{"cases":[{"alternative_route_blocks":[],"case_id":"E14A024__DIRECT","case_type":"DIRECT_POSITIVE","domain":"Refrigerated food logistics","intended_route_id":"B1","omitted_condition_id":null,"remedy_leakage_audit":"The statement presents observations and consequences only; it contains no proposed intervention, recommendation, or named solution pattern.","route_evidence":[{"condition_id":"b023_a06_c1","intended_status":"SATISFIED","scenario_evidence":"Independent product measurements show humidity and spoilage moving progressively farther from the warehouse's preservation goals."},{"condition_id":"b023_a06_c2","intended_status":"SATISFIED","scenario_evidence":"Each controller adjustment intensifies dehumidification without improving actual conditions."},{"condition_id":"b023_a06_c3","intended_status":"SATISFIED","scenario_evidence":"A repeatable week-over-week movement replaces ordinary aisle and shipment fluctuations."},{"condition_id":"b023_a06_c4","intended_status":"SATISFIED","scenario_evidence":"Ice accumulation distorts later readings, lengthens compressor cycles, and magnifies subsequent control errors."}],"scenario_text":"A regional cold-storage operator holds berries from several farms before distribution. Its control system is tuned to keep package-level humidity within a narrow range, using readings from wall-mounted sensors to adjust refrigeration and dehumidification. Since new high-moisture packaging was introduced, the wall readings have continued to look favorable, yet independent checks inside sealed cartons show humidity rising farther above specification each week. Mold losses and rejected pallets are also climbing.\n\nThe controller responds to every favorable wall reading by extending dehumidifier cycles, but those longer cycles create colder surfaces and more condensation inside the new packages. Ordinary differences among aisles have given way to a consistent week-over-week shift toward wetter cartons. Ice accumulating around vents then alters airflow and makes later sensor readings even less representative. Each cycle therefore begins with a larger measurement discrepancy, triggers a stronger adjustment, and leaves more ice and condensation for the next cycle. The warehouse remains busy and tightly monitored, while the stored product moves steadily away from its preservation requirements.","scenario_title":"Cold-Storage Readings Improve as Carton Spoilage Rises","vocabulary_separation_audit":"This case uses physical logistics, refrigeration, packaging, humidity, airflow, and spoilage language; it is distinct from the recruitment and advertising vocabulary of the transfer cases."},{"alternative_route_blocks":[],"case_id":"E14A024__TRANSFER","case_type":"TRANSFER_POSITIVE","domain":"Scholarship recruitment advertising","intended_route_id":"B1","omitted_condition_id":null,"remedy_leakage_audit":"The statement diagnoses campaign behavior without recommending actions, naming an intervention, or describing a solution pattern.","route_evidence":[{"condition_id":"b023_a06_c1","intended_status":"SATISFIED","scenario_evidence":"Verified applications from the intended counties fall progressively farther below the campaign goal."},{"condition_id":"b023_a06_c2","intended_status":"SATISFIED","scenario_evidence":"Weekly bid and audience changes raise reported engagement while further reducing verified applications."},{"condition_id":"b023_a06_c3","intended_status":"SATISFIED","scenario_evidence":"The campaign mix moves consistently toward low-quality placements rather than fluctuating randomly."},{"condition_id":"b023_a06_c4","intended_status":"SATISFIED","scenario_evidence":"Contaminated engagement records enter repeated retraining rounds and increasingly distort later allocation decisions."}],"scenario_text":"A scholarship foundation runs a digital campaign intended to increase completed applications from students in twelve rural counties. Its bidding model was trained when link clicks closely tracked verified application starts. During the current season, several mobile content networks began generating large volumes of cheap engagement from outside the eligible area. Dashboard click-through rates now improve every week, while county-verified applications fall progressively farther below the foundation's enrollment goal.\n\nCampaign managers respond to disappointing application totals with weekly bid and audience changes. Those changes repeatedly send more money toward the placements reporting the strongest click activity, so the next week's verified yield is worse despite apparently stronger engagement. What began as day-to-day variation has become a sustained migration of spending toward the same low-quality networks. The platform also feeds each round of engagement back into model training. As ineligible and automated clicks accumulate in the training history, later forecasts become more confident in the least productive placements, prompting larger reallocations and widening the shortfall after every campaign cycle.","scenario_title":"Rising Click Metrics Accompany a Growing Scholarship Shortfall","vocabulary_separation_audit":"The transfer case replaces industrial sensing and refrigeration language with scholarship eligibility, ad bidding, audience placement, engagement, and application vocabulary."},{"alternative_route_blocks":[],"case_id":"E14A024__NEAR_MISS","case_type":"ONE_LITERAL_NEAR_MISS","domain":"Scholarship recruitment advertising","intended_route_id":"B1","omitted_condition_id":"b023_a06_c1","remedy_leakage_audit":"The statement contains only diagnostic facts, including natural negative evidence for the omitted condition; it offers no remedy or intervention.","route_evidence":[{"condition_id":"b023_a06_c1","intended_status":"CONTRADICTED","scenario_evidence":"Audited applications remain at the same fixed shortfall every week and are explicitly reported as not moving farther from the goal."},{"condition_id":"b023_a06_c2","intended_status":"SATISFIED","scenario_evidence":"Repeated bid and audience revisions fail to change the audited application result."},{"condition_id":"b023_a06_c3","intended_status":"SATISFIED","scenario_evidence":"Spending and recorded engagement move persistently toward low-quality networks."},{"condition_id":"b023_a06_c4","intended_status":"SATISFIED","scenario_evidence":"Successive training rounds accumulate mislabeled engagement and produce increasingly distorted forecasts."}],"scenario_text":"A scholarship foundation advertises to students in twelve rural counties and evaluates the campaign against a weekly goal for completed, eligibility-verified applications. Its bidding model was trained when clicks were a dependable sign of an application start. This season, mobile content networks are supplying growing volumes of cheap activity from outside the eligible area. Reported engagement rises, but audits show exactly 240 verified applications each week against a goal of 300. The shortfall has remained fixed at 60 for eight consecutive weeks; it is explicitly not growing.\n\nManagers have made repeated changes to bids, audience filters, and placement weights, yet none has altered the audited application total. Meanwhile, spending has shifted consistently toward the networks producing the most questionable activity rather than oscillating among sources. Each week's engagement records are added to the model's training history. The resulting buildup of ineligible and automated clicks makes successive forecasts more skewed and more confident, and the campaign's internal allocation records become increasingly detached from applicant quality, even though its measured gap from the weekly application goal remains unchanged.","scenario_title":"Advertising Records Deteriorate While the Application Gap Holds Steady","vocabulary_separation_audit":"The near miss retains the transfer case's recruitment, eligibility, bidding, engagement, and retraining vocabulary while avoiding the cold-storage case's physical-process terminology."}],"experiment_id":"eoa_inverse_innovation_exp14_applicability_retrieval40_20260813","sample_id":"E14A024","schema_version":1}