Recorded-Unrecorded Differentials Without a Global Comparator: The FB-and-FC Handover Instance and Capstone of the Execution Program

When two arms of an experiment differ in exactly one respect — here, whether a monitoring record is retained or discarded before retention — three decisions remain after the data arrive: were the arms genuinely comparable, or did an ordinary asymmetry confound the pair; does the measured contrast track the recorded arm in the declared sense; and who may say what the outcome means. Mill's method of difference leaves its central clause — "all other circumstances remaining the same" — to judgment; whoever holds that judgment after data exist is a Global Comparator, the matched-pair authority, and the last post-data seat standing in this sequence's Execution Program (the Experimenter retired at Paper 44, the Statistician at 48, the Falsifier at 49). This paper — the fiftieth in the sequence, capstone of the Execution Program (Papers 48-50) — retires it by construction: it builds the reference handover instance HD for Paper 44's recorded-unrecorded differential on the circuit-QED co-location envelope, at the executable level, and runs Paper 48's acceptance gate on the deposited pair. Delivered: the delta specification from Paper 42's benchmark environment to 44's design (the monitored-register-as-order-control upgrade, the two-regime protocol, the retention-step toggle, the dummy-write hardware) ; the schema completion to field-level primitives; the certification pipeline as executable functions — Paper 44's two lambda-blind pre-execution campaigns as the S1 gate whose failure is a typed pre-tree disposition (44's "no bound" pre-commitments, never a leaf), and the in-situ audits conjoined into the certificate whose ~40-order comparator-blindness margin is frozen content, cited and never recomputed; the scoring pipeline completing the one slot 48 assigned to this paper — a Ramsey delay-sweep readout-to-rate map over a declared, exactly-enforced grid with symmetric SPAM normalization, frozen edge cases, no clipping anywhere, a frozen per-arm post-exclusion N rule, the signed contrast, and the leaf router implementing 48's differential rows verbatim, including the anti-tracking route (subtyped for diagnosis, never split for authority) ; the concrete exclusion rule list with its veto-impact template; the S1-S4 runbook; and the mock-data challenge, executed in three parts — a deterministic core exhaustive over all four dispositions, every frozen edge case, and the production/harness boundary itself; an auditor-generated adversarial set produced through a recorded, hash-deposited exchange by a party that did not implement the scoring code, whose first execution caught a genuine pre-freeze defect, corrected before hash fixation and reported; and a public-beacon-seeded bulk. All three parts pass against hand-derivable or rule-derived expectations; the routing tables are generated by executing the shipped implementation and are typed non-evidence in the sentence that reports them. The gate passes condition by condition under the coverage completion inherited from Paper 49; the verdict is InstanceStatusD = reference-complete with an empty instance gap register, and the Execution Program closes whole: one template, two instances, every leaf of 44's execution map covered by exactly one instance plus the shared non-execution leaf. Displayed at the close is the channel's frozen comparison content: a certified null already fires the carve and the collapse verdict, together, never separately; a certified tracking contrast already licenses distinctness earned and the first FrameworkForcedShift candidacy — declarable only in the data holder's own successor paper, the single reserved cell that is the Comparator's substantive retirement; everything else is the confounded row. Nothing is measured, carved, or scored; no band or feasibility verdict moves; the data papers stay reserved behind mechanical unlocks; reference-complete never shortens to accepted, and an immutable core guarantees no admissible adoption-time extension can change the margin, the sign convention, the anti-tracking route, or a single claim cell. Comparability has audits, tracking has a router, meaning has a table — nothing is left for an authority to judge.