{"cases":[{"alternative_route_blocks":[],"case_id":"E14A030__DIRECT","case_type":"DIRECT_POSITIVE","domain":"Municipal flood warning operations","intended_route_id":"B1","omitted_condition_id":null,"remedy_leakage_audit":"The statement reports operating history, changed conditions, and consequences without proposing any intervention or naming the underlying solution pattern.","route_evidence":[{"condition_id":"b002_a08_at_a1","intended_status":"SATISFIED","scenario_evidence":"The same river-level trigger previously produced timely, credible warnings but now initiates warnings too early and repeatedly."},{"condition_id":"b002_a08_at_a2","intended_status":"SATISFIED","scenario_evidence":"Upstream development has accelerated runoff, and replacement gauges respond differently to turbulence and debris."},{"condition_id":"b002_a08_at_a3","intended_status":"SATISFIED","scenario_evidence":"Repeated unnecessary warnings consume staff capacity and public attention, while operators are less prepared for the few rapidly developing events."}],"scenario_text":"For more than a decade, a county emergency office has issued river warnings whenever readings at two upstream gauges crossed a fixed level. Historical storms usually produced a gradual rise, so the trigger gave dispatchers enough time to notify low-lying neighborhoods without repeatedly mobilizing them unnecessarily. Over the past three years, extensive upstream development has made runoff faster and more erratic. The county also replaced its aging gauges with devices that react more sharply to brief turbulence and floating debris.\n\nThe same trigger now generates frequent warnings during short spikes that never produce flooding. Dispatchers interrupt other duties, schools prepare buses, and residents receive repeated overnight messages. Public response rates have fallen, and staff increasingly spend the first minutes of each alarm determining whether the reading is credible. Meanwhile, several recent storms rose from ordinary readings to dangerous levels so quickly that the office had little useful lead time. The warning rule remains formally unchanged, but its alerts no longer carry the timing or reliability they once did.","scenario_title":"County River Warnings Lose Their Reliability","vocabulary_separation_audit":"Uses concrete emergency-management language—river gauges, storms, dispatchers, warnings, schools, and residents—without using the archetype name, abstraction labels, or supplied predicate wording."},{"alternative_route_blocks":[],"case_id":"E14A030__TRANSFER","case_type":"TRANSFER_POSITIVE","domain":"Online marketplace fraud review","intended_route_id":"B1","omitted_condition_id":null,"remedy_leakage_audit":"The statement describes observed failures and operational consequences but contains no recommendation, intervention, or solution-oriented language.","route_evidence":[{"condition_id":"b002_a08_at_a1","intended_status":"SATISFIED","scenario_evidence":"A transaction-score cutoff that once separated routine orders from risky ones now classifies many legitimate purchases for review while allowing more coordinated fraud through."},{"condition_id":"b002_a08_at_a2","intended_status":"SATISFIED","scenario_evidence":"The marketplace expanded internationally, changed payment providers, added new merchants, and experienced seasonal volume growth and new fraud behavior."},{"condition_id":"b002_a08_at_a3","intended_status":"SATISFIED","scenario_evidence":"Legitimate orders are delayed, review queues overwhelm analysts, coordinated fraud is missed, and downstream fulfillment and loss decisions deteriorate."}],"scenario_text":"An online marketplace sends an order to its fraud team when a transaction score exceeds 72. When the rule was introduced, most purchases came from domestic repeat customers using a small set of payment methods. The cutoff kept the review queue manageable and captured a useful share of chargeback-prone orders before fulfillment.\n\nThe business has since opened to sellers and buyers in twelve additional countries, adopted a new payment processor, and added merchants offering high-value digital goods. Holiday order volume has also doubled, while organized fraud groups increasingly divide purchases across linked accounts that each appear only moderately risky. The score of 72 remains embedded in the checkout system.\n\nAnalysts now receive large batches of legitimate cross-border orders whose address and device patterns resemble older fraud signals. Many buyers wait hours for release, and some abandon their purchases. At the same time, coordinated account clusters often remain below the cutoff and are fulfilled before their connection becomes visible. Reviewers cannot examine the queue before shipping deadlines, so both customer delays and unrecovered losses have increased.","scenario_title":"Marketplace Review Queue Swells as Trade Expands","vocabulary_separation_audit":"The transfer case shifts from physical hazard warnings to e-commerce transactions and uses distinct vocabulary such as checkout scores, chargebacks, payment processors, account clusters, and fulfillment."},{"alternative_route_blocks":[{"blocked_by_condition_id":"b002_a08_at_a5","condition_set_id":"B2","how_blocked":"The scenario explicitly states that staff apply the score automatically and uniformly, with no discretionary bypass, appeal channel, or parallel handling process."}],"case_id":"E14A030__NEAR_MISS","case_type":"ONE_LITERAL_NEAR_MISS","domain":"Online marketplace fraud review","intended_route_id":"B1","omitted_condition_id":"b002_a08_at_a1","remedy_leakage_audit":"The statement supplies negative evidence and operational facts only; it neither recommends a response nor describes a solution archetype.","route_evidence":[{"condition_id":"b002_a08_at_a1","intended_status":"CONTRADICTED","scenario_evidence":"The score cutoff was launched only after the marketplace had changed and has performed poorly since its first day, so it never had an earlier period of acceptable operational performance."},{"condition_id":"b002_a08_at_a2","intended_status":"SATISFIED","scenario_evidence":"International expansion, a new payment processor, digital-goods merchants, higher volume, and coordinated fraud all changed the operating population relative to the historical data."},{"condition_id":"b002_a08_at_a3","intended_status":"SATISFIED","scenario_evidence":"The cutoff produces unnecessary reviews, customer delays, abandoned purchases, analyst overload, missed coordinated fraud, and increased losses."}],"scenario_text":"An online marketplace launched a fraud-review rule last month: orders with a transaction score above 72 are held before fulfillment. The score had been developed from two-year-old records dominated by domestic repeat buyers and conventional card purchases. Before the rule went live, however, the marketplace had already expanded into twelve additional countries, changed payment processors, added high-value digital-goods merchants, and doubled its seasonal order volume. Fraud groups had also begun spreading purchases across linked accounts.\n\nThe cutoff has performed poorly from its first day in production; there was no earlier period in which it sorted the marketplace’s live orders acceptably. Legitimate cross-border purchases routinely exceed 72 because their addresses, devices, and payment paths resemble patterns in the old records. Analysts cannot clear the resulting holds before fulfillment deadlines, causing delays and abandoned carts. Coordinated account clusters often score below 72 individually and proceed, increasing chargeback losses.\n\nThe checkout system applies the rule automatically to every order. Staff have no discretionary bypass, customers have no appeal channel, and there is no parallel process for handling unusual transactions.","scenario_title":"New Fraud Rule Misfires From Its First Day","vocabulary_separation_audit":"The near miss retains the transfer case’s marketplace setting and comparable operational complexity while explicitly distinguishing a newly deployed, never-successful rule from one whose formerly acceptable performance later changed."}],"experiment_id":"eoa_inverse_innovation_exp14_applicability_retrieval40_20260813","sample_id":"E14A030","schema_version":1}