Data room / Run 1, stage 2: the nine cross-examinations
Institutional investor
run-1/reviews/crossexam-investor.md Open the file as written
CROSS-EXAMINATION — INVESTOR / CREDIT REVIEWER (Stage 2)
Reviewer: senior credit + venture/PE lens. Date: 2026-08-01. Stage 1 position re-adopted in full: investor.md, verdict KILL (return engine below cost of capital at trend HPA; liquidity thesis unfalsifiable until fatal; uncompensated tails; marketing materially inconsistent with spec). Nothing in the other eight reports moves me off that verdict; several reports strengthen it. This document does three things only: (1) disputes findings I believe are wrong or overstated, with arithmetic; (2) identifies peer-proposed fixes that create NEW problems in fund economics, structured finance, and portfolio risk; (3) flags direct conflicts between reviewers.
Method note: all quantitative disputes below were re-verified in Python this session (amortization, waterfall, and PV cross-checks). Where I checked a peer's number and it held, I say so explicitly rather than manufacture a dispute.
================================================================================
PART 1 — DISPUTES
================================================================================
DISPUTE 1 (insurance-actuarial, Finding 1.1 and Verdict bullet 3):
"Distressed sales almost never breach the floor" / "escrow-first 1.0x recovery is
genuinely robust in sale exits even at -45%" — WRONG on spec-standard parameters.
The actuarial report states that on a -45% sand-state replay, "the oldest deal's escrow after 7% selling costs and mortgage payoff still nets ~158K against a 126K capital claim — fully recovered," and concludes floor loss is essentially a foreclosure-only phenomenon. Rebuilt with the spec's own standard deal (625K enrollment, 350K first mortgage at enrollment, 30yr amortization at a deep-lock 3.5%):
- Sale at -45% from ENROLLMENT value: sale 343,750; net of 7% costs 319,688; mortgage balance after 7 years ~297,656; escrow available to the investor ~22,031 against a 126,000 claim. Recovery 0.17x. Not "fully recovered" — an 83% principal loss on a SALE.
- Even granting the oldest deal a 20% pre-peak runup (enrolled 24 months before peak, -45% from peak): escrow to investor ~103,431. Recovery 0.82x. Still breached.
- Netting 158K requires an implied mortgage payoff of ~225,600 — i.e., a mortgage roughly 35% smaller than the spec's, or 20+ years of amortization. The number cannot be produced from the spec's parameters.
The structural error: the actuarial stress was built "fully deployed at the market peak," which means the BULK of the book is peak-enrolled — for those vintages -45% from peak IS -45% from enrollment, and sale exits breach the floor catastrophically, with the shortfall expressly absorbed by the investor (spec Section 33 short-sale provision; my A2). Their oldest-deal example is the best-case vintage presented as representative. Two knock-ons: (a) their Equity Shield shortfall table (0.47x-0.95x reserve coverage) UNDERSTATES the tail — distressed-sale severity is not the modest add-on they modeled but a second ~full-severity channel; their kill on the Shield gets stronger, not weaker; (b) their one pro-product mitigation ("the core contract's escrow-first recovery is robust, so the core contract need not be killed") is unsupported. My Stage 1 C5 (no spread anywhere to absorb ANY credit loss) stands against the corrected numbers with more force. Internal-consistency note: their own Finding 1.2 shows the pause freezes deployment early in a drawdown (~7.5K average deployed); a peak-adjacent vintage would carry ~45K deployed at trough, not the 126K their oldest-deal example assumes. Their loss model cannot have both full deployment and a working pause. Either correction changes their tables.
DISPUTE 2 (insurance-actuarial, Finding 1.1): the floor-loss decomposition omits a
loss channel the estate reviewer establishes as SETTLED LAW.
The actuarial expected-loss basis ("floor loss = foreclosure rate x deployed capital"; normal-decade expected loss 0.5-1%) prices only foreclosure wipeout. Estate-bankruptcy 4.1 establishes (SETTLED LAW: 11 U.S.C. 101(5); Johnson v. Home State Bank) that as drafted — "not a lien," per spec Section 34 — the position is a dischargeable claim in Chapter 7, i.e., a sale-independent ~100%-severity loss channel; red-team F2/M4 shows it is industrializable and CORRELATED (published-roadmap contagion) in a book deliberately selected for equity-rich cash-poor households. Consequence for the actuarial pricing critique: their "fee is 6-18x expected loss" overpricing claim is overstated (true expected loss is higher than the foreclosure-only basis), while their tail-underfunding claim is understated. Same verdict, wrong decomposition — and the decomposition is what a Tier 2 buyer would price from. No new legal claim by me; I adopt estate 4.1's tag.
DISPUTE 3 (retail-customer, V2 and T8): "the HELOC is CHEAPER than Equity Flow for
essentially all exits up to ~7-10 years in any appreciating market" — overstated, and
contradicted by retail's own table.
Retail's own V-table, 3%/yr appreciation (which is trend, not a down market): 7-year hold, Equity Flow net cost 39,068 vs HELOC interest 45,367 — Equity Flow CHEAPER; APR-equivalent 8.2% vs the 8.5% HELOC. 12-year hold at 3%: 82,980 vs 135,729 — Equity Flow cheaper again. The claim is true only for short holds (under ~5 years) and above-trend appreciation (5%+). The correct statement — which is actually a sharper indictment — is the mirror of my B2 grid: the cells where the product is fair-to-cheap for the homeowner (trend HPA, long holds) are EXACTLY the cells where the fund earns 5.7-7.2% gross, below any cost of capital; the cells where the fund earns its marketed 11%+ are the cells where the homeowner is paying 12.8-34.7% APR-equivalent. The two sides of the trade cannot both be adequately paid at any HPA. Retail's framing ("always worse for the homeowner") hands the sponsor an easy rebuttal exhibit (their own 3% row); the pricing-identity framing does not. Self-correction in the same spirit: my Stage 1 B6 ("cheaper alternatives exist for most qualified borrowers") deserves the same qualifier — at trend HPA and 7+ year holds the HELOC advantage inverts; my conversion-cap argument survives because prospects are sold the 5%-HPA story, but B6 as written was one notch too absolute. Also DISPUTE the adjacent framing in V1: calling 5%/yr "a historically normal market" is generous; long-run US nominal HPA is ~3.5-4.5% (my B2). Retail's 23.1% APR at 3% makes the point without the vulnerable assumption.
DISPUTE 4 (red-team, C2): the incumbent "monthly-draw lump sum" wrapper is "strictly
dominant" — overstated on capital economics.
Concur that the wrapper neutralizes the differentiation and inherits none of the tax defect (tax A-2). But "strictly dominant" ignores the one genuine economic innovation in Equity Flow: staged deployment. A lump sum dripped from escrow deploys 100% of capital at day one; Equity Flow's schedule averages ~50% of committed capital outstanding over the term, which is why my B2 grid shows 26-44% gross IRRs on early exits. An incumbent funding the wrapper eats roughly double the time-weighted capital per unit of consumer experience; pricing it at a 15-20% share (red team's number) is achievable only because of the incumbents' rated-securitization cost of capital (actuarial 3.2), not because the wrapper is costless. The correct claim: dominance via cheaper funding, distribution, and legal chassis — likely, and lethal enough — but the "one drafting cycle, strictly better on price" version would not survive an incumbent's own IC memo. This matters because C2 is the load-bearing exhibit for red team's "no right to exist" argument; the honest version still carries it, the inflated version invites rebuttal.
DISPUTE 5 (securities-consumer-finance, Verdict point 1): "why not kill the core
product" rests on three premises that fail on fund economics.
(a) "Lump-sum HEIs operate today at scale under materially similar characterization risk" — contradicted by tax A-2 (Equity Flow is strictly worse-positioned: the monthly cadence and 1.0x floor forfeit even the incumbents' unblessed comfort) and by title 1.2 (the 1.0x floor fails the CFPB Roberts loss-sharing test MORE cleanly than the products already under attack). The risk is not similar; it is a superset. (b) "Expensive but solvable compliance builds" — solvable legally, not economically. The build the panel itself prices (operator 5.1: 1.5-3M pre-revenue; operator verdict: 4-7M and 12-18 months; BD/RIA registration per sec-cf 1.4-1.5) sits against a proof-of-concept platform whose total origination-fee capacity is 630K-1.26M spread over 2-3 years (operator 4.1) and whose fund-level fee stream is undefined (my O1). The fix list exceeds roughly a decade of platform revenue at launch scale. A compliance build the business cannot fund is not a fix; it is a restatement of the kill. (c) "A launch footprint excluding CT, MD, WA... is a coherent strategy" — coherent for a lawyer, an amputation for the fund: add title's Texas exclusion (SETTLED — void homestead encumbrance), Florida's death-exit gutting (title 5.2 — and death is one of three organic exits), Illinois and Maine (title 1.2), and the McGill exposure concentrated in California (red team M5), and the surviving footprint excludes the two largest housing markets and the deepest locked-in/high-appreciation cohorts. Per my M1 funnel math and operator 4.2's conversion rates, the residual footprint cannot reach operator 4.1's break-even volume. The core contract's fix-before-launch verdict implicitly assumes a financeable business survives the fixes. It does not.
DISPUTE 6 (product-operator, Finding 4.1, minor): crediting the platform with a
1.5-2% management fee double-counts against the marketed LP returns.
Operator hedges with "if the platform is also the fund manager," so this is a flag, not an error: no management fee, carry, or servicing fee exists anywhere in the 39-section spec (my O1), and the origination fee is already double-promised (my A15: LP "revenue stream 2" AND the opex budget). Any pro forma that both funds the platform with 375K-1M/yr of management fees and delivers LPs the marketed ~11% is internally inconsistent: my B5 shows a standard 2/20 load takes the best-case cell to ~7.4% net. The platform's solvency and the LP pitch are drawing on the same undefined fee stack; peer models should state which one they are bankrupting.
VERIFIED, NOT DISPUTED (mandated cross-checks)
- Actuarial Tier 2 pricing (4.3, 5.1) vs my Part B4: their point estimate for a year-15 exit at 3% HPA, r=10%, priced at year 7 with the ratchet is 112,466 — identical to my 112K. Their probability-weighted 86,866 (0.69x deployed) is lower than my scenario-based 112-130K solely because their mobility model (mean exit ~24yr on the lock-in-selected book) is longer than my year-15 point assumption — and their distribution is better grounded than my point estimate. I adopt it: my B4/S3 "Tier 2 sale locks in roughly T-bill-or-worse at trend HPA" was, if anything, generous. Their 1.41x-required vs 0.69-1.10x-offered gap is confirmed.
- Retail APR-equivalents (V-table): re-derived 7yr/5% at 12.8% (vs my cruder 12.3% estimate — theirs correctly front-loads the fee netting) and 3yr/5% at 34.7% (consistent with interpolating my B2 grid). Arithmetic stands; only the V2 framing is disputed (Dispute 3).
- Operator cost stack (1.1): 3,150-8,000 fully loaded, midpoint ~5,300, and the minimum-eligibility deal losing ~4,000 at origination — consistent with my U1 (4,000-8,000+) and strictly better evidenced. I adopt their midpoint and their fee-basis-inversion finding (economics invert where the funnel produces volume).
- Estate 4.1, tax P-1/D-1, title 1.2/5.1: outside my domain; noted only that each independently strengthens my Stage 1 D2, A12, A13 and none contradicts my model.
================================================================================
PART 2 — PEER FIXES THAT CREATE NEW PROBLEMS IN MY DOMAIN
================================================================================
FX-1. "Pick the lien" (title Required-fix 1; estate Required-fix 1: record a true
performance deed of trust). The legally correct fix; economically it re-prices the
whole fund.
- It concedes the mortgage frame, and the tax reviewer (D-1, D-4) and estate reviewer (own kill condition) agree the tax/regulatory posture does not survive the concession: the homeowner-side "no income, no interest" pitch dies, and with it origination volume (my A14: CRITICAL at product level). A fund whose asset is originated at one-tenth the projected volume fails on fixed costs alone (operator 4.1).
- Usury becomes a portfolio loss channel, not a footnote: the marketed 11.7% implied yield sits above non-exempt caps in several states, with remedies including interest forfeiture or voiding (tax D-4(d) — STATE-DEPENDENT, adopted). Voided contracts in a recorded-lien book are principal impairments, not disclosure problems.
- Mortgage-recording/doc-stamp taxes on an indeterminate secured amount (title 1.3, tax S-2 — STATE-DEPENDENT, adopted): NY-downstate-style 1.8-2.8% on a 126K commitment is 2,300-3,500 per deal — the ENTIRE origination fee — before the operator's 5,300 cost stack. Deal-level unit economics go from negative to more negative in exactly the high-value coastal markets the sizing tool targets.
- One genuine offset nobody priced: a true recorded lien materially improves the ICA Section 3(c)(5)(C) "qualifying interest" argument that sec-cf 2.2 flags as the load-bearing AMBIGUITY (liens on real estate are the historic core of the exclusion; whether an appreciation-contingent obligation is a loan "fully secured" by real estate remains open — AMBIGUOUS, trending helpful). The fix trades a securities ambiguity for a tax/consumer certainty-of-loss. From the fund's seat that is a bad trade: the ICA problem had a fallback (3(c)(7)); the volume collapse has none.
FX-2. Hold-to-organic-exit restructuring (actuarial fix (ii): evergreen or 25-year
vehicle, re-marketed off the true duration distribution; operator's "30-year
obligation needs a 30-year budget"). Honest, and unfinanceable.
- Fund life: the actuarial mobility model (mean exit ~24yr, 42-56% of positions outstanding past year 25) makes this a 25-30-year vehicle returning, per my B2 grid, 5.7-9.9% gross at 3-5% HPA — before fees, pauses, credit losses, servicing strips. A 2% management fee over 25 years consumes ~50% of committed capital. Net to LPs approaches or undercuts Treasuries with none of the liquidity. There is no LP category for this: insurers fail on RBC (actuarial 3.1), pensions fail on track record (actuarial 3.3), and PE/credit LPs do not underwrite 25-year J-curves at single-digit gross. The fix does not repair the fund; it prices it honestly at zero demand — which is my Stage 1 verdict restated in fund-doc form.
- Evergreen mechanics make it worse: an evergreen needs periodic NAV; these assets have no observable price, sponsor-controlled marks, and a demonstrated wash-trade incentive (red team F11/M3). Auditable NAV does not exist before the secondary market does — the same chicken-and-egg as my S2, relocated into the fund's own financials. Any redemption feature is a run mechanism against unsaleable assets.
FX-3. The consumer-protection stack (retail verdict minimum bar; sec-cf gating items
ii, v-viii; estate fixes 4-5: honest APR/exit-cost disclosure, ratchet and pause
disclosure, all-spouse joinder, capacity screening and counseling, longer rescission,
platform-paid disputes, hardship desk). Each individually correct; jointly they
collapse origination volume and worsen the book.
- Retail's own verdict pt. 4 states the mechanism: honest disclosure removes every HELOC-eligible prospect; what remains is definitionally the no-alternative population. For the fund that is not just lower volume (below operator 4.1 break-even at any plausible funnel per operator 4.2) — it is a WORSE book: my AS1 cash-need selection becomes the whole book, senior-mortgage delinquency (the one full-wipeout path, D1) concentrates, and the correlated-discharge exposure (Dispute 2) deepens. Conversion fixes and credit quality move inversely here; no disclosure regime fixes that.
- Spousal joinder / capacity / counseling add 2-6 weeks and hard attrition to a 30-45 day cycle already losing 30-50% post-application (operator 1.1); CAC per FUNDED deal rises exactly as the fee basis is being cut (FX-4).
FX-4. Operator's economic fixes (servicing strip 25-75bps; origination fee re-based
to home value; pause hysteresis with platform-paid appraisals). Fix the platform by
taxing the fund and the funnel.
- A servicing strip is pure drag on a zero-coupon asset: 90-350/position/yr against positions yielding nothing until exit shaves roughly 30-80bps off deal IRR — the 3%-HPA cells fall from 7.2% toward ~6.5% gross, and the Tier 2 bid (already 0.69-1.10x vs the 1.41x required) drops further because the buyer inherits the strip (operator 6.1).
- Re-basing the fee to home value (incumbent-style 3% of 625K = ~18,750) fixes operator 1.1's per-deal loss by tripling the homeowner's day-one cost — on retail's honest-math page that is a conversion killer, and on early exits it pushes APR-equivalents from 34.7% toward the mid-40s. The platform P&L and the consumer proposition are a strict trade-off at every fee basis; the spec's error was hiding it, not the number chosen.
FX-5. Homeowner counterparty-credit protection (estate fix 6: escrowed or LC-backed
committed capital, springing release on fund default). Destroys the one real
innovation.
Escrowing 84 months of commitments funds 100% of capital day one at cash yields — the staged-deployment IRR advantage (Dispute 4; the only feature that genuinely improves on lump-sum HEIs) disappears, and my B2 grid collapses toward its worst column. An LC on undrawn commitments costs 50-150bps/yr with no revenue line to bear it. And a springing release is my C2 doom-loop with a hair trigger: it converts a funding hiccup into instantaneous book-wide forfeiture. The homeowner deserves this protection — which is the problem: the product's economics only worked by NOT providing it.
FX-6. Pause reform (red team M1; operator 1.4: independent governance, hysteresis,
paused months extend the term, pause-abuse = investor default with forfeiture).
Necessary for the consumer; adverse for the tail. Extending the term for paused months obligates the fund to resume full deployment into homes that have just demonstrated drawdown — anti-selection of capital into the weakest collateral, later in time, with reserves held throughout (my C3). The pause-as-specified was the investor's only credit-risk limiter (actuarial 1.2's own early-cycle math: pause freezes deployment at 6% of commitments); every consumer-fair reform hands that back. There is no parameterization that is simultaneously non-abusive to the homeowner (sec-cf 3.3) and loss-limiting to the fund. This is a design contradiction, not a tuning problem — it belongs on the kill list, not the fix list.
FX-7. Honest offering architecture (sec-cf gating item iii: 506(c) + PPM whose risk
factors match spec Sections 8/16/22/25; actuarial fix (iii): Tier 2 sale as upside
optionality only). Legally clean; commercially self-refuting.
A compliant PPM must state: no secondary market exists (spec Section 8), mean hold ~24 years, 0% nominal return in flat markets for up to 25 years, tax characterization unresolved, bankruptcy treatment SPECULATIVE, fund-manager economics TBD. That document does not raise 50M from anyone the fund can lawfully accept (3(c)(7) QPs per sec-cf 2.3, minus insurers per actuarial 3.1, minus pensions per 3.3). The fix eliminates the fraud by eliminating the fundraise. I flag this not to oppose the fix — it is mandatory — but so the parent process prices its true effect: it is a kill implemented as a disclosure.
FX-8. Duration compensation via a senior fixed-dollar accrual (actuarial fix (iv),
self-flagged). Concur and sharpen: any accrual sufficient to flatten PV-in-exit-year
(their own breakeven: 3.5-5.6pp/yr equivalent) is time-value compensation on
outstanding advances — interest by function (tax D-3, adopted; LIKELY
INTERPRETATION). The structure cannot buy duration protection without buying debt
characterization. This is the same trap as my Stage 1 verdict paragraph: the design
is pinned between its tax posture and its economics, and every fix slides it toward
one wall or the other.
================================================================================
DIRECT CONFLICTS
================================================================================
DC-1. Actuarial (1.1, verdict: "1.0x escrow recovery genuinely robust in sale exits even at -45%; distressed sales almost never breach the floor") vs Title (2.2: waterfall priority unenforceable against third parties; 5.1-5.2: void in TX homestead, defeated by FL homestead at death) and Estate (4.1: non-lien claim discharged in Ch. 7; 8.2: waterfall ignores priming liens). Incompatible because the actuarial recovery model assumes an enforceable, senior-to-costs escrow claim in exactly the sand-state jurisdictions (FL, AZ, NV) and distress events (bankruptcy, death, foreclosure-adjacent sales) where the legal reviewers found the claim unenforceable, dischargeable, or void — and, per my Dispute 1, the arithmetic fails even where the claim IS enforceable.
DC-2. Actuarial (1.1: only material floor-loss mechanism is foreclosure wipeout) vs Estate (4.1, SETTLED LAW: dischargeable claim = sale-independent full-severity loss) and Red team (M4: correlated, industrializable). Incompatible loss models; the Shield sizing and any Tier 2 price both depend on which is right, and estate's is settled law.
DC-3. Operator (verdict: "not a kill — monthly deployment, ride-along decoupling, and the equity-side ratchet are genuinely better-designed than incumbent HEI mechanics"; fixable for 4-7M) vs Tax (verdict: "the defects are not drafting problems; they ARE the product" — the monthly cadence and the 1.0x floor are the two features that destroy the characterization; fixing means removing them). Incompatible because the features the operator proposes to fund and operationalize are the features the tax reviewer requires deleted; both cannot be right about the same clauses.
DC-4. Estate + Title (required fix: record a true security instrument / "pick the lien") vs Tax (D-1: the 1.0x senior secured repayment IS the debt confession; D-4: lien + recharacterization = TILA, usury, licensing cascade) and Sec-cf (3.2: recharacterization voids the arbitration clause and opens 3-year rescission tails). Incompatible fix directions: the bankruptcy/title chassis requires lien status; the tax/regulatory chassis requires its absence. Estate's own kill condition and red team's verdict pt. 1 concede the fix sets are mutually exclusive. This is the central conflict of the whole file.
DC-5. Sec-cf (verdict pt. 1: incumbents operate "under materially similar characterization risk"; core product salvageable with footprint exclusions) vs Tax (A-2: Equity Flow "cannot borrow even the incumbents' unvalidated comfort; strictly worse-positioned") and Title (1.2: the 1.0x floor fails the Roberts loss-sharing test more cleanly than the products already condemned). Incompatible premises about the baseline risk; my Dispute 5 adds that the footprint remedy fails on volume economics.
DC-6. Retail (V2: HELOC cheaper "for essentially all exits up to ~7-10 years in any appreciating market") vs Retail's own V-table and my B2/B6 (at 3% HPA, 7-12 year holds, Equity Flow is cheaper on both dollars and APR). Incompatible with their own numbers; the defensible claim is conditional (short holds / above-trend HPA).
DC-7. Retail (T4 tag: escrow waterfall "LIKELY INTERPRETATION as enforceable") vs Title (2.2: the stated priority order is contradicted by SETTLED closing practice and unenforceable against non-parties). Same clause, opposite tags; title's is the informed one — no closing pays an appreciation share ahead of transfer taxes and commissions.
DC-8. Retail (T5 tag: forced-sale petition viability STATE-DEPENDENT) vs Estate (1.4: probate forced-sale standing for a non-lien holder essentially does not exist — remedy "the courts do not offer," SPECULATIVE as drafted) and Title (4.3: remedy is illusory or self-incriminating). Materially different severity on the same remedy; estate/title's stronger versions are the better-supported ones, and they gut the recovery assumptions any Tier 2 buyer would model.
DC-9. Operator (4.1: platform break-even requires ~750M-1B AUM / 6,000-8,000 positions) vs Sec-cf (2.3: available fund structures cap owners at 100 or QPs-only; the Advisers Act exemption dies past 150M) and Actuarial (3.3: no institutional LPs until seasoning that cannot exist on schedule). Incompatible because the scale operations require is the scale the legal structure and LP market cannot deliver — the platform is squeezed to death between its own reviewers' floors and ceilings.
DC-10. Actuarial (fix (ii): re-cut as an honest 25-year hold-to-maturity vehicle — implicitly financeable) vs My Stage 1 (C1/C4/B5: no LP base exists for 20-30-year illiquid consumer paper at ~6-8% net; that vehicle is the kill scenario, not the fix). Incompatible viability judgments over the identical restructuring; my B2/B5 grid is the pricing, and the actuarial report's own mobility model supplies the duration that makes it unfinanceable.
DC-11. Red team (C2: incumbent monthly-draw wrapper "strictly dominant," priceable at a 15-20% share) vs My B2 and Actuarial 3.2 (staged deployment halves time-weighted capital; the incumbent can only underprice via rated-takeout funding). Incompatible on mechanism: dominance is real but purchased with cheaper capital, not free in the wrapper — matters for anyone modeling the competitive response timeline.
DC-12. Operator (4.1: credits the platform 1.5-2% management fee) vs My O1/A15/B5 (no fee stack exists in the spec; the same undefined fees cannot simultaneously fund the platform and leave LPs the marketed return). Conflict of accounting, not of law: one pro forma or the other is insolvent.
================================================================================
CLOSING POSITION
================================================================================
Verdict unchanged: KILL. Cross-examination moved nothing in the sponsor's favor. The two places I dispute peers hardest — the actuarial reviewer's benign sale-recovery arithmetic (Dispute 1-2) and the retail reviewer's overbroad HELOC claim (Dispute 3) — both resolve, when corrected, into a WORSE picture for the fund: sale exits do breach the floor in the modeled tail, discharge is an unpriced correlated loss channel, and the product's only consumer-fair operating region is precisely its below-cost-of-capital region. The fix lists are pairwise destructive (DC-3, DC-4, DC-10; Part 2 throughout): every repair that satisfies one reviewer detonates a constraint owned by another, and the fixes that are merely expensive are expensive beyond the platform's revenue capacity at any achievable scale (Dispute 5, DC-9). A structure that cannot be simultaneously legal, honest, and financeable is not underspecified — it is overdetermined to fail.
--- end of cross-examination ---