": "anz"
   },
   {
    "code": "NZ",
    "name": "New Zealand",
    "zone": "anz"
   },
   {
    "code": "JP",
    "name": "Japan",
    "zone": "row"
   },
   {
    "code": "SG",
    "name": "Singapore",
    "zone": "row"
   },
   {
    "code": "ZA",
    "name": "South Africa",
    "zone": "row"
   },
   {
    "code": "BR",
    "name": "Brazil",
    "zone": "row"
   },
   {
    "code": "MX",
    "name": "Mexico",
    "zone": "row"
   },
   {
    "code": "IN",
    "name": "India",
    "zone": "row"
   }
  ],
  "defaultCountry": "US",
  "levels": [
   {
    "level": "MAIL",
    "name": "Standard post"
   },
   {
    "level": "PRIORITY_MAIL",
    "name": "Tracked post"
   },
   {
    "level": "EXPEDITED",
    "name": "Courier, two day"
   },
   {
    "level": "EXPRESS",
    "name": "Courier, overnight"
   }
  ],
  "handlingMarkup": 1.06,
  "handlingFlatCents": 95
 },
 "print": {
  "podPackageId": "0600X0900.BW.STD.PB.060UC444.MXX",
  "trim": "6 x 9 in",
  "interior": "black and white",
  "paper": "60# uncoated cream",
  "binding": "perfect bound",
  "finish": "matte laminate cover",
  "returnsDays": 30
 },
 "ebookFreeWithPrint": true,
 "factCount": 252,
 "free": {
  "headlines": [
   {
    "key": "screenplay.strongest_correlation_rebuilt",
    "display": "r = 0.068",
    "value": 0.0678,
    "unit": "correlation",
    "n": 1499,
    "of": "screenplays re-measured with the corrected parser and joined to their outcomes",
    "source": "book/research/refixed_full_corpus.csv",
    "method": "the largest absolute correlation between any rebuilt feature and the composite quality score",
    "caveat": "The instrument was rebuilt, every structural measure changed, and the answer did not. That is what makes the null result a finding rather than a failure."
   },
   {
    "key": "models.r2_budget_only",
    "display": "41%",
    "value": 0.4121,
    "unit": "percent",
    "n": 11427,
    "of": "films with a reported budget of at least $10,000 and a reported worldwide gross of at least $1,000",
    "source": "public/data.json",
    "method": "R-squared of an ordinary least squares fit predicting log worldwide gross from log production budget alone",
    "caveat": "One number, available before anyone reads the script, and it is the benchmark every cleverer model has to beat."
   },
   {
    "key": "models.boxoffice_model_r2",
    "display": "12.7%",
    "value": 0.1267,
    "unit": "percent",
    "n": 1289,
    "of": "films used to fit the shipped box office model",
    "source": "public/boxoffice_model.json",
    "method": "the R-squared reported in the shipped model file",
    "caveat": "This model reads twenty features of the screenplay text. It explains less than a linear fit on the budget alone."
   },
   {
    "key": "corpus.money_coverage_share",
    "display": "15.46%",
    "value": 15.46,
    "unit": "percent",
    "n": 74571,
    "of": "films in the catalogue",
    "source": "public/data.json",
    "method": "share carrying both a reported budget and a reported worldwide gross",
    "caveat": "Every profitability statement in this book, and in every book like it, describes this fraction of the catalogue and not the catalogue. The films that never reported are not a random sample of the ones that did."
   },
   {
    "key": "corpus.total_films",
    "display": "74,571",
    "value": 74571,
    "unit": "count",
    "n": 74571,
    "of": "films in the shipped catalogue",
    "source": "public/data.json",
    "method": "len() of public/data.json",
    "caveat": null
   },
   {
    "key": "screenplay.corpus_size",
    "display": "1,571",
    "value": 1571,
    "unit": "count",
    "n": 1571,
    "of": "graded screenplays that entered the analysis pipeline",
    "source": "pipeline/output/features_with_outcomes.csv",
    "method": "row count of the shipped feature table",
    "caveat": null
   },
   {
    "key": "craft.guides.total",
    "display": "926",
    "value": 926,
    "unit": "count",
    "n": 926,
    "of": "markdown craft guides on disk across the four style directories, excluding the three PROGRESS.md trackers",
    "source": "screenwriter-styles/*.md, directors-styles/*.md, cinematographers-styles/*.md, actor-styles/*.md",
    "method": "Counted *.md files in screenwriter-styles, directors-styles, cinematographers-styles and actor-styles; 929 files exist, three are PROGRESS.md trackers and are excluded.",
    "caveat": "929 files are on disk. The three trackers are not guides and all three misstate the size of the directory they sit in."
   },
   {
    "key": "models.boxoffice_model_interval_95",
    "display": "a factor of 27",
    "value": 27.25,
    "unit": "ratio",
    "n": 1289,
    "of": "films used to fit the shipped box office model",
    "source": "public/boxoffice_model.json",
    "method": "the multiplicative 95 percent interval implied by the residual standard deviation",
    "caveat": "A prediction of $50M carries a 95 percent range from about $2M to $1B. That is not a forecast, it is a genre."
   }
  ],
  "findings": [
   {
    "key": "screenplay.strongest_correlation_rebuilt",
    "display": "r = 0.068",
    "value": 0.0678,
    "unit": "correlation",
    "n": 1499,
    "of": "screenplays re-measured with the corrected parser and joined to their outcomes",
    "source": "book/research/refixed_full_corpus.csv",
    "method": "the largest absolute correlation between any rebuilt feature and the composite quality score",
    "caveat": "The instrument was rebuilt, every structural measure changed, and the answer did not. That is what makes the null result a finding rather than a failure.",
    "chapter": 14,
    "claim": "Nothing in the text of a screenplay predicts how the film was received"
   },
   {
    "key": "models.r2_budget_only",
    "display": "41%",
    "value": 0.4121,
    "unit": "percent",
    "n": 11427,
    "of": "films with a reported budget of at least $10,000 and a reported worldwide gross of at least $1,000",
    "source": "public/data.json",
    "method": "R-squared of an ordinary least squares fit predicting log worldwide gross from log production budget alone",
    "caveat": "One number, available before anyone reads the script, and it is the benchmark every cleverer model has to beat.",
    "chapter": 18,
    "claim": "The budget explains three times more than twenty measurements of the script"
   },
   {
    "key": "corpus.money_coverage_share",
    "display": "15.46%",
    "value": 15.46,
    "unit": "percent",
    "n": 74571,
    "of": "films in the catalogue",
    "source": "public/data.json",
    "method": "share carrying both a reported budget and a reported worldwide gross",
    "caveat": "Every profitability statement in this book, and in every book like it, describes this fraction of the catalogue and not the catalogue. The films that never reported are not a random sample of the ones that did.",
    "chapter": 3,
    "claim": "Six films in seven never report what they made"
   },
   {
    "key": "money.barbell_micro",
    "display": "216%",
    "value": 216.4,
    "unit": "percent",
    "n": 1152,
    "of": "films budgeted under $1M that reported a gross",
    "source": "public/data.json",
    "method": "median percentage return for the cheapest budget band",
    "caveat": null,
    "chapter": 4,
    "claim": "Counting the silent films as failures inverts the cheap end entirely"
   },
   {
    "key": "awards.oscar_profitability_raw",
    "display": "93.4% against 61.4%",
    "value": 93.4,
    "unit": "percent",
    "n": 11427,
    "of": "films with a reported budget and gross, split by whether they won an Academy Award",
    "source": "public/data.json",
    "method": "share of each group whose reported gross exceeds its reported budget",
    "caveat": "Uncontrolled. Award winners are larger, later and better distributed than the median film, so this comparison flatters them. See the matched version.",
    "chapter": 10,
    "claim": "The prestige slate is the most reliably profitable class in the catalogue"
   },
   {
    "key": "genre.horror_median_roi",
    "display": "203%",
    "value": 202.6,
    "unit": "percent",
    "n": 536,
    "of": "horror films with a reported budget and gross",
    "source": "public/data.json",
    "method": "median percentage return among horror films passing the floors",
    "caveat": "Horror also has one of the lowest gross-reporting rates (49.4%), so the same survivorship correction that flattens the micro-budget story applies here.",
    "chapter": 6,
    "claim": "Horror returns more per dollar than anything else that reports"
   },
   {
    "key": "screenplay.no_dialogue_share",
    "display": "74%",
    "value": 74.1,
    "unit": "percent",
    "n": 1571,
    "of": "graded screenplays in the shipped data",
    "source": "pipeline/output/features_with_outcomes.csv",
    "method": "share whose measured dialogue ratio is exactly zero",
    "caveat": "This is the single number that invalidates the published effect sizes.",
    "chapter": 15,
    "claim": "Three quarters of the screenplay corpus read as having no dialogue at all"
   },
   {
    "key": "craft.act_one.disagreement",
    "display": "41 guides say page 30, 28 say page 25",
    "value": null,
    "unit": "text",
    "n": 71,
    "of": "the 71 screenwriter guides that state where act one ends",
    "source": "screenwriter-styles/*.md, directors-styles/*.md, cinematographers-styles/*.md, actor-styles/*.md",
    "method": "The two commonest closing pages for act one and their guide counts.",
    "caveat": "No guide anywhere in the corpus acknowledges that the other number exists. The single most-taught figure in screenwriting is split 41 to 28 inside one corpus written by one process.",
    "chapter": 21,
    "claim": "The most-taught number in screenwriting splits 41 to 28 inside one corpus"
   },
   {
    "key": "corpus.vote_gate_dropped",
    "display": "83,606",
    "value": 83606,
    "unit": "count",
    "n": 158178,
    "of": "rows in the master file, of which these fell below the gate",
    "source": "hollywood_analysis.csv",
    "method": "count of rows with fewer than 500 IMDb votes in hollywood_analysis.csv; the build script keeps only rows at or above 500",
    "caveat": "This single filter is the most consequential decision in the whole dataset. It removes the majority of the source rows and it removes them unevenly.",
    "chapter": 2,
    "claim": "One filter removed the majority of the source rows, unevenly"
   }
  ],
  "faq": [
   {
    "q": "Does this book say screenwriting does not matter?",
    "a": "No. It says that twenty structural measurements of a screenplay text - scene count, dialogue ratio, sentiment arc and seventeen others - carry almost no information about how the finished film was received, across 1,571 produced films. The strongest correlation found was 0.068, which explains under half a per cent of the variance. That is a finding about what those measurements can see, not about whether writing matters. Chapter fourteen states the claim precisely and chapter seventeen states what survives it."
   },
   {
    "q": "Is the analysis reproducible?",
    "a": "Every number in the book resolves to an entry in a machine-readable fact base carrying the value, the sample size, what that sample is a sample of, the source file, the method and the caveat. The manuscript is checked against it before every build: a number without a citation stops the build, and a table that disagrees with the fact it cites stops the build. Appendix A sets out the method in full."
   },
   {
    "q": "Why are so many films missing from the money figures?",
    "a": "Because they never reported. Only 15.46 per cent of the 74,571 films in the catalogue carry both a production budget and a worldwide gross, and the ones that do are not a random sample of the ones that do not. Every profitability statement in this book, and in every book like it, describes that fraction. Part One computes each of them twice - once over the films that reported, once over a bound that treats every non-reporting film as a failure - and prints both."
   },
   {
    "q": "What does the book get wrong?",
    "a": "Three findings this project published before the book existed are withdrawn in chapter sixteen, because the parser that produced them was reading three quarters of the corpus as having no dialogue. Chapter fifteen is the bug, chapter seventeen is the rebuild, and chapter nineteen is what happened between the fix being written and the numbers being corrected. It is the most uncomfortable chapter in the book."
   },
   {
    "q": "Can I check a number without buying the book?",
    "a": "Yes. The correlation matrix that carries the central finding is printed in full on this site, with the sample size and the method beside it, and so is every headline figure. The reference tables in Part Four are the part that needs 200 printed pages."
   },
   {
    "q": "Do you sell through Amazon as well?",
    "a": "Yes, and the paperback costs less here. A retailer takes a cut of every copy; buying direct removes it, and the ebook comes with the paperback at no extra charge, which no retailer can offer."
   },
   {
    "q": "How long does a copy take to arrive?",
    "a": "Each copy is printed when it is ordered, so the window covers manufacturing as well as carriage. Standard post to a United States address is 11 to 15 days; to the United Kingdom, 7 to 9. The delivery page carries every zone, and checkout quotes the real services for the country you name."
   },
   {
    "q": "What is in the reference section?",
    "a": "About a thousand rows: every decade and every genre on one template, the hundred highest master scores, the fifty largest returns and the fifty largest grosses, the highest-scoring film of each of 107 years, and the twenty largest gaps between a budget and a gross. Each table prints the filter that produced it, which is the part most published lists leave out."
   }
  ],
  "limits": [
   {
    "key": "corpus.vote_gate_dropped",
    "display": "83,606",
    "value": 83606,
    "unit": "count",
    "n": 158178,
    "of": "rows in the master file, of which these fell below the gate",
    "source": "hollywood_analysis.csv",
    "method": "count of rows with fewer than 500 IMDb votes in hollywood_analysis.csv; the build script keeps only rows at or above 500",
    "caveat": "This single filter is the most consequential decision in the whole dataset. It removes the majority of the source rows and it removes them unevenly."
   },
   {
    "key": "corpus.money_coverage_share",
    "display": "15.46%",
    "value": 15.46,
    "unit": "percent",
    "n": 74571,
    "of": "films in the catalogue",
    "source": "public/data.json",
    "method": "share carrying both a reported budget and a reported worldwide gross",
    "caveat": "Every profitability statement in this book, and in every book like it, describes this fraction of the catalogue and not the catalogue. The films that never reported are not a random sample of the ones that did."
   },
   {
    "key": "screenplay.tier_circularity_r2",
    "display": "R-squared = 0.92",
    "value": 0.9164,
    "unit": "ratio",
    "n": 1229,
    "of": "graded screenplays carrying a grade and the reception measures",
    "source": "pipeline/output/features_with_outcomes.csv",
    "method": "R-squared of a linear fit predicting the screenplay grade from the finished film's reception: imdb_rating, rt_tomatometer, oscar_wins",
    "caveat": "The grade on these screenplays is almost entirely recoverable from how the finished film was received. Nobody read them. Correlating the text against the grade is therefore partly correlating the text against the film, which is not the same question."
   },
   {
    "key": "models.boxoffice_model_r2",
    "display": "12.7%",
    "value": 0.1267,
    "unit": "percent",
    "n": 1289,
    "of": "films used to fit the shipped box office model",
    "source": "public/boxoffice_model.json",
    "method": "the R-squared reported in the shipped model file",
    "caveat": "This model reads twenty features of the screenplay text. It explains less than a linear fit on the budget alone."
   },
   {
    "key": "craft.table.decade_economics",
    "display": "11 rows",
    "value": null,
    "unit": "list",
    "n": 11427,
    "of": "films with a reported budget of at least $10,000 and a reported worldwide gross of at least $1,000, which is 11,427 of the 74,571 in the catalogue, grouped by release decade",
    "source": "public/data.json",
    "method": "Each film's budget and gross deflated individually by the CPI-U annual average for its release year, rebased to 2025, then the median taken. Author's adjustment; the index is printed in craft.table.cpi_index.",
    "caveat": "The real columns are the author's external assumption, not a repository figure, and nothing in the repository is inflation adjusted. They are also unsafe before about 1970, where a lifetime gross accumulated across re-releases is being deflated at the price level of the release year. The comparison the table supports is the recent one: in real terms the median reporting film's budget peaked in the 1990s and has fallen by more than half since."
   },
   {
    "key": "craft.best_film_per_year.floor_effect",
    "display": "60 of 107 years",
    "value": 60,
    "unit": "count",
    "n": 107,
    "of": "the 107 release years, in how many the highest rated film changes if the 20,000-vote floor is removed",
    "source": "public/data.json",
    "method": "Compared the per-year winner over all films against the per-year winner over films with at least 20,000 votes.",
    "caveat": "The unfiltered winners include a 2026 release rated 10.0 on 512 votes. This number is the argument for the floor, and it is why no \"best film of the year\" claim in this book is made without one."
   },
   {
    "key": "craft.guides.authorship",
    "display": "n/a",
    "value": null,
    "unit": "text",
    "n": 0,
    "of": "the 926 craft guides and their front matter",
    "source": "screenwriter-styles/*.md, directors-styles/*.md, cinematographers-styles/*.md, actor-styles/*.md",
    "method": "Front matter keys inspected; no author or date field exists.",
    "caveat": null
   },
   {
    "key": "craft.reprint.clearance",
    "display": "n/a",
    "value": null,
    "unit": "text",
    "n": 0,
    "of": "the 926 craft guides and the third-party screenplay text inside them",
    "source": "screenwriter-styles/*.md, directors-styles/*.md, cinematographers-styles/*.md, actor-styles/*.md",
    "method": "Deliberately not computed.",
    "caveat": null
   }
  ],
  "theNull": {
   "key": "screenplay.correlation_matrix_shipped",
   "columns": [
    {
     "key": "master_score",
     "label": "Master score"
    },
    {
     "key": "imdb_rating",
     "label": "Audience rating"
    },
    {
     "key": "rt_tomatometer",
     "label": "Critic score"
    },
    {
     "key": "worldwide_gross",
     "label": "Worldwide gross"
    },
    {
     "key": "roi",
     "label": "Return on budget"
    },
    {
     "key": "oscar_wins",
     "label": "Academy Award wins"
    }
   ],
   "rows": [
    {
     "feature": "scene_count",
     "label": "Scene count",
     "cells": [
      {
       "outcome": "master_score",
       "label": "Master score",
       "r": -0.04
      },
      {
       "outcome": "imdb_rating",
       "label": "Audience rating",
       "r": -0.0008
      },
      {
       "outcome": "rt_tomatometer",
       "label": "Critic score",
       "r": -0.0452
      },
      {
       "outcome": "worldwide_gross",
       "label": "Worldwide gross",
       "r": 0.0289
      },
      {
       "outcome": "roi",
       "label": "Return on budget",
       "r": -0.0215
      },
      {
       "outcome": "oscar_wins",
       "label": "Academy Award wins",
       "r": 0.0002
      }
     ],
     "strongest": -0.0452
    },
    {
     "feature": "avg_scene_length",
     "label": "Average scene length",
     "cells": [
      {
       "outcome": "master_score",
       "label": "Master score",
       "r": 0.0595
      },
      {
       "outcome": "imdb_rating",
       "label": "Audience rating",
       "r": 0.0538
      },
      {
       "outcome": "rt_tomatometer",
       "label": "Critic score",
       "r": 0.0351
      },
      {
       "outcome": "worldwide_gross",
       "label": "Worldwide gross",
       "r": 0.0194
      },
      {
       "outcome": "roi",
       "label": "Return on budget",
       "r": 0.0022
      },
      {
       "outcome": "oscar_wins",
       "label": "Academy Award wins",
       "r": 0.0198
      }
     ],
     "strongest": 0.0595
    },
    {
     "feature": "int_ext_ratio",
     "label": "Interior to exterior ratio",
     "cells": [
      {
       "outcome": "master_score",
       "label": "Master score",
       "r": -0.0217
      },
      {
       "outcome": "imdb_rating",
       "label": "Audience rating",
       "r": -0.0397
      },
      {
       "outcome": "rt_tomatometer",
       "label": "Critic score",
       "r": -0.0254
      },
      {
       "outcome": "worldwide_gross",
       "label": "Worldwide gross",
       "r": -0.0683
      },
      {
       "outcome": "roi",
       "label": "Return on budget",
       "r": -0.0279
      },
      {
       "outcome": "oscar_wins",
       "label": "Academy Award wins",
       "r": -0.0222
      }
     ],
     "strongest": -0.0683
    },
    {
     "feature": "scene_length_variance",
     "label": "Scene length variance",
     "cells": [
      {
       "outcome": "master_score",
       "label": "Master score",
       "r": -0.0183
      },
      {
       "outcome": "imdb_rating",
       "label": "Audience rating",
       "r": -0.023
      },
      {
       "outcome": "rt_tomatometer",
       "label": "Critic score",
       "r": -0.0006
      },
      {
       "outcome": "worldwide_gross",
       "label": "Worldwide gross",
       "r": -0.0035
      },
      {
       "outcome": "roi",
       "label": "Return on budget",
       "r": -0.0173
      },
      {
       "outcome": "oscar_wins",
       "label": "Academy Award wins",
       "r": -0.0143
      }
     ],
     "strongest": -0.023
    },
    {
     "feature": "total_pages",
     "label": "Total pages",
     "cells": [
      {
       "outcome": "master_score",
       "label": "Master score",
       "r": 0.0393
      },
      {
       "outcome": "imdb_rating",
       "label": "Audience rating",
       "r": 0.0873
      },
      {
       "outcome": "rt_tomatometer",
       "label": "Critic score",
       "r": 0.0223
      },
      {
       "outcome": "worldwide_gross",
       "label": "Worldwide gross",
       "r": 0.1416
      },
      {
       "outcome": "roi",
       "label": "Return on budget",
       "r": -0.0629
      },
      {
       "outcome": "oscar_wins",
       "label": "Academy Award wins",
       "r": 0.1081
      }
     ],
     "strongest": 0.1416
    },
    {
     "feature": "transition_density",
     "label": "Transition density",
     "cells": [
      {
       "outcome": "master_score",
       "label": "Master score",
       "r": -0.0327
      },
      {
       "outcome": "imdb_rating",
       "label": "Audience rating",
       "r": -0.0243
      },
      {
       "outcome": "rt_tomatometer",
       "label": "Critic score",
       "r": -0.0029
      },
      {
       "outcome": "worldwide_gross",
       "label": "Worldwide gross",
       "r": -0.0551
      },
      {
       "outcome": "roi",
       "label": "Return on budget",
       "r": 0.0105
      },
      {
       "outcome": "oscar_wins",
       "label": "Academy Award wins",
       "r": -0.0041
      }
     ],
     "strongest": -0.0551
    },
    {
     "feature": "unique_character_count",
     "label": "Distinct speaking characters",
     "cells": [
      {
       "outcome": "master_score",
       "label": "Master score",
       "r": 0.0086
      },
      {
       "outcome": "imdb_rating",
       "label": "Audience rating",
       "r": 0.0481
      },
      {
       "outcome": "rt_tomatometer",
       "label": "Critic score",
       "r": 0.0154
      },
      {
       "outcome": "worldwide_gross",
       "label": "Worldwide gross",
       "r": -0.0526
      },
      {
       "outcome": "roi",
       "label": "Return on budget",
       "r": 0.021
      },
      {
       "outcome": "oscar_wins",
       "label": "Academy Award wins",
       "r": 0.0106
      }
     ],
     "strongest": -0.0526
    },
    {
     "feature": "dialogue_ratio",
     "label": "Dialogue ratio",
     "cells": [
      {
       "outcome": "master_score",
       "label": "Master score",
       "r": 0.0018
      },
      {
       "outcome": "imdb_rating",
       "label": "Audience rating",
       "r": 0.0415
      },
      {
       "outcome": "rt_tomatometer",
       "label": "Critic score",
       "r": 0.0279
      },
      {
       "outcome": "worldwide_gross",
       "label": "Worldwide gross",
       "r": -0.0546
      },
      {
       "outcome": "roi",
       "label": "Return on budget",
       "r": 0.0067
      },
      {
       "outcome": "oscar_wins",
       "label": "Academy Award wins",
       "r": 0.0054
      }
     ],
     "strongest": -0.0546
    },
    {
     "feature": "top3_character_dominance",
     "label": "Top three character dominance",
     "cells": [
      {
       "outcome": "master_score",
       "label": "Master score",
       "r": -0.0192
      },
      {
       "outcome": "imdb_rating",
       "label": "Audience rating",
       "r": 0.0235
      },
      {
       "outcome": "rt_tomatometer",
       "label": "Critic score",
       "r": 0.0145
      },
      {
       "outcome": "worldwide_gross",
       "label": "Worldwide gross",
       "r": -0.0454
      },
      {
       "outcome": "roi",
       "label": "Return on budget",
       "r": -0.039
      },
      {
       "outcome": "oscar_wins",
       "label": "Academy Award wins",
       "r": -0.007
      }
     ],
     "strongest": -0.0454
    },
    {
     "feature": "avg_dialogue_length",
     "label": "Average speech length",
     "cells": [
      {
       "outcome": "master_score",
       "label": "Master score",
       "r": -0.0141
      },
      {
       "outcome": "imdb_rating",
       "label": "Audience rating",
       "r": 0.0199
      },
      {
       "outcome": "rt_tomatometer",
       "label": "Critic score",
       "r": 0.0022
      },
      {
       "outcome": "worldwide_gross",
       "label": "Worldwide gross",
       "r": -0.0576
      },
      {
       "outcome": "roi",
       "label": "Return on budget",
       "r": -0.013
      },
      {
       "outcome": "oscar_wins",
       "label": "Academy Award wins",
       "r": -0.017
      }
     ],
     "strongest": -0.0576
    },
    {
     "feature": "character_intro_rate",
     "label": "Character introduction rate",
     "cells": [
      {
       "outcome": "master_score",
       "label": "Master score",
       "r": 0.0073
      },
      {
       "outcome": "imdb_rating",
       "label": "Audience rating",
       "r": 0.0286
      },
      {
       "outcome": "rt_tomatometer",
       "label": "Critic score",
       "r": 0.0093
      },
      {
       "outcome": "worldwide_gross",
       "label": "Worldwide gross",
       "r": -0.0546
      },
      {
       "outcome": "roi",
       "label": "Return on budget",
       "r": 0.0678
      },
      {
       "outcome": "oscar_wins",
       "label": "Academy Award wins",
       "r": -0.0057
      }
     ],
     "strongest": 0.0678
    },
    {
     "feature": "vocabulary_richness",
     "label": "Vocabulary richness",
     "cells": [
      {
       "outcome": "master_score",
       "label": "Master score",
       "r": -0.0683
      },
      {
       "outcome": "imdb_rating",
       "label": "Audience rating",
       "r": -0.0871
      },
      {
       "outcome": "rt_tomatometer",
       "label": "Critic score",
       "r": -0.1343
      },
      {
       "outcome": "worldwide_gross",
       "label": "Worldwide gross",
       "r": 0.0029
      },
      {
       "outcome": "roi",
       "label": "Return on budget",
       "r": -0.0316
      },
      {
       "outcome": "oscar_wins",
       "label": "Academy Award wins",
       "r": -0.0592
      }
     ],
     "strongest": -0.1343
    },
    {
     "feature": "avg_word_length",
     "label": "Average word length",
     "cells": [
      {
       "outcome": "master_score",
       "label": "Master score",
       "r": -0.0485
      },
      {
       "outcome": "imdb_rating",
       "label": "Audience rating",
       "r": -0.0676
      },
      {
       "outcome": "rt_tomatometer",
       "label": "Critic score",
       "r": -0.0881
      },
      {
       "outcome": "worldwide_gross",
       "label": "Worldwide gross",
       "r": 0.0441
      },
      {
       "outcome": "roi",
       "label": "Return on budget",
       "r": -0.0208
      },
      {
       "outcome": "oscar_wins",
       "label": "Academy Award wins",
       "r": -0.0326
      }
     ],
     "strongest": -0.0881
    },
    {
     "feature": "sentiment_mean",
     "label": "Sentiment, mean",
     "cells": [
      {
       "outcome": "master_score",
       "label": "Master score",
       "r": 0.057
      },
      {
       "outcome": "imdb_rating",
       "label": "Audience rating",
       "r": 0.0914
      },
      {
       "outcome": "rt_tomatometer",
       "label": "Critic score",
       "r": 0.1438
      },
      {
       "outcome": "worldwide_gross",
       "label": "Worldwide gross",
       "r": -0.0072
      },
      {
       "outcome": "roi",
       "label": "Return on budget",
       "r": 0.0005
      },
      {
       "outcome": "oscar_wins",
       "label": "Academy Award wins",
       "r": 0.0425
      }
     ],
     "strongest": 0.1438
    },
    {
     "feature": "sentiment_variance",
     "label": "Sentiment, variance",
     "cells": [
      {
       "outcome": "master_score",
       "label": "Master score",
       "r": 0.0016
      },
      {
       "outcome": "imdb_rating",
       "label": "Audience rating",
       "r": -0.0324
      },
      {
       "outcome": "rt_tomatometer",
       "label": "Critic score",
       "r": -0.0237
      },
      {
       "outcome": "worldwide_gross",
       "label": "Worldwide gross",
       "r": 0.0522
      },
      {
       "outcome": "roi",
       "label": "Return on budget",
       "r": 0.0319
      },
      {
       "outcome": "oscar_wins",
       "label": "Academy Award wins",
       "r": -0.0099
      }
     ],
     "strongest": 0.0522
    },
    {
     "feature": "sentiment_arc_slope",
     "label": "Sentiment arc slope",
     "cells": [
      {
       "outcome": "master_score",
       "label": "Master score",
       "r": 0.022
      },
      {
       "outcome": "imdb_rating",
       "label": "Audience rating",
       "r": 0.0442
      },
      {
       "outcome": "rt_tomatometer",
       "label": "Critic score",
       "r": 0.0113
      },
      {
       "outcome": "worldwide_gross",
       "label": "Worldwide gross",
       "r": -0.0237
      },
      {
       "outcome": "roi",
       "label": "Return on budget",
       "r": -0.0145
      },
      {
       "outcome": "oscar_wins",
       "label": "Academy Award wins",
       "r": -0.0216
      }
     ],
     "strongest": 0.0442
    },
    {
     "feature": "action_ratio",
     "label": "Action ratio",
     "cells": [
      {
       "outcome": "master_score",
       "label": "Master score",
       "r": 0.0058
      },
      {
       "outcome": "imdb_rating",
       "label": "Audience rating",
       "r": -0.035
      },
      {
       "outcome": "rt_tomatometer",
       "label": "Critic score",
       "r": -0.0287
      },
      {
       "outcome": "worldwide_gross",
       "label": "Worldwide gross",
       "r": 0.0614
      },
      {
       "outcome": "roi",
       "label": "Return on budget",
       "r": -0.0035
      },
      {
       "outcome": "oscar_wins",
       "label": "Academy Award wins",
       "r": -0.0095
      }
     ],
     "strongest": 0.0614
    },
    {
     "feature": "caps_density",
     "label": "Capitals density",
     "cells": [
      {
       "outcome": "master_score",
       "label": "Master score",
       "r": -0.0135
      },
      {
       "outcome": "imdb_rating",
       "label": "Audience rating",
       "r": -0.0126
      },
      {
       "outcome": "rt_tomatometer",
       "label": "Critic score",
       "r": -0.0327
      },
      {
       "outcome": "worldwide_gross",
       "label": "Worldwide gross",
       "r": 0.1595
      },
      {
       "outcome": "roi",
       "label": "Return on budget",
       "r": -0.0512
      },
      {
       "outcome": "oscar_wins",
       "label": "Academy Award wins",
       "r": -0.0199
      }
     ],
     "strongest": 0.1595
    },
    {
     "feature": "exclamation_density",
     "label": "Exclamation density",
     "cells": [
      {
       "outcome": "master_score",
       "label": "Master score",
       "r": 0.0262
      },
      {
       "outcome": "imdb_rating",
       "label": "Audience rating",
       "r": 0.0138
      },
      {
       "outcome": "rt_tomatometer",
       "label": "Critic score",
       "r": 0.0442
      },
      {
       "outcome": "worldwide_gross",
       "label": "Worldwide gross",
       "r": 0.1566
      },
      {
       "outcome": "roi",
       "label": "Return on budget",
       "r": 0.0187
      },
      {
       "outcome": "oscar_wins",
       "label": "Academy Award wins",
       "r": 0.0196
      }
     ],
     "strongest": 0.1566
    },
    {
     "feature": "question_density",
     "label": "Question density",
     "cells": [
      {
       "outcome": "master_score",
       "label": "Master score",
       "r": 0.0361
      },
      {
       "outcome": "imdb_rating",
       "label": "Audience rating",
       "r": 0.0592
      },
      {
       "outcome": "rt_tomatometer",
       "label": "Critic score",
       "r": 0.1216
      },
      {
       "outcome": "worldwide_gross",
       "label": "Worldwide gross",
       "r": -0.0958
      },
      {
       "outcome": "roi",
       "label": "Return on budget",
       "r": 0.0326
      },
      {
       "outcome": "oscar_wins",
       "label": "Academy Award wins",
       "r": -0.0543
      }
     ],
     "strongest": 0.1216
    }
   ],
   "n": 1571,
   "of": "graded screenplays, correlating each structural feature against each outcome measure",
   "method": "Pearson correlation per pair, over rows where both are present",
   "caveat": null,
   "rebuilt": {
    "key": "screenplay.strongest_correlation_rebuilt",
    "display": "r = 0.068",
    "value": 0.0678,
    "unit": "correlation",
    "n": 1499,
    "of": "screenplays re-measured with the corrected parser and joined to their outcomes",
    "source": "book/research/refixed_full_corpus.csv",
    "method": "the largest absolute correlation between any rebuilt feature and the composite quality score",
    "caveat": "The instrument was rebuilt, every structural measure changed, and the answer did not. That is what makes the null result a finding rather than a failure."
   },
   "benchmark": {
    "key": "models.r2_budget_only",
    "display": "41%",
    "value": 0.4121,
    "unit": "percent",
    "n": 11427,
    "of": "films with a reported budget of at least $10,000 and a reported worldwide gross of at least $1,000",
    "source": "public/data.json",
    "method": "R-squared of an ordinary least squares fit predicting log worldwide gross from log production budget alone",
    "caveat": "One number, available before anyone reads the script, and it is the benchmark every cleverer model has to beat."
   }
  },
  "features": [
   {
    "key": "scene_count",
    "label": "Scene count",
    "unit": "scenes",
    "definition": "The number of scene headings detected. A scene heading is a line the parser reads as a slugline, so this counts formatting, not narrative units."
   },
   {
    "key": "avg_scene_length",
    "label": "Average scene length",
    "unit": "lines",
    "definition": "Mean of action lines plus dialogue lines per scene. Measured in lines rather than pages or seconds."
   },
   {
    "key": "int_ext_ratio",
    "label": "Interior to exterior ratio",
    "unit": "ratio",
    "definition": "Interior scenes divided by exterior scenes, from the INT. and EXT. markers in the sluglines. A screenplay with no exterior scenes divides by one rather than by zero, so the value is the interior count."
   },
   {
    "key": "scene_length_variance",
    "label": "Scene length variance",
    "unit": "lines squared",
    "definition": "Sample variance of the scene lengths above. High values mean a mix of very long and very short scenes."
   },
   {
    "key": "total_pages",
    "label": "Total pages",
    "unit": "pages",
    "definition": "Total lines divided by fifty-six, the standard lines per screenplay page. An estimate from the text, not a page count from a PDF."
   },
   {
    "key": "transition_density",
    "label": "Transition density",
    "unit": "per page",
    "definition": "Count of lines beginning FADE IN, FADE OUT, FADE TO, CUT TO, DISSOLVE, SMASH CUT or MATCH CUT, divided by the estimated page count."
   },
   {
    "key": "unique_character_count",
    "label": "Distinct speaking characters",
    "unit": "characters",
    "definition": "The number of distinct character cues the parser found. A character who appears under two spellings counts twice; one who never speaks does not count at all."
   },
   {
    "key": "dialogue_ratio",
    "label": "Dialogue ratio",
    "unit": "share of lines",
    "definition": "Dialogue lines divided by total lines. This is the feature that read exactly zero for three quarters of the corpus before the parser was repaired, and chapter [!ch:seventy-four-percent] is what that did."
   },
   {
    "key": "top3_character_dominance",
    "label": "Top three character dominance",
    "unit": "share",
    "definition": "Words spoken by the three largest parts, divided by all spoken words. A two-hander scores near one; an ensemble scores low."
   },
   {
    "key": "avg_dialogue_length",
    "label": "Average speech length",
    "unit": "words",
    "definition": "Spoken words divided by the number of dialogue blocks. The average length of one uninterrupted speech."
   },
   {
    "key": "character_intro_rate",
    "label": "Character introduction rate",
    "unit": "per page",
    "definition": "Distinct speaking characters divided by the estimated page count. Named as a rate, computed as a density: it does not measure when characters arrive, only how many there are per page."
   },
   {
    "key": "vocabulary_richness",
    "label": "Vocabulary richness",
    "unit": "ratio",
    "definition": "Distinct words divided by total words, over the first five thousand words. A type-token ratio, sampled at a fixed length because the raw ratio falls as a document gets longer and would otherwise measure length."
   },
   {
    "key": "avg_word_length",
    "label": "Average word length",
    "unit": "characters",
    "definition": "Mean length in characters of every alphabetic word in the document, action and dialogue together."
   },
   {
    "key": "sentiment_mean",
    "label": "Sentiment, mean",
    "unit": "score",
    "definition": "Mean VADER compound sentiment over the dialogue blocks, on a scale from minus one to plus one. VADER is a lexicon tuned on social media text, and applying it to screen dialogue is an assumption rather than a measurement."
   },
   {
    "key": "sentiment_variance",
    "label": "Sentiment, variance",
    "unit": "score squared",
    "definition": "Sample variance of the same block scores. Read as emotional range."
   },
   {
    "key": "sentiment_arc_slope",
    "label": "Sentiment arc slope",
    "unit": "score per tenth",
    "definition": "The document is cut into ten equal chunks, each chunk scored, and a straight line fitted across them. Positive means the dialogue gets more positive from start to end. A straight line through ten points is a crude summary of an emotional arc and is treated as one."
   },
   {
    "key": "action_ratio",
    "label": "Action ratio",
    "unit": "share of lines",
    "definition": "Non-blank lines that are not dialogue, divided by all non-blank lines. Roughly the inverse of the dialogue ratio, and it fails in the same way when the parser fails."
   },
   {
    "key": "caps_density",
    "label": "Capitals density",
    "unit": "share of words",
    "definition": "Words of two or more capital letters, divided by total words. Intended to capture emphasis and sound cues; it also captures character cues and sluglines, so it is partly a measure of formatting."
   },
   {
    "key": "exclamation_density",
    "label": "Exclamation density",
    "unit": "share of words",
    "definition": "Exclamation marks divided by total words, across the whole document."
   },
   {
    "key": "question_density",
    "label": "Question density",
    "unit": "share of words",
    "definition": "Question marks divided by total words, across the whole document."
   }
  ]
 },
 "art": {
  "hero-ledger": {
   "alt": "A long ledger table under one low lamp, ruled columns running away into darkness",
   "aspect": "21:9"
  },
  "card-book": {
   "alt": "A thick unmarked paperback standing on end on a dark table",
   "aspect": "4:3"
  },
  "banner-the-null": {
   "alt": "A single taut horizontal wire over a dark grid, lit from one side",
   "aspect": "16:9"
  },
  "banner-contents": {
   "alt": "A wooden card index drawer pulled open, packed with blank cards",
   "aspect": "16:9"
  },
  "banner-printing": {
   "alt": "A guillotine blade above a stack of freshly cut cream paper",
   "aspect": "16:9"
  },
  "banner-limits": {
   "alt": "A wide flat drawer standing mostly empty, a few objects at one end",
   "aspect": "16:9"
  },
  "tile-corpus": {
   "alt": "Hundreds of film cans stacked in a dark store, one lit",
   "aspect": "4:3"
  },
  "tile-screenplays": {
   "alt": "A tall stack of loose brass-bound paper on a dark table",
   "aspect": "4:3"
  },
  "tile-guides": {
   "alt": "Identical unmarked pamphlets fanned across a dark surface",
   "aspect": "4:3"
  },
  "tile-money": {
   "alt": "A bank of small drawers, most shut, a few standing open and empty",
   "aspect": "4:3"
  },
  "banner-rule": {
   "alt": "A machinist's spirit level on a dark slab, its bubble dead centre",
   "aspect": "16:9"
  },
  "banner-questions": {
   "alt": "An empty reading chair and side table under one lamp, in darkness",
   "aspect": "16:9"
  },
  "spread-open": {
   "alt": "A thick book lying open, both pages blank cream stock, lit from one side",
   "aspect": "16:9"
  },
  "divider-desk": {
   "alt": "A long dark desk under one lamp, a closed book and loose paper on it",
   "aspect": "21:9"
  },
  "tile-calipers": {
   "alt": "Precision measuring instruments laid out in a row on dark felt",
   "aspect": "4:3"
  },
  "tile-parcel": {
   "alt": "A single book-shaped parcel in brown paper and string on a dark bench",
   "aspect": "4:3"
  },
  "tile-press": {
   "alt": "The glue wheel and clamp of a perfect binding machine, close",
   "aspect": "4:3"
  },
  "tile-stack": {
   "alt": "Finished paperbacks stacked square on a dark bench, spines outward",
   "aspect": "4:3"
  },
  "tile-archive": {
   "alt": "A shelf of unmarked box files, one gap where a box is missing",
   "aspect": "4:3"
  },
  "social-card": {
   "alt": "A single strip of blank film on a light box, curling into darkness",
   "aspect": "16:9"
  }
 }
}