{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp04_retrieval_first_paired20_20260802","cell_id":"computability_boundary_mapping__public_administration_policy","considered_ids":["H1","H4","H5"],"selected_id":"H5","dispositions":[{"hypothesis_id":"H1","lane":"COLLISION","reason":"The core mechanism substantially collides with Cedar, Rego, Catala, OpenFisca, and DMN. Its residual claim combines several unvalidated requirements—benefits-domain deployment, legislative-correctness certification, and 80% historical-construct retention—rather than isolating one well-supported distinction."},{"hypothesis_id":"H4","lane":"REJECT","reason":"OpenFisca already implements Cartesian-product tax-benefit simulations, while NIST establishes constrained exhaustive testing and coverage measurement. The remaining integration claim is weakly differentiated and lacks support for the proposed discovery-rate benefit."},{"hypothesis_id":"H5","lane":"FINALIST","reason":"Existing sources directly establish procurement assurance cases, contractual gates, independent review, and change-triggered reassessment, leaving a narrow, falsifiable residual: a mandatory computability-specific subcase for exact terminating guarantees. This is a material and potentially repairable distinction despite only moderate evidence of the recurring procurement problem."}],"selection_reason":"H5 offers the clearest differentiated opportunity: it adds one bounded, testable computability-classification obligation to a recognizable and directly sourced public-procurement assurance workflow. H1 and H4 depend on broader combinations whose principal mechanisms already have substantial close analogues."}