{"cases":[{"alternative_route_blocks":[{"blocked_by_condition_id":"b037_a07_c5","condition_set_id":"B3","how_blocked":"The evaluation uses one fixed performance score under unchanged eligibility limits, rather than several competing objectives, constraints, or proxy measures."}],"case_id":"E14A031__DIRECT","case_type":"DIRECT_POSITIVE","domain":"industrial polymer formulation","intended_route_id":"B2","omitted_condition_id":null,"remedy_leakage_audit":"The statement reports seed-dependent outcomes and observed trial behavior without recommending any intervention or naming an archetype.","route_evidence":[{"condition_id":"b037_a07_c2","intended_status":"SATISFIED","scenario_evidence":"Identical runs beginning from the legacy and experimental recipes repeatedly finish at materially different performance scores."},{"condition_id":"b037_a07_c3","intended_status":"SATISFIED","scenario_evidence":"The legacy-derived recipe admits no beneficial permitted single adjustment, while a superior recipe differs in coupled dimensions whose isolated changes are rejected."}],"scenario_text":"A manufacturer uses an automated laboratory to refine a heat-resistant coating. Each run begins with an approved recipe, changes one ingredient level or one curing setting, and retains a trial only when a fixed durability score rises. With the same materials, equipment, eligibility limits, and number of trials, runs seeded from the long-used commercial recipe repeatedly finish near 82 points, while runs seeded from an unusual experimental recipe finish near 94. The 82-point recipe cannot be improved by any permitted single adjustment: every nearby trial scores the same or worse. Yet the 94-point recipe has been reproduced and meets all the same limits. Records show that it differs mainly through a coupled resin-ratio and catalyst-setting combination. Changing either component alone from the 82-point recipe lowers the score, although the complete combination scores substantially higher. The engineering group needs to account for why endpoint quality depends so heavily on the initial recipe despite identical evaluation rules and laboratory capability.","scenario_title":"Divergent coating trials from approved recipes","vocabulary_separation_audit":"This positive uses recipes, ingredients, curing settings, and durability measurements; the transfer positive instead uses manuscript fragments, melodic continuity, and archival ordering."},{"alternative_route_blocks":[{"blocked_by_condition_id":"b037_a07_c5","condition_set_id":"B3","how_blocked":"All candidate orders are judged by one fixed fit score, with no competing objectives, constraint tradeoffs, or proxy rankings."}],"case_id":"E14A031__TRANSFER","case_type":"TRANSFER_POSITIVE","domain":"digital reconstruction of medieval music manuscripts","intended_route_id":"B2","omitted_condition_id":null,"remedy_leakage_audit":"The statement gives observed ordering and scoring facts only; it neither recommends a procedure nor names the underlying diagnostic pattern.","route_evidence":[{"condition_id":"b037_a07_c2","intended_status":"SATISFIED","scenario_evidence":"Editors using different initial fragment orders reach stable final reconstructions with scores of 71 and 91 under otherwise identical conditions."},{"condition_id":"b037_a07_c3","intended_status":"SATISFIED","scenario_evidence":"No single-fragment relocation improves the 71-point order, while the superior order differs through interdependent placements whose partial adoption reduces the score."}],"scenario_text":"A digital archive is reconstructing the page order of a fragmented medieval chant manuscript. Editors begin from a proposed sequence and repeatedly relocate one fragment, keeping a revision only when a single fixed fit score increases. The score combines the system's already-established evidence about ink joins and melodic continuity into one published number. Teams have the same fragment scans, scoring rule, edit allowance, and working time. Those beginning from the catalogue's traditional order consistently finish with a 71-point sequence; those beginning from a recently discovered marginal-note ordering finish near 91. In the 71-point sequence, moving any one fragment to any other position leaves the score unchanged or lowers it. The 91-point sequence is nevertheless reproducible from the same pieces. It differs in three interdependent placements, and archived comparisons show that each corresponding placement made alone scores worse than 71 before all three are present. Curators need to explain why equally experienced editors obtain sharply different final reconstructions from different opening orders.","scenario_title":"Conflicting reconstructions of a chant folio","vocabulary_separation_audit":"The manuscript setting and its fragment-ordering language are substantially separated from the formulation vocabulary of the direct case."},{"alternative_route_blocks":[{"blocked_by_condition_id":"b037_a07_c1","condition_set_id":"B1","how_blocked":"Every ordering below the maximum has an allowed one-fragment relocation with a strictly higher score, so progress never stalls below a better ordering."},{"blocked_by_condition_id":"b037_a07_c5","condition_set_id":"B3","how_blocked":"The archive uses exactly one fixed evidentiary score and reports no competing goals, constraint tradeoffs, or proxy measures."}],"case_id":"E14A031__NEAR_MISS","case_type":"ONE_LITERAL_NEAR_MISS","domain":"digital reconstruction of medieval music manuscripts","intended_route_id":"B2","omitted_condition_id":"b037_a07_c3","remedy_leakage_audit":"The statement supplies explicit negative structural evidence and a time-limit explanation without proposing any corrective action or naming an archetype.","route_evidence":[{"condition_id":"b037_a07_c2","intended_status":"SATISFIED","scenario_evidence":"Within the fixed twelve-edit session, distant and close initial orders finish at substantially different scores."},{"condition_id":"b037_a07_c3","intended_status":"CONTRADICTED","scenario_evidence":"Complete enumeration establishes that every nonmaximum order has a strictly improving permitted relocation, no equal-score shelf exists, and successive improvements reach the unique maximum without a worsening or multi-fragment step."}],"scenario_text":"A digital archive is reconstructing the page order of a second fragmented medieval chant manuscript. Editors start from supplied sequences and may make twelve one-fragment relocations, retaining a move only when the archive's single fixed evidentiary score rises. All editors use identical scans, rules, and working time. A starting order already close to the best known sequence usually finishes near 96 points, while a distant catalogue order finishes near 78 within the twelve-edit allowance, so the opening sequence strongly predicts the quality delivered at the deadline. However, this manuscript's scoring table has been completely enumerated. For every ordering below 96, at least one permitted one-fragment relocation produces a strictly higher score. There are no stretches of equal-scoring orders, and following improving relocations from any starting order eventually reaches the unique 96-point sequence without any temporary decrease or coordinated movement of several fragments. The archive uses no secondary ranking, competing objective, or changing constraint. Managers need to explain the large deadline-time disparity while respecting these verified properties of the ordering table.","scenario_title":"Deadline disparity in a fully enumerated folio","vocabulary_separation_audit":"The near miss retains the transfer case's archival vocabulary and comparable reconstruction difficulty; its distinguishing evidence concerns continuous strict improvement rather than introducing direct-case formulation terms."}],"experiment_id":"eoa_inverse_innovation_exp14_applicability_retrieval40_20260813","sample_id":"E14A031","schema_version":1}