{"judgments":[{"pair_id":"E12MQ019","scores_a":{"structural_fidelity":5,"domain_fidelity":4,"causal_coherence":5,"operational_specificity":5,"testability":5,"practicality":4,"contrivance_risk":2},"scores_b":{"structural_fidelity":5,"domain_fidelity":5,"causal_coherence":5,"operational_specificity":5,"testability":5,"practicality":5,"contrivance_risk":1},"winner":"B","reason":"Both cleanly invert routine audit-work initiation while retaining auditor judgment. B better bounds eligibility and uses a more practical retrospective replay."},{"pair_id":"E12MQ001","scores_a":{"structural_fidelity":5,"domain_fidelity":5,"causal_coherence":5,"operational_specificity":5,"testability":5,"practicality":5,"contrivance_risk":1},"scores_b":{"structural_fidelity":5,"domain_fidelity":4,"causal_coherence":5,"operational_specificity":5,"testability":5,"practicality":3,"contrivance_risk":3},"winner":"A","reason":"A directly addresses a recovery-driven CI herd with health-gated admission, coalescing, fairness, and a discriminating replay. B fits structurally but is a more burdensome retrofit with certification and fixed-sequencing risks."},{"pair_id":"E12MQ005","scores_a":{"structural_fidelity":3,"domain_fidelity":3,"causal_coherence":4,"operational_specificity":5,"testability":5,"practicality":1,"contrivance_risk":5},"scores_b":{"structural_fidelity":4,"domain_fidelity":4,"causal_coherence":5,"operational_specificity":5,"testability":5,"practicality":2,"contrivance_risk":4},"winner":"B","reason":"B more directly changes the withdrawal-and-recording choice set and preserves a verifiable transaction witness. A has a sound occupancy invariant but is an overly narrow, impractical physical substitute for inventory counting."},{"pair_id":"E12MQ013","scores_a":{"structural_fidelity":5,"domain_fidelity":4,"causal_coherence":5,"operational_specificity":5,"testability":5,"practicality":4,"contrivance_risk":2},"scores_b":{"structural_fidelity":5,"domain_fidelity":5,"causal_coherence":5,"operational_specificity":5,"testability":5,"practicality":5,"contrivance_risk":1},"winner":"B","reason":"Both are strong decomposition hypotheses. B more directly addresses a common software entanglement problem with coherent ownership, enforceable boundaries, and a low-risk shadow-first test."},{"pair_id":"E12MQ007","scores_a":{"structural_fidelity":5,"domain_fidelity":5,"causal_coherence":5,"operational_specificity":5,"testability":5,"practicality":5,"contrivance_risk":1},"scores_b":{"structural_fidelity":5,"domain_fidelity":5,"causal_coherence":5,"operational_specificity":5,"testability":4,"practicality":5,"contrivance_risk":1},"winner":"A","reason":"A preserves the relational mechanism while using an equal-content rotating-expert comparator, giving it a substantially stronger falsification of the stable-mentor claim."},{"pair_id":"E12MQ021","scores_a":{"structural_fidelity":3,"domain_fidelity":4,"causal_coherence":5,"operational_specificity":5,"testability":5,"practicality":2,"contrivance_risk":5},"scores_b":{"structural_fidelity":5,"domain_fidelity":5,"causal_coherence":5,"operational_specificity":5,"testability":5,"practicality":4,"contrivance_risk":1},"winner":"B","reason":"B directly builds a bounded, competent, reviewable authority foundation for a real binding registry decision. A demonstrates physical interlocking but cannot establish the social legitimacy and representation it nominally targets."}]}