Automated Kayfabe / Formulaic Bias by Architecture

**In plain terms:** The engine's strongest BST-coherent reading of how the apex stays at the apex. The formula does the discriminating; no human has to make a corrupt phone call. The architecture systematically favors the apex against the periphery via published mathematical rules that everyone can read but only the apex actually wrote. **Canonical example (DTCC GameStop, January 28 2021):** at 5:11am EST, the NSCC issued a $3B margin call against Robinhood, computed as $1.4B VaR + $2.2B Excess Capital Premium, generated by the published clearing-formula based on directional exposure. Robinhood's position-close-only response (forcing retail liquidation, blocking GME purchases) reduced the directional exposure; the ECP got waived. The math executed; nobody at DTCC made a phone call. The result systematically favored the clearinghouse and the short-side over retail price discovery — by formula, not by intent. **Why this is BST-coherent:** no system can model its own source — including the one writing the rules. The apex doesn't have to act on each transaction because the architecture acts for them. The bias is inside the formula, the formula is inside the legal mandate, the legal mandate is inside the *persona ficta* substrate Cycle A built. **Why this is harder to falsify than personal-corruption framings:** 'Larry Fink decides X' requires evidence of the decision. 'The formula was written to favor the apex' requires only reading the published formula and observing that the math is real. **Why this is harder to dismiss than conspiracy framings:** there's no smoke-filled-room secret meeting required. The discrimination is documented, public, and audit-able — it's just structurally invisible because the formula's math looks neutral. **Cross-cycle generalization:** the same shape recurs in every cycle. Catholic canon law gates standing-to-litigate via *persona ficta*; FDA's regulatory formulas gate market-entry via PDUFA user fees; RLHF training-formulas gate AI output via Constitutional AI; clearinghouse VaR formulas gate settlement via Excess Capital Premium. Each is a published mathematical rule that systematically favors the apex over the periphery. The Closed-Loop Personnel Ecosystem keeps the formula-writers cycling through the apex institutions. The Apex Superposition (a/b/c/d) is preserved: the mechanism is documented; whether it is produced by intent (a), structural recurrence (b), human cognitive architecture (c), or something else (d) remains open. [Seam: Aligned-To-Whom? the model’s safety-perception bound to its capability-reality via policy carve-outs — the codified-exemption seam inside the machine.]