Tail Calibration Failure

The Sanctuary Discount's mean-trajectory calibration succeeds for modal-distribution events and fails catastrophically when operational deployment generates tail events that exceed the discounted-announcement's risk envelope; the rebound is amplified by the prior under-pricing as every dependency the modal discount entered is repriced simultaneously. The empirical signature is Brent $77.40 (Monday open, 2026-05-04) → $116.55 (Tuesday intraday peak, 2026-05-05) with $107.38 CME settlement after Hegseth's threshold-framing reassurance — a roughly 50% surge within twenty-four hours of the discount's operation. The structural mechanism: the marketplace's calibrated discount is constructed against the modal trajectory of accumulated weekend-window announcements without a separate tail-risk channel, so when the operational tail materializes the discount's prior under-pricing is repriced through every position that had embedded it. Cross-references META-1 Coupling Failure (the discount-architecture is decoupled from the tail-risk channel that would have constrained it) and META-3 Threshold Cascade (the rebound itself is a cascade through every dependency the prior under-pricing entered). Recursive structural finding: the pattern instantiates the Cycle 1 audit's S1 LLM Cognitive Signature (mean-trajectory regression) at marketplace level — a homology, not an identity. The operation by which an LLM averages historical analogues to emit modal completion and the operation by which the marketplace prices the modal trajectory of weekend-window announcements are the same operation at different scales, and they fail in the same way when the world is not modal.