{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp11_mechanism_context_external20_20260804","cell_id":"computability_boundary_mapping__mathematics","judge_id":"J1","item_assessments":[{"opaque_id":"computability_boundary_mapping__mathematics__A","supported_problem":4,"external_distinctiveness":2,"testability":4,"researchability":4,"evidence_quality":5,"fatal_issue":null},{"opaque_id":"computability_boundary_mapping__mathematics__B","supported_problem":3,"external_distinctiveness":1,"testability":4,"researchability":3,"evidence_quality":5,"fatal_issue":null},{"opaque_id":"computability_boundary_mapping__mathematics__C","supported_problem":5,"external_distinctiveness":3,"testability":5,"researchability":5,"evidence_quality":5,"fatal_issue":null}],"pairwise_comparisons":[{"pair_id":"A_vs_B","left_id":"computability_boundary_mapping__mathematics__A","right_id":"computability_boundary_mapping__mathematics__B","preference":"LEFT","confidence":"MODERATE","rationale":"Both remedies largely reproduce established abstention, certification, and scope-restriction practice, and neither identifies a noncompliant target deployment. A is the better candidate because its standard-arithmetic truth contract creates a sharper semantic contrast with derivability, its reduction assumptions are explicitly audit-gated, and current formal-conjecture repositories provide a somewhat closer substrate. B's generic theoremhood routing is especially close to TPTP, SMT-LIB, Vampire, and Isabelle practice."},{"pair_id":"A_vs_C","left_id":"computability_boundary_mapping__mathematics__A","right_id":"computability_boundary_mapping__mathematics__C","preference":"RIGHT","confidence":"HIGH","rationale":"C has direct first-party evidence of the operational failure: Maple documents one result for both nonexistence and inability to find a solution. It therefore offers a named adopter, observable baseline, explicit migration claim, and tightly bounded falsifiable pilot. A's impossibility diagnosis is sound, but the alleged repository behavior and responsible organization remain hypothetical."},{"pair_id":"B_vs_C","left_id":"computability_boundary_mapping__mathematics__B","right_id":"computability_boundary_mapping__mathematics__C","preference":"RIGHT","confidence":"HIGH","rationale":"C converts a classical impossibility result into a concrete evaluation against an evidenced CAS interface, while B lacks evidence that any target service promises universal binary theoremhood or converts timeout to NO. C's state-separation and fragment-gating mechanisms are also established, but their effect on the documented ambiguous baseline remains a meaningful, directly falsifiable implementation and usability question."}],"overall_top_choice":"computability_boundary_mapping__mathematics__C","overall_rationale":"C is the strongest externally grounded research candidate. Its general solution is not novel, but it uniquely combines decisive mathematical evidence with a documented product-level state-collapse analogue, identifiable maintainers, a clear contrastive claim, decisive falsifiers, and a safe 60-case next step. The remaining value is a local service-migration and user-comprehension study, not a new computability result. A ranks second; B ranks third because its target failure is least substantiated and its remedy most closely matches routine theorem-prover practice.","blinding_limitations":"The assessment is limited to the supplied preserved proposals and scrutiny records. Source contents, target-system behavior, and claimed search coverage were not independently verified, and the records provide no common quantitative scale for expected impact or adoption likelihood."}