{"generated_at":"2026-10-07T01:39:09.448584+00:00","date":"2026-10-06","totals":{"cloud_requests":2080,"local_predictions":480,"known_cloud_cost_usd":0.44000975000000087,"unknown_cloud_charges":0,"cloud_failures":30,"preflight_cost_usd":1.23e-05},"typed":[{"cohort":"transfer","variant":"archive-jev-zero","historical":true,"n":160,"correct":113,"oracle_correct":113,"failures":0,"posterior":{"mean":0.7037037037037037,"low":0.6313448698431301,"high":0.7712987720707516,"prob_above_bar":2.5093413697136984e-05,"jeffreys_mean":0.7049689440993789,"a":114,"b":48},"oracle_posterior":{"mean":0.7037037037037037,"low":0.6313448698431301,"high":0.7712987720707516,"prob_above_bar":2.5093413697136984e-05,"jeffreys_mean":0.7049689440993789,"a":114,"b":48},"cost_usd":0.028876260000000025,"known_cost_usd":0.028876260000000025,"unknown_costs":0,"median_ms":338.381,"p95_ms":437.8461999999999,"missing_latency":0,"tasks":{"approval_policy":{"n":40,"correct":30,"oracle_correct":30,"failures":0,"posterior":{"mean":0.7380952380952381,"low":0.5969537633763495,"high":0.8577869966591287,"prob_above_bar":0.06820156519850526,"jeffreys_mean":0.7439024390243902,"a":31,"b":11},"oracle_posterior":{"mean":0.7380952380952381,"low":0.5969537633763495,"high":0.8577869966591287,"prob_above_bar":0.06820156519850526,"jeffreys_mean":0.7439024390243902,"a":31,"b":11},"cost_usd":0.0064210440000000025,"known_cost_usd":0.0064210440000000025,"unknown_costs":0,"median_ms":327.119,"p95_ms":367.41904999999997,"missing_latency":0},"project_selection":{"n":40,"correct":3,"oracle_correct":3,"failures":0,"posterior":{"mean":0.09523809523809523,"low":0.027227466798360835,"high":0.19924597837666297,"prob_above_bar":1.6871847800495888e-26,"jeffreys_mean":0.08536585365853659,"a":4,"b":38},"oracle_posterior":{"mean":0.09523809523809523,"low":0.027227466798360835,"high":0.19924597837666297,"prob_above_bar":1.6871847800495888e-26,"jeffreys_mean":0.08536585365853659,"a":4,"b":38},"cost_usd":0.013521774,"known_cost_usd":0.013521774,"unknown_costs":0,"median_ms":411.3465,"p95_ms":491.4419,"missing_latency":0},"stateful_config":{"n":40,"correct":40,"oracle_correct":40,"failures":0,"posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"oracle_posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"cost_usd":0.005747406000000001,"known_cost_usd":0.005747406000000001,"unknown_costs":0,"median_ms":325.58799999999997,"p95_ms":396.7456499999999,"missing_latency":0},"ticket_triage":{"n":40,"correct":40,"oracle_correct":40,"failures":0,"posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"oracle_posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"cost_usd":0.0031860360000000006,"known_cost_usd":0.0031860360000000006,"unknown_costs":0,"median_ms":328.52049999999997,"p95_ms":380.2961999999999,"missing_latency":0}},"gate":{"threshold":0.66,"calibration_accepted":29,"accepted":108,"errors":0,"n":160,"posterior":{"mean":0.990909090909091,"low":0.9667233352014916,"high":0.9997677535056477,"prob_above_bar":0.9999999976598474,"jeffreys_mean":0.9954128440366973,"a":109,"b":1}},"risk_curve":[{"threshold":1.0,"accepted":28,"errors":0,"coverage":0.175,"risk":0.0},{"threshold":0.99,"accepted":43,"errors":0,"coverage":0.26875,"risk":0.0},{"threshold":0.98,"accepted":57,"errors":0,"coverage":0.35625,"risk":0.0},{"threshold":0.97,"accepted":61,"errors":0,"coverage":0.38125,"risk":0.0},{"threshold":0.96,"accepted":63,"errors":0,"coverage":0.39375,"risk":0.0},{"threshold":0.95,"accepted":67,"errors":0,"coverage":0.41875,"risk":0.0},{"threshold":0.94,"accepted":73,"errors":0,"coverage":0.45625,"risk":0.0},{"threshold":0.93,"accepted":78,"errors":0,"coverage":0.4875,"risk":0.0},{"threshold":0.92,"accepted":79,"errors":0,"coverage":0.49375,"risk":0.0},{"threshold":0.91,"accepted":82,"errors":0,"coverage":0.5125,"risk":0.0},{"threshold":0.9,"accepted":83,"errors":0,"coverage":0.51875,"risk":0.0},{"threshold":0.89,"accepted":85,"errors":0,"coverage":0.53125,"risk":0.0},{"threshold":0.88,"accepted":86,"errors":0,"coverage":0.5375,"risk":0.0},{"threshold":0.87,"accepted":87,"errors":0,"coverage":0.54375,"risk":0.0},{"threshold":0.85,"accepted":88,"errors":0,"coverage":0.55,"risk":0.0},{"threshold":0.83,"accepted":89,"errors":0,"coverage":0.55625,"risk":0.0},{"threshold":0.82,"accepted":91,"errors":0,"coverage":0.56875,"risk":0.0},{"threshold":0.81,"accepted":92,"errors":0,"coverage":0.575,"risk":0.0},{"threshold":0.79,"accepted":94,"errors":0,"coverage":0.5875,"risk":0.0},{"threshold":0.78,"accepted":96,"errors":0,"coverage":0.6,"risk":0.0},{"threshold":0.77,"accepted":97,"errors":0,"coverage":0.60625,"risk":0.0},{"threshold":0.76,"accepted":100,"errors":0,"coverage":0.625,"risk":0.0},{"threshold":0.75,"accepted":102,"errors":0,"coverage":0.6375,"risk":0.0},{"threshold":0.74,"accepted":103,"errors":0,"coverage":0.64375,"risk":0.0},{"threshold":0.72,"accepted":104,"errors":0,"coverage":0.65,"risk":0.0},{"threshold":0.71,"accepted":105,"errors":0,"coverage":0.65625,"risk":0.0},{"threshold":0.68,"accepted":106,"errors":0,"coverage":0.6625,"risk":0.0},{"threshold":0.66,"accepted":108,"errors":0,"coverage":0.675,"risk":0.0},{"threshold":0.64,"accepted":109,"errors":1,"coverage":0.68125,"risk":0.009174311926605505},{"threshold":0.63,"accepted":110,"errors":1,"coverage":0.6875,"risk":0.00909090909090909},{"threshold":0.61,"accepted":111,"errors":1,"coverage":0.69375,"risk":0.009009009009009009},{"threshold":0.6,"accepted":112,"errors":1,"coverage":0.7,"risk":0.008928571428571428},{"threshold":0.59,"accepted":114,"errors":3,"coverage":0.7125,"risk":0.02631578947368421},{"threshold":0.5757575757575757,"accepted":115,"errors":4,"coverage":0.71875,"risk":0.034782608695652174},{"threshold":0.5700000000000001,"accepted":116,"errors":5,"coverage":0.725,"risk":0.04310344827586207},{"threshold":0.54,"accepted":118,"errors":5,"coverage":0.7375,"risk":0.0423728813559322},{"threshold":0.52,"accepted":119,"errors":6,"coverage":0.74375,"risk":0.05042016806722689},{"threshold":0.51,"accepted":121,"errors":8,"coverage":0.75625,"risk":0.06611570247933884},{"threshold":0.5,"accepted":125,"errors":12,"coverage":0.78125,"risk":0.096},{"threshold":0.48,"accepted":126,"errors":13,"coverage":0.7875,"risk":0.10317460317460317},{"threshold":0.45,"accepted":127,"errors":14,"coverage":0.79375,"risk":0.11023622047244094},{"threshold":0.44,"accepted":128,"errors":15,"coverage":0.8,"risk":0.1171875},{"threshold":0.42,"accepted":129,"errors":16,"coverage":0.80625,"risk":0.12403100775193798},{"threshold":0.4141414141414141,"accepted":130,"errors":17,"coverage":0.8125,"risk":0.13076923076923078},{"threshold":0.41000000000000003,"accepted":132,"errors":19,"coverage":0.825,"risk":0.14393939393939395},{"threshold":0.41,"accepted":133,"errors":20,"coverage":0.83125,"risk":0.15037593984962405},{"threshold":0.404040404040404,"accepted":134,"errors":21,"coverage":0.8375,"risk":0.15671641791044777},{"threshold":0.4,"accepted":136,"errors":23,"coverage":0.85,"risk":0.16911764705882354},{"threshold":0.39999999999999997,"accepted":137,"errors":24,"coverage":0.85625,"risk":0.17518248175182483},{"threshold":0.39,"accepted":138,"errors":25,"coverage":0.8625,"risk":0.18115942028985507},{"threshold":0.38,"accepted":139,"errors":26,"coverage":0.86875,"risk":0.18705035971223022},{"threshold":0.37,"accepted":141,"errors":28,"coverage":0.88125,"risk":0.19858156028368795},{"threshold":0.32999999999999996,"accepted":142,"errors":29,"coverage":0.8875,"risk":0.20422535211267606},{"threshold":0.32,"accepted":143,"errors":30,"coverage":0.89375,"risk":0.2097902097902098},{"threshold":0.31999999999999995,"accepted":144,"errors":31,"coverage":0.9,"risk":0.2152777777777778},{"threshold":0.31,"accepted":145,"errors":32,"coverage":0.90625,"risk":0.2206896551724138},{"threshold":0.2828282828282829,"accepted":146,"errors":33,"coverage":0.9125,"risk":0.22602739726027396},{"threshold":0.26,"accepted":148,"errors":35,"coverage":0.925,"risk":0.23648648648648649},{"threshold":0.24242424242424246,"accepted":149,"errors":36,"coverage":0.93125,"risk":0.24161073825503357},{"threshold":0.23,"accepted":152,"errors":39,"coverage":0.95,"risk":0.2565789473684211},{"threshold":0.21000000000000002,"accepted":154,"errors":41,"coverage":0.9625,"risk":0.2662337662337662},{"threshold":0.19999999999999996,"accepted":155,"errors":42,"coverage":0.96875,"risk":0.2709677419354839},{"threshold":0.1919191919191919,"accepted":156,"errors":43,"coverage":0.975,"risk":0.27564102564102566},{"threshold":0.19,"accepted":157,"errors":44,"coverage":0.98125,"risk":0.2802547770700637},{"threshold":0.18000000000000002,"accepted":158,"errors":45,"coverage":0.9875,"risk":0.2848101265822785},{"threshold":0.18,"accepted":159,"errors":46,"coverage":0.99375,"risk":0.2893081761006289},{"threshold":0.1414141414141414,"accepted":160,"errors":47,"coverage":1.0,"risk":0.29375}]},{"cohort":"transfer","variant":"archive-luna-zero","historical":true,"n":160,"correct":159,"oracle_correct":159,"failures":0,"posterior":{"mean":0.9876543209876543,"low":0.965880874845245,"high":0.9984920381266038,"prob_above_bar":0.9999999999940713,"jeffreys_mean":0.9906832298136646,"a":160,"b":2},"oracle_posterior":{"mean":0.9876543209876543,"low":0.965880874845245,"high":0.9984920381266038,"prob_above_bar":0.9999999999940713,"jeffreys_mean":0.9906832298136646,"a":160,"b":2},"cost_usd":0.044340200000000024,"known_cost_usd":0.044340200000000024,"unknown_costs":0,"median_ms":1263.6695,"p95_ms":2613.7951999999987,"missing_latency":0,"tasks":{"approval_policy":{"n":40,"correct":40,"oracle_correct":40,"failures":0,"posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"oracle_posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"cost_usd":0.010808,"known_cost_usd":0.010808,"unknown_costs":0,"median_ms":1493.4875000000002,"p95_ms":1944.7879499999995,"missing_latency":0},"project_selection":{"n":40,"correct":39,"oracle_correct":39,"failures":0,"posterior":{"mean":0.9523809523809523,"low":0.8714459796158537,"high":0.9940368817589572,"prob_above_bar":0.9947837699178925,"jeffreys_mean":0.9634146341463414,"a":40,"b":2},"oracle_posterior":{"mean":0.9523809523809523,"low":0.8714459796158537,"high":0.9940368817589572,"prob_above_bar":0.9947837699178925,"jeffreys_mean":0.9634146341463414,"a":40,"b":2},"cost_usd":0.016797199999999998,"known_cost_usd":0.016797199999999998,"unknown_costs":0,"median_ms":2041.201,"p95_ms":3202.0152,"missing_latency":0},"stateful_config":{"n":40,"correct":40,"oracle_correct":40,"failures":0,"posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"oracle_posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"cost_usd":0.009196200000000003,"known_cost_usd":0.009196200000000003,"unknown_costs":0,"median_ms":1079.7514999999999,"p95_ms":1374.259649999999,"missing_latency":0},"ticket_triage":{"n":40,"correct":40,"oracle_correct":40,"failures":0,"posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"oracle_posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"cost_usd":0.007538799999999999,"known_cost_usd":0.007538799999999999,"unknown_costs":0,"median_ms":1139.103,"p95_ms":1476.5196499999995,"missing_latency":0}},"gate":null,"risk_curve":[]},{"cohort":"transfer","variant":"archive-luna-taught","historical":true,"n":160,"correct":159,"oracle_correct":159,"failures":0,"posterior":{"mean":0.9876543209876543,"low":0.965880874845245,"high":0.9984920381266038,"prob_above_bar":0.9999999999940713,"jeffreys_mean":0.9906832298136646,"a":160,"b":2},"oracle_posterior":{"mean":0.9876543209876543,"low":0.965880874845245,"high":0.9984920381266038,"prob_above_bar":0.9999999999940713,"jeffreys_mean":0.9906832298136646,"a":160,"b":2},"cost_usd":0.06951739999999998,"known_cost_usd":0.06951739999999998,"unknown_costs":0,"median_ms":1300.182,"p95_ms":2953.77265,"missing_latency":0,"tasks":{"approval_policy":{"n":40,"correct":40,"oracle_correct":40,"failures":0,"posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"oracle_posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"cost_usd":0.01546,"known_cost_usd":0.01546,"unknown_costs":0,"median_ms":1549.9470000000001,"p95_ms":1922.3335499999998,"missing_latency":0},"project_selection":{"n":40,"correct":39,"oracle_correct":39,"failures":0,"posterior":{"mean":0.9523809523809523,"low":0.8714459796158537,"high":0.9940368817589572,"prob_above_bar":0.9947837699178925,"jeffreys_mean":0.9634146341463414,"a":40,"b":2},"oracle_posterior":{"mean":0.9523809523809523,"low":0.8714459796158537,"high":0.9940368817589572,"prob_above_bar":0.9947837699178925,"jeffreys_mean":0.9634146341463414,"a":40,"b":2},"cost_usd":0.0274404,"known_cost_usd":0.0274404,"unknown_costs":0,"median_ms":1977.321,"p95_ms":3461.0818,"missing_latency":0},"stateful_config":{"n":40,"correct":40,"oracle_correct":40,"failures":0,"posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"oracle_posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"cost_usd":0.015085400000000004,"known_cost_usd":0.015085400000000004,"unknown_costs":0,"median_ms":1121.9850000000001,"p95_ms":1476.1362999999997,"missing_latency":0},"ticket_triage":{"n":40,"correct":40,"oracle_correct":40,"failures":0,"posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"oracle_posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"cost_usd":0.011531600000000001,"known_cost_usd":0.011531600000000001,"unknown_costs":0,"median_ms":1100.9569999999999,"p95_ms":1415.10575,"missing_latency":0}},"gate":null,"risk_curve":[]},{"cohort":"transfer","variant":"archive-jev-taught","historical":true,"n":160,"correct":114,"oracle_correct":114,"failures":0,"posterior":{"mean":0.7098765432098766,"low":0.6378775325055042,"high":0.7769675646523236,"prob_above_bar":5.40936680759272e-05,"jeffreys_mean":0.7111801242236024,"a":115,"b":47},"oracle_posterior":{"mean":0.7098765432098766,"low":0.6378775325055042,"high":0.7769675646523236,"prob_above_bar":5.40936680759272e-05,"jeffreys_mean":0.7111801242236024,"a":115,"b":47},"cost_usd":0.037375380000000034,"known_cost_usd":0.037375380000000034,"unknown_costs":0,"median_ms":384.7895,"p95_ms":485.3124999999999,"missing_latency":0,"tasks":{"approval_policy":{"n":40,"correct":25,"oracle_correct":25,"failures":0,"posterior":{"mean":0.6190476190476191,"low":0.46936254803283345,"high":0.7579890328213116,"prob_above_bar":0.0005349652670281795,"jeffreys_mean":0.6219512195121951,"a":26,"b":16},"oracle_posterior":{"mean":0.6190476190476191,"low":0.46936254803283345,"high":0.7579890328213116,"prob_above_bar":0.0005349652670281795,"jeffreys_mean":0.6219512195121951,"a":26,"b":16},"cost_usd":0.008166564000000001,"known_cost_usd":0.008166564000000001,"unknown_costs":0,"median_ms":387.148,"p95_ms":485.24235,"missing_latency":0},"project_selection":{"n":40,"correct":9,"oracle_correct":9,"failures":0,"posterior":{"mean":0.23809523809523808,"low":0.12363248839584409,"high":0.3761374654340026,"prob_above_bar":9.019608107716316e-18,"jeffreys_mean":0.23170731707317074,"a":10,"b":32},"oracle_posterior":{"mean":0.23809523809523808,"low":0.12363248839584409,"high":0.3761374654340026,"prob_above_bar":9.019608107716316e-18,"jeffreys_mean":0.23170731707317074,"a":10,"b":32},"cost_usd":0.017100173999999992,"known_cost_usd":0.017100173999999992,"unknown_costs":0,"median_ms":412.089,"p95_ms":499.31844999999976,"missing_latency":0},"stateful_config":{"n":40,"correct":40,"oracle_correct":40,"failures":0,"posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"oracle_posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"cost_usd":0.007669326000000002,"known_cost_usd":0.007669326000000002,"unknown_costs":0,"median_ms":392.537,"p95_ms":459.3990499999999,"missing_latency":0},"ticket_triage":{"n":40,"correct":40,"oracle_correct":40,"failures":0,"posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"oracle_posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"cost_usd":0.004439315999999999,"known_cost_usd":0.004439315999999999,"unknown_costs":0,"median_ms":318.0905,"p95_ms":385.13154999999995,"missing_latency":0}},"gate":{"threshold":0.85,"calibration_accepted":30,"accepted":109,"errors":6,"n":160,"posterior":{"mean":0.9369369369369369,"low":0.8850533748777115,"high":0.9740343856463762,"prob_above_bar":0.9996344091750515,"jeffreys_mean":0.9409090909090909,"a":104,"b":7}},"risk_curve":[{"threshold":1.0,"accepted":31,"errors":0,"coverage":0.19375,"risk":0.0},{"threshold":0.99,"accepted":57,"errors":0,"coverage":0.35625,"risk":0.0},{"threshold":0.98,"accepted":70,"errors":0,"coverage":0.4375,"risk":0.0},{"threshold":0.97,"accepted":84,"errors":3,"coverage":0.525,"risk":0.03571428571428571},{"threshold":0.96,"accepted":89,"errors":3,"coverage":0.55625,"risk":0.033707865168539325},{"threshold":0.95,"accepted":92,"errors":3,"coverage":0.575,"risk":0.03260869565217391},{"threshold":0.94,"accepted":95,"errors":3,"coverage":0.59375,"risk":0.031578947368421054},{"threshold":0.9393939393939393,"accepted":96,"errors":3,"coverage":0.6,"risk":0.03125},{"threshold":0.93,"accepted":99,"errors":5,"coverage":0.61875,"risk":0.050505050505050504},{"threshold":0.92,"accepted":102,"errors":6,"coverage":0.6375,"risk":0.058823529411764705},{"threshold":0.91,"accepted":103,"errors":6,"coverage":0.64375,"risk":0.05825242718446602},{"threshold":0.9,"accepted":105,"errors":6,"coverage":0.65625,"risk":0.05714285714285714},{"threshold":0.88,"accepted":107,"errors":6,"coverage":0.66875,"risk":0.056074766355140186},{"threshold":0.87,"accepted":108,"errors":6,"coverage":0.675,"risk":0.05555555555555555},{"threshold":0.86,"accepted":109,"errors":6,"coverage":0.68125,"risk":0.05504587155963303},{"threshold":0.84,"accepted":110,"errors":7,"coverage":0.6875,"risk":0.06363636363636363},{"threshold":0.78,"accepted":111,"errors":8,"coverage":0.69375,"risk":0.07207207207207207},{"threshold":0.72,"accepted":112,"errors":8,"coverage":0.7,"risk":0.07142857142857142},{"threshold":0.69,"accepted":113,"errors":8,"coverage":0.70625,"risk":0.07079646017699115},{"threshold":0.6599999999999999,"accepted":114,"errors":9,"coverage":0.7125,"risk":0.07894736842105263},{"threshold":0.64,"accepted":115,"errors":10,"coverage":0.71875,"risk":0.08695652173913043},{"threshold":0.63,"accepted":116,"errors":10,"coverage":0.725,"risk":0.08620689655172414},{"threshold":0.62,"accepted":117,"errors":11,"coverage":0.73125,"risk":0.09401709401709402},{"threshold":0.61,"accepted":119,"errors":12,"coverage":0.74375,"risk":0.10084033613445378},{"threshold":0.6,"accepted":120,"errors":13,"coverage":0.75,"risk":0.10833333333333334},{"threshold":0.59,"accepted":122,"errors":13,"coverage":0.7625,"risk":0.10655737704918032},{"threshold":0.5757575757575758,"accepted":123,"errors":14,"coverage":0.76875,"risk":0.11382113821138211},{"threshold":0.57,"accepted":124,"errors":14,"coverage":0.775,"risk":0.11290322580645161},{"threshold":0.56,"accepted":125,"errors":15,"coverage":0.78125,"risk":0.12},{"threshold":0.54,"accepted":126,"errors":15,"coverage":0.7875,"risk":0.11904761904761904},{"threshold":0.53,"accepted":129,"errors":17,"coverage":0.80625,"risk":0.13178294573643412},{"threshold":0.51,"accepted":130,"errors":18,"coverage":0.8125,"risk":0.13846153846153847},{"threshold":0.5,"accepted":133,"errors":21,"coverage":0.83125,"risk":0.15789473684210525},{"threshold":0.48,"accepted":134,"errors":22,"coverage":0.8375,"risk":0.16417910447761194},{"threshold":0.46,"accepted":135,"errors":23,"coverage":0.84375,"risk":0.17037037037037037},{"threshold":0.45,"accepted":136,"errors":24,"coverage":0.85,"risk":0.17647058823529413},{"threshold":0.44,"accepted":137,"errors":24,"coverage":0.85625,"risk":0.17518248175182483},{"threshold":0.42,"accepted":138,"errors":25,"coverage":0.8625,"risk":0.18115942028985507},{"threshold":0.3899999999999999,"accepted":139,"errors":26,"coverage":0.86875,"risk":0.18705035971223022},{"threshold":0.38000000000000006,"accepted":140,"errors":27,"coverage":0.875,"risk":0.19285714285714287},{"threshold":0.38,"accepted":141,"errors":28,"coverage":0.88125,"risk":0.19858156028368795},{"threshold":0.35000000000000003,"accepted":142,"errors":29,"coverage":0.8875,"risk":0.20422535211267606},{"threshold":0.34,"accepted":143,"errors":30,"coverage":0.89375,"risk":0.2097902097902098},{"threshold":0.32999999999999996,"accepted":144,"errors":31,"coverage":0.9,"risk":0.2152777777777778},{"threshold":0.3,"accepted":145,"errors":32,"coverage":0.90625,"risk":0.2206896551724138},{"threshold":0.26,"accepted":147,"errors":34,"coverage":0.91875,"risk":0.23129251700680273},{"threshold":0.24999999999999994,"accepted":148,"errors":35,"coverage":0.925,"risk":0.23648648648648649},{"threshold":0.24000000000000002,"accepted":150,"errors":36,"coverage":0.9375,"risk":0.24},{"threshold":0.21999999999999995,"accepted":151,"errors":37,"coverage":0.94375,"risk":0.24503311258278146},{"threshold":0.20999999999999996,"accepted":153,"errors":39,"coverage":0.95625,"risk":0.2549019607843137},{"threshold":0.2,"accepted":156,"errors":42,"coverage":0.975,"risk":0.2692307692307692},{"threshold":0.18999999999999995,"accepted":157,"errors":43,"coverage":0.98125,"risk":0.27388535031847133},{"threshold":0.18,"accepted":158,"errors":44,"coverage":0.9875,"risk":0.27848101265822783},{"threshold":0.17,"accepted":159,"errors":45,"coverage":0.99375,"risk":0.2830188679245283},{"threshold":0.14,"accepted":160,"errors":46,"coverage":1.0,"risk":0.2875}]},{"cohort":"transfer","variant":"archive-fable-taught","historical":true,"n":160,"correct":160,"oracle_correct":160,"failures":0,"posterior":{"mean":0.9938271604938271,"low":0.9773481985863577,"high":0.9998427588979373,"prob_above_bar":0.9999999999998215,"jeffreys_mean":0.9968944099378882,"a":161,"b":1},"oracle_posterior":{"mean":0.9938271604938271,"low":0.9773481985863577,"high":0.9998427588979373,"prob_above_bar":0.9999999999998215,"jeffreys_mean":0.9968944099378882,"a":161,"b":1},"cost_usd":4.775030000000004,"known_cost_usd":4.775030000000004,"unknown_costs":0,"median_ms":3581.4545,"p95_ms":6007.1586499999985,"missing_latency":0,"tasks":{"approval_policy":{"n":40,"correct":40,"oracle_correct":40,"failures":0,"posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"oracle_posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"cost_usd":1.0930299999999997,"known_cost_usd":1.0930299999999997,"unknown_costs":0,"median_ms":3460.938,"p95_ms":5587.231749999996,"missing_latency":0},"project_selection":{"n":40,"correct":40,"oracle_correct":40,"failures":0,"posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"oracle_posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"cost_usd":1.7808400000000004,"known_cost_usd":1.7808400000000004,"unknown_costs":0,"median_ms":4365.3395,"p95_ms":6368.094399999998,"missing_latency":0},"stateful_config":{"n":40,"correct":40,"oracle_correct":40,"failures":0,"posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"oracle_posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"cost_usd":1.0807099999999998,"known_cost_usd":1.0807099999999998,"unknown_costs":0,"median_ms":3588.9435,"p95_ms":4845.5485,"missing_latency":0},"ticket_triage":{"n":40,"correct":40,"oracle_correct":40,"failures":0,"posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"oracle_posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"cost_usd":0.8204500000000002,"known_cost_usd":0.8204500000000002,"unknown_costs":0,"median_ms":3315.6895,"p95_ms":4215.505549999999,"missing_latency":0}},"gate":null,"risk_curve":[]},{"cohort":"transfer","variant":"archive-fable-zero","historical":true,"n":160,"correct":160,"oracle_correct":160,"failures":0,"posterior":{"mean":0.9938271604938271,"low":0.9773481985863577,"high":0.9998427588979373,"prob_above_bar":0.9999999999998215,"jeffreys_mean":0.9968944099378882,"a":161,"b":1},"oracle_posterior":{"mean":0.9938271604938271,"low":0.9773481985863577,"high":0.9998427588979373,"prob_above_bar":0.9999999999998215,"jeffreys_mean":0.9968944099378882,"a":161,"b":1},"cost_usd":2.9184800000000006,"known_cost_usd":2.9184800000000006,"unknown_costs":0,"median_ms":3656.989,"p95_ms":6139.851899999999,"missing_latency":0,"tasks":{"approval_policy":{"n":40,"correct":40,"oracle_correct":40,"failures":0,"posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"oracle_posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"cost_usd":0.71528,"known_cost_usd":0.71528,"unknown_costs":0,"median_ms":3631.058,"p95_ms":5287.432499999999,"missing_latency":0},"project_selection":{"n":40,"correct":40,"oracle_correct":40,"failures":0,"posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"oracle_posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"cost_usd":1.1046899999999997,"known_cost_usd":1.1046899999999997,"unknown_costs":0,"median_ms":5012.871,"p95_ms":6774.38545,"missing_latency":0},"stateful_config":{"n":40,"correct":40,"oracle_correct":40,"failures":0,"posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"oracle_posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"cost_usd":0.62271,"known_cost_usd":0.62271,"unknown_costs":0,"median_ms":3301.1465,"p95_ms":4204.341649999999,"missing_latency":0},"ticket_triage":{"n":40,"correct":40,"oracle_correct":40,"failures":0,"posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"oracle_posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"cost_usd":0.47580000000000006,"known_cost_usd":0.47580000000000006,"unknown_costs":0,"median_ms":3545.9995,"p95_ms":4341.009699999999,"missing_latency":0}},"gate":null,"risk_curve":[]},{"cohort":"transfer","variant":"responses-zero","historical":false,"n":160,"correct":159,"oracle_correct":159,"failures":0,"posterior":{"mean":0.9876543209876543,"low":0.965880874845245,"high":0.9984920381266038,"prob_above_bar":0.9999999999940713,"jeffreys_mean":0.9906832298136646,"a":160,"b":2},"oracle_posterior":{"mean":0.9876543209876543,"low":0.965880874845245,"high":0.9984920381266038,"prob_above_bar":0.9999999999940713,"jeffreys_mean":0.9906832298136646,"a":160,"b":2},"cost_usd":0.0196351,"known_cost_usd":0.0196351,"unknown_costs":0,"median_ms":1503.8489999715239,"p95_ms":3334.9866420496255,"missing_latency":0,"tasks":{"approval_policy":{"n":40,"correct":40,"oracle_correct":40,"failures":0,"posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"oracle_posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"cost_usd":0.0047518000000000005,"known_cost_usd":0.0047518000000000005,"unknown_costs":0,"median_ms":1459.5508335041814,"p95_ms":2031.2413830019063,"missing_latency":0},"project_selection":{"n":40,"correct":39,"oracle_correct":39,"failures":0,"posterior":{"mean":0.9523809523809523,"low":0.8714459796158537,"high":0.9940368817589572,"prob_above_bar":0.9947837699178925,"jeffreys_mean":0.9634146341463414,"a":40,"b":2},"oracle_posterior":{"mean":0.9523809523809523,"low":0.8714459796158537,"high":0.9940368817589572,"prob_above_bar":0.9947837699178925,"jeffreys_mean":0.9634146341463414,"a":40,"b":2},"cost_usd":0.007618899999999999,"known_cost_usd":0.007618899999999999,"unknown_costs":0,"median_ms":2374.3420409737155,"p95_ms":4132.891841721721,"missing_latency":0},"stateful_config":{"n":40,"correct":40,"oracle_correct":40,"failures":0,"posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"oracle_posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"cost_usd":0.003986899999999999,"known_cost_usd":0.003986899999999999,"unknown_costs":0,"median_ms":1419.0963745350018,"p95_ms":2116.489477822324,"missing_latency":0},"ticket_triage":{"n":40,"correct":40,"oracle_correct":40,"failures":0,"posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"oracle_posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"cost_usd":0.0032775,"known_cost_usd":0.0032775,"unknown_costs":0,"median_ms":1355.7497080182657,"p95_ms":1787.2658378037158,"missing_latency":0}},"gate":null,"risk_curve":[]},{"cohort":"transfer","variant":"decisions-zero","historical":false,"n":160,"correct":105,"oracle_correct":105,"failures":9,"posterior":{"mean":0.654320987654321,"low":0.5796181523793575,"high":0.7254142946892794,"prob_above_bar":1.9944492788286512e-08,"jeffreys_mean":0.65527950310559,"a":106,"b":56},"oracle_posterior":{"mean":0.654320987654321,"low":0.5796181523793575,"high":0.7254142946892794,"prob_above_bar":1.9944492788286512e-08,"jeffreys_mean":0.65527950310559,"a":106,"b":56},"cost_usd":0.06308109999999997,"known_cost_usd":0.06308109999999997,"unknown_costs":0,"median_ms":153.28293747734278,"p95_ms":231.07070795958856,"missing_latency":0,"tasks":{"approval_policy":{"n":40,"correct":27,"oracle_correct":27,"failures":0,"posterior":{"mean":0.6666666666666666,"low":0.5191336131610755,"high":0.799165886282918,"prob_above_bar":0.00498707173522759,"jeffreys_mean":0.6707317073170732,"a":28,"b":14},"oracle_posterior":{"mean":0.6666666666666666,"low":0.5191336131610755,"high":0.799165886282918,"prob_above_bar":0.00498707173522759,"jeffreys_mean":0.6707317073170732,"a":28,"b":14},"cost_usd":0.013631800000000003,"known_cost_usd":0.013631800000000003,"unknown_costs":0,"median_ms":159.72579206572846,"p95_ms":235.9721587330568,"missing_latency":0},"project_selection":{"n":40,"correct":2,"oracle_correct":2,"failures":9,"posterior":{"mean":0.07142857142857142,"low":0.015351471533416648,"high":0.1653333534178303,"prob_above_bar":2.58163561613729e-28,"jeffreys_mean":0.06097560975609756,"a":3,"b":39},"oracle_posterior":{"mean":0.07142857142857142,"low":0.015351471533416648,"high":0.1653333534178303,"prob_above_bar":2.58163561613729e-28,"jeffreys_mean":0.06097560975609756,"a":3,"b":39},"cost_usd":0.029885400000000013,"known_cost_usd":0.029885400000000013,"unknown_costs":0,"median_ms":164.75381253985688,"p95_ms":244.77394583518617,"missing_latency":0},"stateful_config":{"n":40,"correct":36,"oracle_correct":36,"failures":0,"posterior":{"mean":0.8809523809523809,"low":0.7686854506905387,"high":0.9591932737525728,"prob_above_bar":0.8359654394005773,"jeffreys_mean":0.8902439024390244,"a":37,"b":5},"oracle_posterior":{"mean":0.8809523809523809,"low":0.7686854506905387,"high":0.9591932737525728,"prob_above_bar":0.8359654394005773,"jeffreys_mean":0.8902439024390244,"a":37,"b":5},"cost_usd":0.012693900000000003,"known_cost_usd":0.012693900000000003,"unknown_costs":0,"median_ms":151.0874375817366,"p95_ms":202.8006860928144,"missing_latency":0},"ticket_triage":{"n":40,"correct":40,"oracle_correct":40,"failures":0,"posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"oracle_posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"cost_usd":0.006870000000000001,"known_cost_usd":0.006870000000000001,"unknown_costs":0,"median_ms":141.13618741976097,"p95_ms":219.30630410788575,"missing_latency":0}},"gate":{"threshold":0.65,"calibration_accepted":23,"minimum":4,"planned_calibration":48,"accepted":86,"errors":1,"n":160,"posterior":{"mean":0.9772727272727273,"low":0.9376156797458327,"high":0.9972037589615068,"prob_above_bar":0.9999976228892455,"jeffreys_mean":0.9827586206896551,"a":86,"b":2}},"risk_curve":[{"threshold":1.0,"accepted":12,"errors":0,"coverage":0.075,"risk":0.0},{"threshold":0.99,"accepted":28,"errors":0,"coverage":0.175,"risk":0.0},{"threshold":0.98,"accepted":33,"errors":0,"coverage":0.20625,"risk":0.0},{"threshold":0.97,"accepted":39,"errors":0,"coverage":0.24375,"risk":0.0},{"threshold":0.96,"accepted":40,"errors":0,"coverage":0.25,"risk":0.0},{"threshold":0.95,"accepted":44,"errors":0,"coverage":0.275,"risk":0.0},{"threshold":0.94,"accepted":48,"errors":0,"coverage":0.3,"risk":0.0},{"threshold":0.92,"accepted":51,"errors":0,"coverage":0.31875,"risk":0.0},{"threshold":0.91,"accepted":56,"errors":0,"coverage":0.35,"risk":0.0},{"threshold":0.9,"accepted":57,"errors":0,"coverage":0.35625,"risk":0.0},{"threshold":0.89,"accepted":60,"errors":0,"coverage":0.375,"risk":0.0},{"threshold":0.88,"accepted":61,"errors":0,"coverage":0.38125,"risk":0.0},{"threshold":0.87,"accepted":63,"errors":0,"coverage":0.39375,"risk":0.0},{"threshold":0.8500000000000001,"accepted":64,"errors":0,"coverage":0.4,"risk":0.0},{"threshold":0.85,"accepted":66,"errors":0,"coverage":0.4125,"risk":0.0},{"threshold":0.84,"accepted":68,"errors":0,"coverage":0.425,"risk":0.0},{"threshold":0.82,"accepted":69,"errors":0,"coverage":0.43125,"risk":0.0},{"threshold":0.81,"accepted":70,"errors":0,"coverage":0.4375,"risk":0.0},{"threshold":0.8,"accepted":72,"errors":0,"coverage":0.45,"risk":0.0},{"threshold":0.78,"accepted":74,"errors":0,"coverage":0.4625,"risk":0.0},{"threshold":0.77,"accepted":76,"errors":0,"coverage":0.475,"risk":0.0},{"threshold":0.76,"accepted":77,"errors":0,"coverage":0.48125,"risk":0.0},{"threshold":0.74,"accepted":78,"errors":1,"coverage":0.4875,"risk":0.01282051282051282},{"threshold":0.73,"accepted":79,"errors":1,"coverage":0.49375,"risk":0.012658227848101266},{"threshold":0.72,"accepted":80,"errors":1,"coverage":0.5,"risk":0.0125},{"threshold":0.71,"accepted":81,"errors":1,"coverage":0.50625,"risk":0.012345679012345678},{"threshold":0.69,"accepted":82,"errors":1,"coverage":0.5125,"risk":0.012195121951219513},{"threshold":0.67,"accepted":85,"errors":1,"coverage":0.53125,"risk":0.011764705882352941},{"threshold":0.66,"accepted":86,"errors":1,"coverage":0.5375,"risk":0.011627906976744186},{"threshold":0.62,"accepted":87,"errors":1,"coverage":0.54375,"risk":0.011494252873563218},{"threshold":0.6,"accepted":89,"errors":2,"coverage":0.55625,"risk":0.02247191011235955},{"threshold":0.59,"accepted":91,"errors":2,"coverage":0.56875,"risk":0.02197802197802198},{"threshold":0.58,"accepted":93,"errors":3,"coverage":0.58125,"risk":0.03225806451612903},{"threshold":0.55,"accepted":94,"errors":3,"coverage":0.5875,"risk":0.031914893617021274},{"threshold":0.54,"accepted":96,"errors":3,"coverage":0.6,"risk":0.03125},{"threshold":0.53,"accepted":98,"errors":5,"coverage":0.6125,"risk":0.05102040816326531},{"threshold":0.52,"accepted":100,"errors":6,"coverage":0.625,"risk":0.06},{"threshold":0.51,"accepted":103,"errors":7,"coverage":0.64375,"risk":0.06796116504854369},{"threshold":0.5,"accepted":104,"errors":7,"coverage":0.65,"risk":0.0673076923076923},{"threshold":0.49,"accepted":105,"errors":7,"coverage":0.65625,"risk":0.06666666666666667},{"threshold":0.47,"accepted":107,"errors":7,"coverage":0.66875,"risk":0.06542056074766354},{"threshold":0.46,"accepted":108,"errors":8,"coverage":0.675,"risk":0.07407407407407407},{"threshold":0.45,"accepted":109,"errors":9,"coverage":0.68125,"risk":0.08256880733944955},{"threshold":0.44,"accepted":112,"errors":11,"coverage":0.7,"risk":0.09821428571428571},{"threshold":0.43,"accepted":113,"errors":11,"coverage":0.70625,"risk":0.09734513274336283},{"threshold":0.41,"accepted":114,"errors":12,"coverage":0.7125,"risk":0.10526315789473684},{"threshold":0.4,"accepted":116,"errors":13,"coverage":0.725,"risk":0.11206896551724138},{"threshold":0.39,"accepted":117,"errors":14,"coverage":0.73125,"risk":0.11965811965811966},{"threshold":0.37,"accepted":118,"errors":15,"coverage":0.7375,"risk":0.1271186440677966},{"threshold":0.36999999999999994,"accepted":119,"errors":15,"coverage":0.74375,"risk":0.12605042016806722},{"threshold":0.36000000000000004,"accepted":120,"errors":16,"coverage":0.75,"risk":0.13333333333333333},{"threshold":0.33999999999999997,"accepted":121,"errors":17,"coverage":0.75625,"risk":0.14049586776859505},{"threshold":0.31,"accepted":122,"errors":18,"coverage":0.7625,"risk":0.14754098360655737},{"threshold":0.30999999999999994,"accepted":123,"errors":19,"coverage":0.76875,"risk":0.15447154471544716},{"threshold":0.3,"accepted":124,"errors":20,"coverage":0.775,"risk":0.16129032258064516},{"threshold":0.29,"accepted":125,"errors":21,"coverage":0.78125,"risk":0.168},{"threshold":0.27999999999999997,"accepted":127,"errors":22,"coverage":0.79375,"risk":0.1732283464566929},{"threshold":0.26999999999999996,"accepted":128,"errors":23,"coverage":0.8,"risk":0.1796875},{"threshold":0.25,"accepted":129,"errors":24,"coverage":0.80625,"risk":0.18604651162790697},{"threshold":0.23999999999999994,"accepted":130,"errors":25,"coverage":0.8125,"risk":0.19230769230769232},{"threshold":0.22999999999999995,"accepted":131,"errors":26,"coverage":0.81875,"risk":0.1984732824427481},{"threshold":0.21999999999999995,"accepted":132,"errors":27,"coverage":0.825,"risk":0.20454545454545456},{"threshold":0.17,"accepted":133,"errors":28,"coverage":0.83125,"risk":0.21052631578947367},{"threshold":0.16999999999999998,"accepted":135,"errors":30,"coverage":0.84375,"risk":0.2222222222222222},{"threshold":0.16,"accepted":136,"errors":31,"coverage":0.85,"risk":0.22794117647058823},{"threshold":0.15999999999999998,"accepted":137,"errors":32,"coverage":0.85625,"risk":0.23357664233576642},{"threshold":0.14999999999999997,"accepted":139,"errors":34,"coverage":0.86875,"risk":0.2446043165467626},{"threshold":0.13999999999999999,"accepted":140,"errors":35,"coverage":0.875,"risk":0.25},{"threshold":0.12999999999999998,"accepted":142,"errors":37,"coverage":0.8875,"risk":0.2605633802816901},{"threshold":0.12999999999999995,"accepted":143,"errors":38,"coverage":0.89375,"risk":0.26573426573426573},{"threshold":0.12,"accepted":144,"errors":39,"coverage":0.9,"risk":0.2708333333333333},{"threshold":0.11999999999999997,"accepted":146,"errors":41,"coverage":0.9125,"risk":0.2808219178082192},{"threshold":0.11999999999999994,"accepted":147,"errors":42,"coverage":0.91875,"risk":0.2857142857142857},{"threshold":0.10999999999999997,"accepted":148,"errors":43,"coverage":0.925,"risk":0.2905405405405405},{"threshold":0.09999999999999996,"accepted":149,"errors":44,"coverage":0.93125,"risk":0.2953020134228188},{"threshold":0.08999999999999998,"accepted":150,"errors":45,"coverage":0.9375,"risk":0.3},{"threshold":0.08999999999999996,"accepted":151,"errors":46,"coverage":0.94375,"risk":0.304635761589404}]},{"cohort":"transfer","variant":"decisions-taught","historical":false,"n":160,"correct":75,"oracle_correct":75,"failures":3,"posterior":{"mean":0.4691358024691358,"low":0.3929865924494434,"high":0.5460070510662318,"prob_above_bar":2.1062017746952228e-26,"jeffreys_mean":0.468944099378882,"a":76,"b":86},"oracle_posterior":{"mean":0.4691358024691358,"low":0.3929865924494434,"high":0.5460070510662318,"prob_above_bar":2.1062017746952228e-26,"jeffreys_mean":0.468944099378882,"a":76,"b":86},"cost_usd":0.07530910000000003,"known_cost_usd":0.07530910000000003,"unknown_costs":0,"median_ms":154.930979013443,"p95_ms":220.70157326525074,"missing_latency":0,"tasks":{"approval_policy":{"n":40,"correct":9,"oracle_correct":9,"failures":0,"posterior":{"mean":0.23809523809523808,"low":0.12363248839584409,"high":0.3761374654340026,"prob_above_bar":9.019608107716316e-18,"jeffreys_mean":0.23170731707317074,"a":10,"b":32},"oracle_posterior":{"mean":0.23809523809523808,"low":0.12363248839584409,"high":0.3761374654340026,"prob_above_bar":9.019608107716316e-18,"jeffreys_mean":0.23170731707317074,"a":10,"b":32},"cost_usd":0.015867799999999998,"known_cost_usd":0.015867799999999998,"unknown_costs":0,"median_ms":150.77272953931242,"p95_ms":243.1787291425279,"missing_latency":0},"project_selection":{"n":40,"correct":0,"oracle_correct":0,"failures":3,"posterior":{"mean":0.023809523809523808,"low":0.0006173168932865188,"high":0.08604383629402843,"prob_above_bar":1.2468055713982848e-32,"jeffreys_mean":0.012195121951219513,"a":1,"b":41},"oracle_posterior":{"mean":0.023809523809523808,"low":0.0006173168932865188,"high":0.08604383629402843,"prob_above_bar":1.2468055713982848e-32,"jeffreys_mean":0.012195121951219513,"a":1,"b":41},"cost_usd":0.03471740000000001,"known_cost_usd":0.03471740000000001,"unknown_costs":0,"median_ms":170.07181298686191,"p95_ms":217.62170207221058,"missing_latency":0},"stateful_config":{"n":40,"correct":27,"oracle_correct":27,"failures":0,"posterior":{"mean":0.6666666666666666,"low":0.5191336131610755,"high":0.799165886282918,"prob_above_bar":0.00498707173522759,"jeffreys_mean":0.6707317073170732,"a":28,"b":14},"oracle_posterior":{"mean":0.6666666666666666,"low":0.5191336131610755,"high":0.799165886282918,"prob_above_bar":0.00498707173522759,"jeffreys_mean":0.6707317073170732,"a":28,"b":14},"cost_usd":0.015589899999999999,"known_cost_usd":0.015589899999999999,"unknown_costs":0,"median_ms":156.21372952591628,"p95_ms":221.8680316174868,"missing_latency":0},"ticket_triage":{"n":40,"correct":39,"oracle_correct":39,"failures":0,"posterior":{"mean":0.9523809523809523,"low":0.8714459796158537,"high":0.9940368817589572,"prob_above_bar":0.9947837699178925,"jeffreys_mean":0.9634146341463414,"a":40,"b":2},"oracle_posterior":{"mean":0.9523809523809523,"low":0.8714459796158537,"high":0.9940368817589572,"prob_above_bar":0.9947837699178925,"jeffreys_mean":0.9634146341463414,"a":40,"b":2},"cost_usd":0.009134,"known_cost_usd":0.009134,"unknown_costs":0,"median_ms":146.73639595275745,"p95_ms":203.3023125259205,"missing_latency":0}},"gate":{"threshold":0.67,"calibration_accepted":21,"minimum":4,"planned_calibration":48,"accepted":56,"errors":3,"n":160,"posterior":{"mean":0.9310344827586207,"low":0.8538009356894765,"high":0.9805495793052907,"prob_above_bar":0.9904834673515662,"jeffreys_mean":0.9385964912280702,"a":54,"b":4}},"risk_curve":[{"threshold":1.0,"accepted":12,"errors":0,"coverage":0.075,"risk":0.0},{"threshold":0.99,"accepted":21,"errors":0,"coverage":0.13125,"risk":0.0},{"threshold":0.97,"accepted":23,"errors":0,"coverage":0.14375,"risk":0.0},{"threshold":0.96,"accepted":25,"errors":0,"coverage":0.15625,"risk":0.0},{"threshold":0.95,"accepted":26,"errors":0,"coverage":0.1625,"risk":0.0},{"threshold":0.94,"accepted":30,"errors":0,"coverage":0.1875,"risk":0.0},{"threshold":0.93,"accepted":31,"errors":0,"coverage":0.19375,"risk":0.0},{"threshold":0.92,"accepted":34,"errors":0,"coverage":0.2125,"risk":0.0},{"threshold":0.9,"accepted":35,"errors":0,"coverage":0.21875,"risk":0.0},{"threshold":0.89,"accepted":37,"errors":0,"coverage":0.23125,"risk":0.0},{"threshold":0.88,"accepted":38,"errors":0,"coverage":0.2375,"risk":0.0},{"threshold":0.87,"accepted":39,"errors":0,"coverage":0.24375,"risk":0.0},{"threshold":0.85,"accepted":41,"errors":0,"coverage":0.25625,"risk":0.0},{"threshold":0.84,"accepted":42,"errors":0,"coverage":0.2625,"risk":0.0},{"threshold":0.83,"accepted":43,"errors":0,"coverage":0.26875,"risk":0.0},{"threshold":0.81,"accepted":44,"errors":0,"coverage":0.275,"risk":0.0},{"threshold":0.8,"accepted":45,"errors":0,"coverage":0.28125,"risk":0.0},{"threshold":0.79,"accepted":46,"errors":1,"coverage":0.2875,"risk":0.021739130434782608},{"threshold":0.76,"accepted":48,"errors":1,"coverage":0.3,"risk":0.020833333333333332},{"threshold":0.75,"accepted":52,"errors":1,"coverage":0.325,"risk":0.019230769230769232},{"threshold":0.73,"accepted":53,"errors":1,"coverage":0.33125,"risk":0.018867924528301886},{"threshold":0.71,"accepted":54,"errors":2,"coverage":0.3375,"risk":0.037037037037037035},{"threshold":0.69,"accepted":55,"errors":2,"coverage":0.34375,"risk":0.03636363636363636},{"threshold":0.68,"accepted":56,"errors":3,"coverage":0.35,"risk":0.05357142857142857},{"threshold":0.65,"accepted":57,"errors":3,"coverage":0.35625,"risk":0.05263157894736842},{"threshold":0.64,"accepted":60,"errors":4,"coverage":0.375,"risk":0.06666666666666667},{"threshold":0.63,"accepted":61,"errors":4,"coverage":0.38125,"risk":0.06557377049180328},{"threshold":0.6200000000000001,"accepted":62,"errors":4,"coverage":0.3875,"risk":0.06451612903225806},{"threshold":0.61,"accepted":64,"errors":5,"coverage":0.4,"risk":0.078125},{"threshold":0.6,"accepted":65,"errors":6,"coverage":0.40625,"risk":0.09230769230769231},{"threshold":0.59,"accepted":66,"errors":6,"coverage":0.4125,"risk":0.09090909090909091},{"threshold":0.5800000000000001,"accepted":67,"errors":6,"coverage":0.41875,"risk":0.08955223880597014},{"threshold":0.58,"accepted":69,"errors":7,"coverage":0.43125,"risk":0.10144927536231885},{"threshold":0.5700000000000001,"accepted":70,"errors":8,"coverage":0.4375,"risk":0.11428571428571428},{"threshold":0.57,"accepted":72,"errors":9,"coverage":0.45,"risk":0.125},{"threshold":0.56,"accepted":76,"errors":12,"coverage":0.475,"risk":0.15789473684210525},{"threshold":0.55,"accepted":79,"errors":13,"coverage":0.49375,"risk":0.16455696202531644},{"threshold":0.54,"accepted":81,"errors":14,"coverage":0.50625,"risk":0.1728395061728395},{"threshold":0.53,"accepted":84,"errors":15,"coverage":0.525,"risk":0.17857142857142858},{"threshold":0.51,"accepted":87,"errors":18,"coverage":0.54375,"risk":0.20689655172413793},{"threshold":0.5,"accepted":88,"errors":19,"coverage":0.55,"risk":0.2159090909090909},{"threshold":0.4999999999999998,"accepted":89,"errors":20,"coverage":0.55625,"risk":0.2247191011235955},{"threshold":0.49,"accepted":91,"errors":22,"coverage":0.56875,"risk":0.24175824175824176},{"threshold":0.48,"accepted":93,"errors":23,"coverage":0.58125,"risk":0.24731182795698925},{"threshold":0.47,"accepted":94,"errors":24,"coverage":0.5875,"risk":0.2553191489361702},{"threshold":0.46,"accepted":97,"errors":26,"coverage":0.60625,"risk":0.26804123711340205},{"threshold":0.45,"accepted":98,"errors":27,"coverage":0.6125,"risk":0.2755102040816326},{"threshold":0.4499999999999999,"accepted":99,"errors":28,"coverage":0.61875,"risk":0.2828282828282828},{"threshold":0.43,"accepted":100,"errors":28,"coverage":0.625,"risk":0.28},{"threshold":0.42,"accepted":103,"errors":31,"coverage":0.64375,"risk":0.30097087378640774},{"threshold":0.3999999999999999,"accepted":104,"errors":32,"coverage":0.65,"risk":0.3076923076923077},{"threshold":0.39,"accepted":105,"errors":33,"coverage":0.65625,"risk":0.3142857142857143},{"threshold":0.38,"accepted":107,"errors":35,"coverage":0.66875,"risk":0.32710280373831774},{"threshold":0.37,"accepted":109,"errors":36,"coverage":0.68125,"risk":0.3302752293577982},{"threshold":0.36,"accepted":111,"errors":38,"coverage":0.69375,"risk":0.34234234234234234},{"threshold":0.35,"accepted":112,"errors":38,"coverage":0.7,"risk":0.3392857142857143},{"threshold":0.34,"accepted":116,"errors":41,"coverage":0.725,"risk":0.35344827586206895},{"threshold":0.33,"accepted":117,"errors":42,"coverage":0.73125,"risk":0.358974358974359},{"threshold":0.32,"accepted":118,"errors":43,"coverage":0.7375,"risk":0.3644067796610169},{"threshold":0.31999999999999995,"accepted":120,"errors":45,"coverage":0.75,"risk":0.375},{"threshold":0.31,"accepted":121,"errors":46,"coverage":0.75625,"risk":0.38016528925619836},{"threshold":0.30999999999999994,"accepted":122,"errors":47,"coverage":0.7625,"risk":0.38524590163934425},{"threshold":0.29,"accepted":124,"errors":49,"coverage":0.775,"risk":0.3951612903225806},{"threshold":0.26999999999999996,"accepted":126,"errors":51,"coverage":0.7875,"risk":0.40476190476190477},{"threshold":0.25999999999999995,"accepted":127,"errors":52,"coverage":0.79375,"risk":0.4094488188976378},{"threshold":0.25000000000000006,"accepted":128,"errors":53,"coverage":0.8,"risk":0.4140625},{"threshold":0.25,"accepted":129,"errors":54,"coverage":0.80625,"risk":0.4186046511627907},{"threshold":0.24,"accepted":130,"errors":55,"coverage":0.8125,"risk":0.4230769230769231},{"threshold":0.23999999999999994,"accepted":131,"errors":56,"coverage":0.81875,"risk":0.42748091603053434},{"threshold":0.23,"accepted":132,"errors":57,"coverage":0.825,"risk":0.4318181818181818},{"threshold":0.22,"accepted":133,"errors":58,"coverage":0.83125,"risk":0.43609022556390975},{"threshold":0.21999999999999995,"accepted":135,"errors":60,"coverage":0.84375,"risk":0.4444444444444444},{"threshold":0.20999999999999994,"accepted":136,"errors":61,"coverage":0.85,"risk":0.4485294117647059},{"threshold":0.2,"accepted":138,"errors":63,"coverage":0.8625,"risk":0.45652173913043476},{"threshold":0.19999999999999996,"accepted":139,"errors":64,"coverage":0.86875,"risk":0.460431654676259},{"threshold":0.18999999999999995,"accepted":140,"errors":65,"coverage":0.875,"risk":0.4642857142857143},{"threshold":0.17999999999999997,"accepted":141,"errors":66,"coverage":0.88125,"risk":0.46808510638297873},{"threshold":0.16999999999999998,"accepted":142,"errors":67,"coverage":0.8875,"risk":0.47183098591549294},{"threshold":0.16999999999999993,"accepted":143,"errors":68,"coverage":0.89375,"risk":0.4755244755244755},{"threshold":0.16,"accepted":144,"errors":69,"coverage":0.9,"risk":0.4791666666666667},{"threshold":0.15999999999999998,"accepted":145,"errors":70,"coverage":0.90625,"risk":0.4827586206896552},{"threshold":0.15,"accepted":148,"errors":73,"coverage":0.925,"risk":0.49324324324324326},{"threshold":0.13999999999999999,"accepted":150,"errors":75,"coverage":0.9375,"risk":0.5},{"threshold":0.11999999999999997,"accepted":153,"errors":78,"coverage":0.95625,"risk":0.5098039215686274},{"threshold":0.10999999999999997,"accepted":154,"errors":79,"coverage":0.9625,"risk":0.512987012987013},{"threshold":0.09999999999999998,"accepted":155,"errors":80,"coverage":0.96875,"risk":0.5161290322580645},{"threshold":0.08999999999999998,"accepted":156,"errors":81,"coverage":0.975,"risk":0.5192307692307693},{"threshold":0.05999999999999997,"accepted":157,"errors":82,"coverage":0.98125,"risk":0.5222929936305732}]},{"cohort":"transfer","variant":"responses-taught","historical":false,"n":160,"correct":160,"oracle_correct":160,"failures":0,"posterior":{"mean":0.9938271604938271,"low":0.9773481985863577,"high":0.9998427588979373,"prob_above_bar":0.9999999999998215,"jeffreys_mean":0.9968944099378882,"a":161,"b":1},"oracle_posterior":{"mean":0.9938271604938271,"low":0.9773481985863577,"high":0.9998427588979373,"prob_above_bar":0.9999999999998215,"jeffreys_mean":0.9968944099378882,"a":161,"b":1},"cost_usd":0.038330874999999986,"known_cost_usd":0.038330874999999986,"unknown_costs":0,"median_ms":1459.1994999791496,"p95_ms":3647.018483444115,"missing_latency":0,"tasks":{"approval_policy":{"n":40,"correct":40,"oracle_correct":40,"failures":0,"posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"oracle_posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"cost_usd":0.008391250000000001,"known_cost_usd":0.008391250000000001,"unknown_costs":0,"median_ms":1416.7104375083,"p95_ms":2010.186208289815,"missing_latency":0},"project_selection":{"n":40,"correct":40,"oracle_correct":40,"failures":0,"posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"oracle_posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"cost_usd":0.015158249999999998,"known_cost_usd":0.015158249999999998,"unknown_costs":0,"median_ms":2515.287291491404,"p95_ms":4262.771606288151,"missing_latency":0},"stateful_config":{"n":40,"correct":40,"oracle_correct":40,"failures":0,"posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"oracle_posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"cost_usd":0.008314875,"known_cost_usd":0.008314875,"unknown_costs":0,"median_ms":1426.616666954942,"p95_ms":1970.376760634826,"missing_latency":0},"ticket_triage":{"n":40,"correct":40,"oracle_correct":40,"failures":0,"posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"oracle_posterior":{"mean":0.9761904761904762,"low":0.9139561637059715,"high":0.9993826831067135,"prob_above_bar":0.9994330184693362,"jeffreys_mean":0.9878048780487805,"a":41,"b":1},"cost_usd":0.006466500000000001,"known_cost_usd":0.006466500000000001,"unknown_costs":0,"median_ms":1248.0595005326904,"p95_ms":1868.1841753539618,"missing_latency":0}},"gate":null,"risk_curve":[]},{"cohort":"applicability","variant":"archive-luna-fallback","historical":true,"n":240,"correct":236,"oracle_correct":237,"failures":0,"posterior":{"mean":0.9793388429752066,"low":0.9580513834044089,"high":0.9932300389537785,"prob_above_bar":0.9999999999999802,"jeffreys_mean":0.9813278008298755,"a":237,"b":5},"oracle_posterior":{"mean":0.9834710743801653,"low":0.9640542920064572,"high":0.9954596919995788,"prob_above_bar":0.9999999999999983,"jeffreys_mean":0.9854771784232366,"a":238,"b":4},"cost_usd":0.06632319999999998,"known_cost_usd":0.06632319999999998,"unknown_costs":0,"median_ms":1877.83,"p95_ms":3002.114849999999,"missing_latency":0,"tasks":{"approval_policy":{"n":60,"correct":59,"oracle_correct":60,"failures":0,"posterior":{"mean":0.967741935483871,"low":0.9120118770772547,"high":0.9960043968326835,"prob_above_bar":0.9998047828706151,"jeffreys_mean":0.9754098360655737,"a":60,"b":2},"oracle_posterior":{"mean":0.9838709677419355,"low":0.9413187760677074,"high":0.9995850400871603,"prob_above_bar":0.9999852108235314,"jeffreys_mean":0.9918032786885246,"a":61,"b":1},"cost_usd":0.0160718,"known_cost_usd":0.0160718,"unknown_costs":0,"median_ms":2034.1799999999998,"p95_ms":2678.254499999999,"missing_latency":0},"project_selection":{"n":60,"correct":58,"oracle_correct":58,"failures":0,"posterior":{"mean":0.9516129032258065,"low":0.8865278328118185,"high":0.9897406848316792,"prob_above_bar":0.998722215153117,"jeffreys_mean":0.9590163934426229,"a":59,"b":3},"oracle_posterior":{"mean":0.9516129032258065,"low":0.8865278328118185,"high":0.9897406848316792,"prob_above_bar":0.998722215153117,"jeffreys_mean":0.9590163934426229,"a":59,"b":3},"cost_usd":0.024019200000000008,"known_cost_usd":0.024019200000000008,"unknown_costs":0,"median_ms":2211.0515,"p95_ms":3848.7558999999987,"missing_latency":0},"stateful_config":{"n":60,"correct":60,"oracle_correct":60,"failures":0,"posterior":{"mean":0.9838709677419355,"low":0.9413187760677074,"high":0.9995850400871603,"prob_above_bar":0.9999852108235314,"jeffreys_mean":0.9918032786885246,"a":61,"b":1},"oracle_posterior":{"mean":0.9838709677419355,"low":0.9413187760677074,"high":0.9995850400871603,"prob_above_bar":0.9999852108235314,"jeffreys_mean":0.9918032786885246,"a":61,"b":1},"cost_usd":0.015066,"known_cost_usd":0.015066,"unknown_costs":0,"median_ms":2008.417,"p95_ms":2486.9952,"missing_latency":0},"ticket_triage":{"n":60,"correct":59,"oracle_correct":59,"failures":0,"posterior":{"mean":0.967741935483871,"low":0.9120118770772547,"high":0.9960043968326835,"prob_above_bar":0.9998047828706151,"jeffreys_mean":0.9754098360655737,"a":60,"b":2},"oracle_posterior":{"mean":0.967741935483871,"low":0.9120118770772547,"high":0.9960043968326835,"prob_above_bar":0.9998047828706151,"jeffreys_mean":0.9754098360655737,"a":60,"b":2},"cost_usd":0.011166200000000003,"known_cost_usd":0.011166200000000003,"unknown_costs":0,"median_ms":1368.6925,"p95_ms":1919.9273,"missing_latency":0}},"gate":null,"risk_curve":[]},{"cohort":"applicability","variant":"archive-jev-zero","historical":true,"n":240,"correct":180,"oracle_correct":181,"failures":0,"posterior":{"mean":0.7479338842975206,"low":0.6914743294014548,"high":0.8005125728538232,"prob_above_bar":0.0004220985394667399,"jeffreys_mean":0.7489626556016598,"a":181,"b":61},"oracle_posterior":{"mean":0.7520661157024794,"low":0.6958749649973335,"high":0.8043117891311331,"prob_above_bar":0.0007309007080465509,"jeffreys_mean":0.7531120331950207,"a":182,"b":60},"cost_usd":0.043413972000000016,"known_cost_usd":0.043413972000000016,"unknown_costs":0,"median_ms":349.9155,"p95_ms":445.89439999999996,"missing_latency":0,"tasks":{"approval_policy":{"n":60,"correct":52,"oracle_correct":53,"failures":0,"posterior":{"mean":0.8548387096774194,"low":0.7578413113166378,"high":0.9302477020274412,"prob_above_bar":0.70732716817573,"jeffreys_mean":0.860655737704918,"a":53,"b":9},"oracle_posterior":{"mean":0.8709677419354839,"low":0.7777507914635026,"high":0.9416394941602595,"prob_above_bar":0.8188192065400973,"jeffreys_mean":0.8770491803278688,"a":54,"b":8},"cost_usd":0.009632154000000004,"known_cost_usd":0.009632154000000004,"unknown_costs":0,"median_ms":331.8185,"p95_ms":429.39965,"missing_latency":0},"project_selection":{"n":60,"correct":11,"oracle_correct":11,"failures":0,"posterior":{"mean":0.1935483870967742,"low":0.10599421987043235,"high":0.299786377267992,"prob_above_bar":7.274006101906912e-29,"jeffreys_mean":0.1885245901639344,"a":12,"b":50},"oracle_posterior":{"mean":0.1935483870967742,"low":0.10599421987043235,"high":0.299786377267992,"prob_above_bar":7.274006101906912e-29,"jeffreys_mean":0.1885245901639344,"a":12,"b":50},"cost_usd":0.020313762000000003,"known_cost_usd":0.020313762000000003,"unknown_costs":0,"median_ms":407.3965,"p95_ms":486.5369499999998,"missing_latency":0},"stateful_config":{"n":60,"correct":58,"oracle_correct":58,"failures":0,"posterior":{"mean":0.9516129032258065,"low":0.8865278328118185,"high":0.9897406848316792,"prob_above_bar":0.998722215153117,"jeffreys_mean":0.9590163934426229,"a":59,"b":3},"oracle_posterior":{"mean":0.9516129032258065,"low":0.8865278328118185,"high":0.9897406848316792,"prob_above_bar":0.998722215153117,"jeffreys_mean":0.9590163934426229,"a":59,"b":3},"cost_usd":0.008669724000000002,"known_cost_usd":0.008669724000000002,"unknown_costs":0,"median_ms":332.51800000000003,"p95_ms":413.56095,"missing_latency":0},"ticket_triage":{"n":60,"correct":59,"oracle_correct":59,"failures":0,"posterior":{"mean":0.967741935483871,"low":0.9120118770772547,"high":0.9960043968326835,"prob_above_bar":0.9998047828706151,"jeffreys_mean":0.9754098360655737,"a":60,"b":2},"oracle_posterior":{"mean":0.967741935483871,"low":0.9120118770772547,"high":0.9960043968326835,"prob_above_bar":0.9998047828706151,"jeffreys_mean":0.9754098360655737,"a":60,"b":2},"cost_usd":0.0047983320000000015,"known_cost_usd":0.0047983320000000015,"unknown_costs":0,"median_ms":330.5425,"p95_ms":407.15799999999996,"missing_latency":0}},"gate":{"threshold":0.72,"calibration_accepted":72,"accepted":153,"errors":1,"n":240,"posterior":{"mean":0.9870967741935484,"low":0.9643529687093082,"high":0.9984233246043949,"prob_above_bar":0.9999999999796523,"jeffreys_mean":0.9902597402597403,"a":153,"b":2}},"risk_curve":[{"threshold":1.0,"accepted":43,"errors":0,"coverage":0.17916666666666667,"risk":0.0},{"threshold":0.99,"accepted":62,"errors":0,"coverage":0.25833333333333336,"risk":0.0},{"threshold":0.98,"accepted":73,"errors":0,"coverage":0.30416666666666664,"risk":0.0},{"threshold":0.97,"accepted":82,"errors":0,"coverage":0.3416666666666667,"risk":0.0},{"threshold":0.96,"accepted":92,"errors":0,"coverage":0.38333333333333336,"risk":0.0},{"threshold":0.95,"accepted":100,"errors":0,"coverage":0.4166666666666667,"risk":0.0},{"threshold":0.94,"accepted":108,"errors":0,"coverage":0.45,"risk":0.0},{"threshold":0.9393939393939393,"accepted":109,"errors":0,"coverage":0.45416666666666666,"risk":0.0},{"threshold":0.93,"accepted":112,"errors":0,"coverage":0.4666666666666667,"risk":0.0},{"threshold":0.92,"accepted":120,"errors":0,"coverage":0.5,"risk":0.0},{"threshold":0.91,"accepted":122,"errors":0,"coverage":0.5083333333333333,"risk":0.0},{"threshold":0.9,"accepted":123,"errors":0,"coverage":0.5125,"risk":0.0},{"threshold":0.89,"accepted":126,"errors":0,"coverage":0.525,"risk":0.0},{"threshold":0.88,"accepted":128,"errors":0,"coverage":0.5333333333333333,"risk":0.0},{"threshold":0.86,"accepted":133,"errors":0,"coverage":0.5541666666666667,"risk":0.0},{"threshold":0.85,"accepted":136,"errors":0,"coverage":0.5666666666666667,"risk":0.0},{"threshold":0.84,"accepted":138,"errors":0,"coverage":0.575,"risk":0.0},{"threshold":0.8200000000000001,"accepted":139,"errors":0,"coverage":0.5791666666666667,"risk":0.0},{"threshold":0.81,"accepted":140,"errors":0,"coverage":0.5833333333333334,"risk":0.0},{"threshold":0.78,"accepted":144,"errors":0,"coverage":0.6,"risk":0.0},{"threshold":0.77,"accepted":145,"errors":0,"coverage":0.6041666666666666,"risk":0.0},{"threshold":0.76,"accepted":146,"errors":0,"coverage":0.6083333333333333,"risk":0.0},{"threshold":0.75,"accepted":148,"errors":1,"coverage":0.6166666666666667,"risk":0.006756756756756757},{"threshold":0.74,"accepted":150,"errors":1,"coverage":0.625,"risk":0.006666666666666667},{"threshold":0.73,"accepted":151,"errors":1,"coverage":0.6291666666666667,"risk":0.006622516556291391},{"threshold":0.72,"accepted":153,"errors":1,"coverage":0.6375,"risk":0.006535947712418301},{"threshold":0.71,"accepted":154,"errors":1,"coverage":0.6416666666666667,"risk":0.006493506493506494},{"threshold":0.7,"accepted":156,"errors":1,"coverage":0.65,"risk":0.00641025641025641},{"threshold":0.6900000000000001,"accepted":158,"errors":2,"coverage":0.6583333333333333,"risk":0.012658227848101266},{"threshold":0.69,"accepted":160,"errors":3,"coverage":0.6666666666666666,"risk":0.01875},{"threshold":0.67,"accepted":163,"errors":4,"coverage":0.6791666666666667,"risk":0.024539877300613498},{"threshold":0.66,"accepted":166,"errors":4,"coverage":0.6916666666666667,"risk":0.024096385542168676},{"threshold":0.64,"accepted":169,"errors":5,"coverage":0.7041666666666667,"risk":0.029585798816568046},{"threshold":0.63,"accepted":171,"errors":6,"coverage":0.7125,"risk":0.03508771929824561},{"threshold":0.61,"accepted":174,"errors":7,"coverage":0.725,"risk":0.040229885057471264},{"threshold":0.6,"accepted":177,"errors":8,"coverage":0.7375,"risk":0.04519774011299435},{"threshold":0.58,"accepted":179,"errors":10,"coverage":0.7458333333333333,"risk":0.055865921787709494},{"threshold":0.56,"accepted":182,"errors":12,"coverage":0.7583333333333333,"risk":0.06593406593406594},{"threshold":0.5555555555555556,"accepted":183,"errors":12,"coverage":0.7625,"risk":0.06557377049180328},{"threshold":0.5555555555555555,"accepted":184,"errors":13,"coverage":0.7666666666666667,"risk":0.07065217391304347},{"threshold":0.55,"accepted":186,"errors":14,"coverage":0.775,"risk":0.07526881720430108},{"threshold":0.54,"accepted":187,"errors":15,"coverage":0.7791666666666667,"risk":0.08021390374331551},{"threshold":0.53,"accepted":191,"errors":17,"coverage":0.7958333333333333,"risk":0.08900523560209424},{"threshold":0.52,"accepted":192,"errors":18,"coverage":0.8,"risk":0.09375},{"threshold":0.51,"accepted":197,"errors":22,"coverage":0.8208333333333333,"risk":0.1116751269035533},{"threshold":0.5,"accepted":200,"errors":23,"coverage":0.8333333333333334,"risk":0.115},{"threshold":0.49,"accepted":201,"errors":24,"coverage":0.8375,"risk":0.11940298507462686},{"threshold":0.48,"accepted":202,"errors":24,"coverage":0.8416666666666667,"risk":0.1188118811881188},{"threshold":0.45,"accepted":204,"errors":26,"coverage":0.85,"risk":0.12745098039215685},{"threshold":0.43,"accepted":205,"errors":27,"coverage":0.8541666666666666,"risk":0.13170731707317074},{"threshold":0.42,"accepted":207,"errors":29,"coverage":0.8625,"risk":0.14009661835748793},{"threshold":0.4,"accepted":208,"errors":30,"coverage":0.8666666666666667,"risk":0.14423076923076922},{"threshold":0.38,"accepted":210,"errors":32,"coverage":0.875,"risk":0.1523809523809524},{"threshold":0.36999999999999994,"accepted":212,"errors":34,"coverage":0.8833333333333333,"risk":0.16037735849056603},{"threshold":0.36,"accepted":213,"errors":35,"coverage":0.8875,"risk":0.1643192488262911},{"threshold":0.3434343434343434,"accepted":214,"errors":36,"coverage":0.8916666666666667,"risk":0.16822429906542055},{"threshold":0.33333333333333326,"accepted":215,"errors":37,"coverage":0.8958333333333334,"risk":0.17209302325581396},{"threshold":0.32999999999999996,"accepted":216,"errors":38,"coverage":0.9,"risk":0.17592592592592593},{"threshold":0.32,"accepted":217,"errors":38,"coverage":0.9041666666666667,"risk":0.17511520737327188},{"threshold":0.30000000000000004,"accepted":218,"errors":39,"coverage":0.9083333333333333,"risk":0.17889908256880735},{"threshold":0.29000000000000004,"accepted":219,"errors":40,"coverage":0.9125,"risk":0.182648401826484},{"threshold":0.27,"accepted":222,"errors":43,"coverage":0.925,"risk":0.19369369369369369},{"threshold":0.2626262626262626,"accepted":223,"errors":44,"coverage":0.9291666666666667,"risk":0.19730941704035873},{"threshold":0.25999999999999995,"accepted":224,"errors":45,"coverage":0.9333333333333333,"risk":0.20089285714285715},{"threshold":0.24999999999999994,"accepted":225,"errors":46,"coverage":0.9375,"risk":0.20444444444444446},{"threshold":0.24242424242424243,"accepted":226,"errors":47,"coverage":0.9416666666666667,"risk":0.2079646017699115},{"threshold":0.22999999999999998,"accepted":227,"errors":48,"coverage":0.9458333333333333,"risk":0.21145374449339208},{"threshold":0.22,"accepted":232,"errors":52,"coverage":0.9666666666666667,"risk":0.22413793103448276},{"threshold":0.21000000000000002,"accepted":233,"errors":53,"coverage":0.9708333333333333,"risk":0.22746781115879827},{"threshold":0.21,"accepted":234,"errors":54,"coverage":0.975,"risk":0.23076923076923078},{"threshold":0.202020202020202,"accepted":235,"errors":55,"coverage":0.9791666666666666,"risk":0.23404255319148937},{"threshold":0.19999999999999998,"accepted":236,"errors":56,"coverage":0.9833333333333333,"risk":0.23728813559322035},{"threshold":0.19,"accepted":237,"errors":57,"coverage":0.9875,"risk":0.24050632911392406},{"threshold":0.18999999999999995,"accepted":238,"errors":58,"coverage":0.9916666666666667,"risk":0.24369747899159663},{"threshold":0.17171717171717168,"accepted":239,"errors":59,"coverage":0.9958333333333333,"risk":0.24686192468619247},{"threshold":0.13999999999999996,"accepted":240,"errors":60,"coverage":1.0,"risk":0.25}]},{"cohort":"applicability","variant":"responses-zero","historical":false,"n":240,"correct":236,"oracle_correct":235,"failures":0,"posterior":{"mean":0.9793388429752066,"low":0.9580513834044089,"high":0.9932300389537785,"prob_above_bar":0.9999999999999802,"jeffreys_mean":0.9813278008298755,"a":237,"b":5},"oracle_posterior":{"mean":0.9752066115702479,"low":0.9522504798872883,"high":0.9908099171423539,"prob_above_bar":0.9999999999998084,"jeffreys_mean":0.9771784232365145,"a":236,"b":6},"cost_usd":0.029699700000000002,"known_cost_usd":0.029699700000000002,"unknown_costs":0,"median_ms":1535.9118329943158,"p95_ms":2894.9222771159834,"missing_latency":0,"tasks":{"approval_policy":{"n":60,"correct":56,"oracle_correct":55,"failures":0,"posterior":{"mean":0.9193548387096774,"low":0.8405310453592895,"high":0.972848639604444,"prob_above_bar":0.9821156263666959,"jeffreys_mean":0.9262295081967213,"a":57,"b":5},"oracle_posterior":{"mean":0.9032258064516129,"low":0.8189880888902016,"high":0.9630400051221482,"prob_above_bar":0.9539610712241788,"jeffreys_mean":0.9098360655737705,"a":56,"b":6},"cost_usd":0.007348599999999998,"known_cost_usd":0.007348599999999998,"unknown_costs":0,"median_ms":1418.5676669585519,"p95_ms":2194.6083481831006,"missing_latency":0},"project_selection":{"n":60,"correct":60,"oracle_correct":60,"failures":0,"posterior":{"mean":0.9838709677419355,"low":0.9413187760677074,"high":0.9995850400871603,"prob_above_bar":0.9999852108235314,"jeffreys_mean":0.9918032786885246,"a":61,"b":1},"oracle_posterior":{"mean":0.9838709677419355,"low":0.9413187760677074,"high":0.9995850400871603,"prob_above_bar":0.9999852108235314,"jeffreys_mean":0.9918032786885246,"a":61,"b":1},"cost_usd":0.010786400000000002,"known_cost_usd":0.010786400000000002,"unknown_costs":0,"median_ms":2161.5795204997994,"p95_ms":3693.027524970238,"missing_latency":0},"stateful_config":{"n":60,"correct":60,"oracle_correct":60,"failures":0,"posterior":{"mean":0.9838709677419355,"low":0.9413187760677074,"high":0.9995850400871603,"prob_above_bar":0.9999852108235314,"jeffreys_mean":0.9918032786885246,"a":61,"b":1},"oracle_posterior":{"mean":0.9838709677419355,"low":0.9413187760677074,"high":0.9995850400871603,"prob_above_bar":0.9999852108235314,"jeffreys_mean":0.9918032786885246,"a":61,"b":1},"cost_usd":0.006398500000000002,"known_cost_usd":0.006398500000000002,"unknown_costs":0,"median_ms":1531.3605830306187,"p95_ms":1997.4111540592264,"missing_latency":0},"ticket_triage":{"n":60,"correct":60,"oracle_correct":60,"failures":0,"posterior":{"mean":0.9838709677419355,"low":0.9413187760677074,"high":0.9995850400871603,"prob_above_bar":0.9999852108235314,"jeffreys_mean":0.9918032786885246,"a":61,"b":1},"oracle_posterior":{"mean":0.9838709677419355,"low":0.9413187760677074,"high":0.9995850400871603,"prob_above_bar":0.9999852108235314,"jeffreys_mean":0.9918032786885246,"a":61,"b":1},"cost_usd":0.005166199999999999,"known_cost_usd":0.005166199999999999,"unknown_costs":0,"median_ms":1327.935291454196,"p95_ms":2112.894129072082,"missing_latency":0}},"gate":null,"risk_curve":[]},{"cohort":"applicability","variant":"decisions-zero","historical":false,"n":240,"correct":157,"oracle_correct":158,"failures":11,"posterior":{"mean":0.6528925619834711,"low":0.5919024035250231,"high":0.7114888747534849,"prob_above_bar":6.357167496034258e-12,"jeffreys_mean":0.6535269709543569,"a":158,"b":84},"oracle_posterior":{"mean":0.6570247933884298,"low":0.5961720178351235,"high":0.7154190433220925,"prob_above_bar":1.7076093454253663e-11,"jeffreys_mean":0.6576763485477178,"a":159,"b":83},"cost_usd":0.09482120000000001,"known_cost_usd":0.09482120000000001,"unknown_costs":0,"median_ms":155.86422896012664,"p95_ms":216.34925599209961,"missing_latency":0,"tasks":{"approval_policy":{"n":60,"correct":43,"oracle_correct":44,"failures":0,"posterior":{"mean":0.7096774193548387,"low":0.5917125332438207,"high":0.814833807668532,"prob_above_bar":0.008939638710481472,"jeffreys_mean":0.7131147540983607,"a":44,"b":18},"oracle_posterior":{"mean":0.7258064516129032,"low":0.6092941471591897,"high":0.8285279157397194,"prob_above_bar":0.019345804465377234,"jeffreys_mean":0.7295081967213115,"a":45,"b":17},"cost_usd":0.020448099999999997,"known_cost_usd":0.020448099999999997,"unknown_costs":0,"median_ms":148.87825003825128,"p95_ms":190.35327077144754,"missing_latency":0},"project_selection":{"n":60,"correct":5,"oracle_correct":5,"failures":10,"posterior":{"mean":0.0967741935483871,"low":0.036959994877851884,"high":0.1810119111097984,"prob_above_bar":6.452617352117452e-38,"jeffreys_mean":0.09016393442622951,"a":6,"b":56},"oracle_posterior":{"mean":0.0967741935483871,"low":0.036959994877851884,"high":0.1810119111097984,"prob_above_bar":6.452617352117452e-38,"jeffreys_mean":0.09016393442622951,"a":6,"b":56},"cost_usd":0.044870400000000005,"known_cost_usd":0.044870400000000005,"unknown_costs":0,"median_ms":169.4648130214773,"p95_ms":258.7569813360459,"missing_latency":0},"stateful_config":{"n":60,"correct":50,"oracle_correct":50,"failures":1,"posterior":{"mean":0.8225806451612904,"low":0.719114954272785,"high":0.9063902302056501,"prob_above_bar":0.4394489639989438,"jeffreys_mean":0.8278688524590164,"a":51,"b":11},"oracle_posterior":{"mean":0.8225806451612904,"low":0.719114954272785,"high":0.9063902302056501,"prob_above_bar":0.4394489639989438,"jeffreys_mean":0.8278688524590164,"a":51,"b":11},"cost_usd":0.019149,"known_cost_usd":0.019149,"unknown_costs":0,"median_ms":154.24135403009132,"p95_ms":208.69518747203978,"missing_latency":0},"ticket_triage":{"n":60,"correct":59,"oracle_correct":59,"failures":0,"posterior":{"mean":0.967741935483871,"low":0.9120118770772547,"high":0.9960043968326835,"prob_above_bar":0.9998047828706151,"jeffreys_mean":0.9754098360655737,"a":60,"b":2},"oracle_posterior":{"mean":0.967741935483871,"low":0.9120118770772547,"high":0.9960043968326835,"prob_above_bar":0.9998047828706151,"jeffreys_mean":0.9754098360655737,"a":60,"b":2},"cost_usd":0.010353700000000002,"known_cost_usd":0.010353700000000002,"unknown_costs":0,"median_ms":147.68635405926034,"p95_ms":190.55208571371614,"missing_latency":0}},"gate":{"threshold":0.71,"calibration_accepted":50,"minimum":12,"planned_calibration":96,"accepted":113,"errors":2,"n":240,"posterior":{"mean":0.9739130434782609,"low":0.9380613067700988,"high":0.9945398216817188,"prob_above_bar":0.9999997353183869,"jeffreys_mean":0.9780701754385965,"a":112,"b":3}},"risk_curve":[{"threshold":1.0,"accepted":28,"errors":0,"coverage":0.11666666666666667,"risk":0.0},{"threshold":0.99,"accepted":49,"errors":0,"coverage":0.20416666666666666,"risk":0.0},{"threshold":0.98,"accepted":57,"errors":0,"coverage":0.2375,"risk":0.0},{"threshold":0.97,"accepted":63,"errors":0,"coverage":0.2625,"risk":0.0},{"threshold":0.96,"accepted":67,"errors":0,"coverage":0.2791666666666667,"risk":0.0},{"threshold":0.95,"accepted":72,"errors":0,"coverage":0.3,"risk":0.0},{"threshold":0.94,"accepted":80,"errors":0,"coverage":0.3333333333333333,"risk":0.0},{"threshold":0.93,"accepted":86,"errors":0,"coverage":0.35833333333333334,"risk":0.0},{"threshold":0.92,"accepted":88,"errors":0,"coverage":0.36666666666666664,"risk":0.0},{"threshold":0.91,"accepted":90,"errors":1,"coverage":0.375,"risk":0.011111111111111112},{"threshold":0.9,"accepted":91,"errors":2,"coverage":0.37916666666666665,"risk":0.02197802197802198},{"threshold":0.89,"accepted":93,"errors":2,"coverage":0.3875,"risk":0.021505376344086023},{"threshold":0.88,"accepted":94,"errors":2,"coverage":0.39166666666666666,"risk":0.02127659574468085},{"threshold":0.87,"accepted":95,"errors":2,"coverage":0.3958333333333333,"risk":0.021052631578947368},{"threshold":0.86,"accepted":97,"errors":2,"coverage":0.4041666666666667,"risk":0.020618556701030927},{"threshold":0.85,"accepted":99,"errors":2,"coverage":0.4125,"risk":0.020202020202020204},{"threshold":0.84,"accepted":100,"errors":2,"coverage":0.4166666666666667,"risk":0.02},{"threshold":0.83,"accepted":101,"errors":2,"coverage":0.42083333333333334,"risk":0.019801980198019802},{"threshold":0.82,"accepted":102,"errors":2,"coverage":0.425,"risk":0.0196078431372549},{"threshold":0.8,"accepted":104,"errors":2,"coverage":0.43333333333333335,"risk":0.019230769230769232},{"threshold":0.79,"accepted":106,"errors":2,"coverage":0.44166666666666665,"risk":0.018867924528301886},{"threshold":0.75,"accepted":107,"errors":2,"coverage":0.44583333333333336,"risk":0.018691588785046728},{"threshold":0.74,"accepted":108,"errors":2,"coverage":0.45,"risk":0.018518518518518517},{"threshold":0.73,"accepted":110,"errors":2,"coverage":0.4583333333333333,"risk":0.01818181818181818},{"threshold":0.72,"accepted":111,"errors":2,"coverage":0.4625,"risk":0.018018018018018018},{"threshold":0.71,"accepted":113,"errors":2,"coverage":0.4708333333333333,"risk":0.017699115044247787},{"threshold":0.7,"accepted":115,"errors":2,"coverage":0.4791666666666667,"risk":0.017391304347826087},{"threshold":0.68,"accepted":119,"errors":3,"coverage":0.49583333333333335,"risk":0.025210084033613446},{"threshold":0.67,"accepted":122,"errors":5,"coverage":0.5083333333333333,"risk":0.040983606557377046},{"threshold":0.66,"accepted":123,"errors":5,"coverage":0.5125,"risk":0.04065040650406504},{"threshold":0.65,"accepted":127,"errors":6,"coverage":0.5291666666666667,"risk":0.047244094488188976},{"threshold":0.64,"accepted":129,"errors":6,"coverage":0.5375,"risk":0.046511627906976744},{"threshold":0.62,"accepted":130,"errors":6,"coverage":0.5416666666666666,"risk":0.046153846153846156},{"threshold":0.61,"accepted":132,"errors":7,"coverage":0.55,"risk":0.05303030303030303},{"threshold":0.6000000000000001,"accepted":133,"errors":7,"coverage":0.5541666666666667,"risk":0.05263157894736842},{"threshold":0.6,"accepted":135,"errors":8,"coverage":0.5625,"risk":0.05925925925925926},{"threshold":0.59,"accepted":138,"errors":11,"coverage":0.575,"risk":0.07971014492753623},{"threshold":0.57,"accepted":139,"errors":11,"coverage":0.5791666666666667,"risk":0.07913669064748201},{"threshold":0.56,"accepted":142,"errors":12,"coverage":0.5916666666666667,"risk":0.08450704225352113},{"threshold":0.54,"accepted":145,"errors":13,"coverage":0.6041666666666666,"risk":0.0896551724137931},{"threshold":0.53,"accepted":150,"errors":13,"coverage":0.625,"risk":0.08666666666666667},{"threshold":0.52,"accepted":152,"errors":14,"coverage":0.6333333333333333,"risk":0.09210526315789473},{"threshold":0.5,"accepted":155,"errors":16,"coverage":0.6458333333333334,"risk":0.1032258064516129},{"threshold":0.49000000000000005,"accepted":156,"errors":16,"coverage":0.65,"risk":0.10256410256410256},{"threshold":0.49,"accepted":157,"errors":16,"coverage":0.6541666666666667,"risk":0.10191082802547771},{"threshold":0.48,"accepted":160,"errors":17,"coverage":0.6666666666666666,"risk":0.10625},{"threshold":0.47,"accepted":163,"errors":19,"coverage":0.6791666666666667,"risk":0.1165644171779141},{"threshold":0.46,"accepted":167,"errors":22,"coverage":0.6958333333333333,"risk":0.1317365269461078},{"threshold":0.45,"accepted":169,"errors":23,"coverage":0.7041666666666667,"risk":0.13609467455621302},{"threshold":0.4499999999999999,"accepted":170,"errors":23,"coverage":0.7083333333333334,"risk":0.13529411764705881},{"threshold":0.44,"accepted":173,"errors":25,"coverage":0.7208333333333333,"risk":0.14450867052023122},{"threshold":0.43,"accepted":177,"errors":27,"coverage":0.7375,"risk":0.15254237288135594},{"threshold":0.42,"accepted":179,"errors":27,"coverage":0.7458333333333333,"risk":0.15083798882681565},{"threshold":0.41,"accepted":180,"errors":28,"coverage":0.75,"risk":0.15555555555555556},{"threshold":0.4,"accepted":182,"errors":30,"coverage":0.7583333333333333,"risk":0.16483516483516483},{"threshold":0.3999999999999999,"accepted":183,"errors":31,"coverage":0.7625,"risk":0.16939890710382513},{"threshold":0.39,"accepted":186,"errors":33,"coverage":0.775,"risk":0.1774193548387097},{"threshold":0.38,"accepted":187,"errors":34,"coverage":0.7791666666666667,"risk":0.18181818181818182},{"threshold":0.3799999999999999,"accepted":188,"errors":34,"coverage":0.7833333333333333,"risk":0.18085106382978725},{"threshold":0.37,"accepted":189,"errors":34,"coverage":0.7875,"risk":0.17989417989417988},{"threshold":0.3499999999999999,"accepted":190,"errors":34,"coverage":0.7916666666666666,"risk":0.17894736842105263},{"threshold":0.32999999999999996,"accepted":191,"errors":35,"coverage":0.7958333333333333,"risk":0.18324607329842932},{"threshold":0.31999999999999995,"accepted":192,"errors":36,"coverage":0.8,"risk":0.1875},{"threshold":0.29,"accepted":193,"errors":37,"coverage":0.8041666666666667,"risk":0.19170984455958548},{"threshold":0.2899999999999999,"accepted":194,"errors":38,"coverage":0.8083333333333333,"risk":0.1958762886597938},{"threshold":0.28,"accepted":195,"errors":39,"coverage":0.8125,"risk":0.2},{"threshold":0.26999999999999996,"accepted":196,"errors":40,"coverage":0.8166666666666667,"risk":0.20408163265306123},{"threshold":0.25999999999999995,"accepted":197,"errors":41,"coverage":0.8208333333333333,"risk":0.20812182741116753},{"threshold":0.24,"accepted":198,"errors":41,"coverage":0.825,"risk":0.20707070707070707},{"threshold":0.23999999999999994,"accepted":199,"errors":42,"coverage":0.8291666666666667,"risk":0.21105527638190955},{"threshold":0.22999999999999995,"accepted":201,"errors":44,"coverage":0.8375,"risk":0.21890547263681592},{"threshold":0.22,"accepted":202,"errors":45,"coverage":0.8416666666666667,"risk":0.22277227722772278},{"threshold":0.20000000000000004,"accepted":203,"errors":46,"coverage":0.8458333333333333,"risk":0.22660098522167488},{"threshold":0.2,"accepted":205,"errors":48,"coverage":0.8541666666666666,"risk":0.23414634146341465},{"threshold":0.19999999999999996,"accepted":207,"errors":50,"coverage":0.8625,"risk":0.24154589371980675},{"threshold":0.19,"accepted":208,"errors":51,"coverage":0.8666666666666667,"risk":0.24519230769230768},{"threshold":0.18999999999999995,"accepted":209,"errors":52,"coverage":0.8708333333333333,"risk":0.24880382775119617},{"threshold":0.18999999999999992,"accepted":211,"errors":54,"coverage":0.8791666666666667,"risk":0.2559241706161137},{"threshold":0.16999999999999998,"accepted":212,"errors":55,"coverage":0.8833333333333333,"risk":0.25943396226415094},{"threshold":0.16,"accepted":213,"errors":56,"coverage":0.8875,"risk":0.26291079812206575},{"threshold":0.15999999999999998,"accepted":214,"errors":57,"coverage":0.8916666666666667,"risk":0.26635514018691586},{"threshold":0.15,"accepted":215,"errors":58,"coverage":0.8958333333333334,"risk":0.26976744186046514},{"threshold":0.14999999999999997,"accepted":216,"errors":59,"coverage":0.9,"risk":0.27314814814814814},{"threshold":0.14,"accepted":217,"errors":60,"coverage":0.9041666666666667,"risk":0.2764976958525346},{"threshold":0.13999999999999999,"accepted":218,"errors":61,"coverage":0.9083333333333333,"risk":0.2798165137614679},{"threshold":0.12999999999999998,"accepted":220,"errors":63,"coverage":0.9166666666666666,"risk":0.2863636363636364},{"threshold":0.12999999999999995,"accepted":221,"errors":64,"coverage":0.9208333333333333,"risk":0.2895927601809955},{"threshold":0.11999999999999997,"accepted":223,"errors":66,"coverage":0.9291666666666667,"risk":0.29596412556053814},{"threshold":0.11,"accepted":224,"errors":67,"coverage":0.9333333333333333,"risk":0.29910714285714285},{"threshold":0.10999999999999997,"accepted":226,"errors":69,"coverage":0.9416666666666667,"risk":0.3053097345132743},{"threshold":0.10999999999999995,"accepted":227,"errors":70,"coverage":0.9458333333333333,"risk":0.30837004405286345},{"threshold":0.09999999999999998,"accepted":228,"errors":71,"coverage":0.95,"risk":0.31140350877192985},{"threshold":0.08999999999999996,"accepted":229,"errors":72,"coverage":0.9541666666666667,"risk":0.314410480349345}]}],"routing":[{"cohort":"routing","task":"feedback","variant":"qwen3:0.6b","historical":false,"n":48,"correct":41,"oracle_correct":41,"failures":0,"posterior":{"mean":0.84,"low":0.7275784873044024,"high":0.9267772400920117,"prob_above_bar":0.5847767159387696,"jeffreys_mean":0.8469387755102041,"a":42,"b":8},"oracle_posterior":{"mean":0.84,"low":0.7275784873044024,"high":0.9267772400920117,"prob_above_bar":0.5847767159387696,"jeffreys_mean":0.8469387755102041,"a":42,"b":8},"cost_usd":0,"known_cost_usd":0,"unknown_costs":0,"median_ms":160.21518752677366,"p95_ms":164.918825001223,"missing_latency":0,"cost_basis":"single-call mean, three calls per item for cloud; local hardware/energy unknown"},{"cohort":"routing","task":"feedback","variant":"archive-qwen3:0.6b","historical":true,"n":48,"correct":41,"oracle_correct":41,"failures":0,"posterior":{"mean":0.84,"low":0.7275784873044024,"high":0.9267772400920117,"prob_above_bar":0.5847767159387696,"jeffreys_mean":0.8469387755102041,"a":42,"b":8},"oracle_posterior":{"mean":0.84,"low":0.7275784873044024,"high":0.9267772400920117,"prob_above_bar":0.5847767159387696,"jeffreys_mean":0.8469387755102041,"a":42,"b":8},"cost_usd":0,"known_cost_usd":0,"unknown_costs":0,"median_ms":null,"p95_ms":null,"missing_latency":48,"cost_basis":"single-call mean, three calls per item for cloud; local hardware/energy unknown"},{"cohort":"routing","task":"feedback","variant":"qwen3:1.7b","historical":false,"n":48,"correct":47,"oracle_correct":47,"failures":0,"posterior":{"mean":0.96,"low":0.891458236031593,"high":0.9950181581650287,"prob_above_bar":0.9985758927442414,"jeffreys_mean":0.9693877551020408,"a":48,"b":2},"oracle_posterior":{"mean":0.96,"low":0.891458236031593,"high":0.9950181581650287,"prob_above_bar":0.9985758927442414,"jeffreys_mean":0.9693877551020408,"a":48,"b":2},"cost_usd":0,"known_cost_usd":0,"unknown_costs":0,"median_ms":189.15956252021715,"p95_ms":193.9093187567778,"missing_latency":0,"cost_basis":"single-call mean, three calls per item for cloud; local hardware/energy unknown"},{"cohort":"routing","task":"feedback","variant":"archive-qwen3:1.7b","historical":true,"n":48,"correct":47,"oracle_correct":47,"failures":0,"posterior":{"mean":0.96,"low":0.891458236031593,"high":0.9950181581650287,"prob_above_bar":0.9985758927442414,"jeffreys_mean":0.9693877551020408,"a":48,"b":2},"oracle_posterior":{"mean":0.96,"low":0.891458236031593,"high":0.9950181581650287,"prob_above_bar":0.9985758927442414,"jeffreys_mean":0.9693877551020408,"a":48,"b":2},"cost_usd":0,"known_cost_usd":0,"unknown_costs":0,"median_ms":null,"p95_ms":null,"missing_latency":48,"cost_basis":"single-call mean, three calls per item for cloud; local hardware/energy unknown"},{"cohort":"routing","task":"feedback","variant":"qwen3:4b","historical":false,"n":48,"correct":48,"oracle_correct":48,"failures":0,"posterior":{"mean":0.98,"low":0.9274807363357226,"high":0.9994834435024067,"prob_above_bar":0.9998681382170594,"jeffreys_mean":0.9897959183673469,"a":49,"b":1},"oracle_posterior":{"mean":0.98,"low":0.9274807363357226,"high":0.9994834435024067,"prob_above_bar":0.9998681382170594,"jeffreys_mean":0.9897959183673469,"a":49,"b":1},"cost_usd":0,"known_cost_usd":0,"unknown_costs":0,"median_ms":313.7360834516585,"p95_ms":324.6793459111359,"missing_latency":0,"cost_basis":"single-call mean, three calls per item for cloud; local hardware/energy unknown"},{"cohort":"routing","task":"feedback","variant":"archive-qwen3:4b","historical":true,"n":48,"correct":48,"oracle_correct":48,"failures":0,"posterior":{"mean":0.98,"low":0.9274807363357226,"high":0.9994834435024067,"prob_above_bar":0.9998681382170594,"jeffreys_mean":0.9897959183673469,"a":49,"b":1},"oracle_posterior":{"mean":0.98,"low":0.9274807363357226,"high":0.9994834435024067,"prob_above_bar":0.9998681382170594,"jeffreys_mean":0.9897959183673469,"a":49,"b":1},"cost_usd":0,"known_cost_usd":0,"unknown_costs":0,"median_ms":null,"p95_ms":null,"missing_latency":48,"cost_basis":"single-call mean, three calls per item for cloud; local hardware/energy unknown"},{"cohort":"routing","task":"feedback","variant":"qwen3:8b","historical":false,"n":48,"correct":48,"oracle_correct":48,"failures":0,"posterior":{"mean":0.98,"low":0.9274807363357226,"high":0.9994834435024067,"prob_above_bar":0.9998681382170594,"jeffreys_mean":0.9897959183673469,"a":49,"b":1},"oracle_posterior":{"mean":0.98,"low":0.9274807363357226,"high":0.9994834435024067,"prob_above_bar":0.9998681382170594,"jeffreys_mean":0.9897959183673469,"a":49,"b":1},"cost_usd":0,"known_cost_usd":0,"unknown_costs":0,"median_ms":348.81539549678564,"p95_ms":356.86345012509264,"missing_latency":0,"cost_basis":"single-call mean, three calls per item for cloud; local hardware/energy unknown"},{"cohort":"routing","task":"feedback","variant":"archive-qwen3:8b","historical":true,"n":48,"correct":48,"oracle_correct":48,"failures":0,"posterior":{"mean":0.98,"low":0.9274807363357226,"high":0.9994834435024067,"prob_above_bar":0.9998681382170594,"jeffreys_mean":0.9897959183673469,"a":49,"b":1},"oracle_posterior":{"mean":0.98,"low":0.9274807363357226,"high":0.9994834435024067,"prob_above_bar":0.9998681382170594,"jeffreys_mean":0.9897959183673469,"a":49,"b":1},"cost_usd":0,"known_cost_usd":0,"unknown_costs":0,"median_ms":null,"p95_ms":null,"missing_latency":48,"cost_basis":"single-call mean, three calls per item for cloud; local hardware/energy unknown"},{"cohort":"routing","task":"feedback","variant":"qwen3-coder:30b","historical":false,"n":48,"correct":47,"oracle_correct":47,"failures":0,"posterior":{"mean":0.96,"low":0.891458236031593,"high":0.9950181581650287,"prob_above_bar":0.9985758927442414,"jeffreys_mean":0.9693877551020408,"a":48,"b":2},"oracle_posterior":{"mean":0.96,"low":0.891458236031593,"high":0.9950181581650287,"prob_above_bar":0.9985758927442414,"jeffreys_mean":0.9693877551020408,"a":48,"b":2},"cost_usd":0,"known_cost_usd":0,"unknown_costs":0,"median_ms":323.5358750098385,"p95_ms":361.50316885323264,"missing_latency":0,"cost_basis":"single-call mean, three calls per item for cloud; local hardware/energy unknown"},{"cohort":"routing","task":"feedback","variant":"archive-qwen3-coder:30b","historical":true,"n":48,"correct":47,"oracle_correct":47,"failures":0,"posterior":{"mean":0.96,"low":0.891458236031593,"high":0.9950181581650287,"prob_above_bar":0.9985758927442414,"jeffreys_mean":0.9693877551020408,"a":48,"b":2},"oracle_posterior":{"mean":0.96,"low":0.891458236031593,"high":0.9950181581650287,"prob_above_bar":0.9985758927442414,"jeffreys_mean":0.9693877551020408,"a":48,"b":2},"cost_usd":0,"known_cost_usd":0,"unknown_costs":0,"median_ms":null,"p95_ms":null,"missing_latency":48,"cost_basis":"single-call mean, three calls per item for cloud; local hardware/energy unknown"},{"cohort":"routing","task":"feedback","variant":"decisions","historical":false,"n":48,"correct":46,"oracle_correct":46,"failures":0,"posterior":{"mean":0.94,"low":0.8602128129775362,"high":0.9871915786817211,"prob_above_bar":0.9923731144747153,"jeffreys_mean":0.9489795918367347,"a":47,"b":3},"oracle_posterior":{"mean":0.94,"low":0.8602128129775362,"high":0.9871915786817211,"prob_above_bar":0.9923731144747153,"jeffreys_mean":0.9489795918367347,"a":47,"b":3},"cost_usd":0.0009116000000000001,"known_cost_usd":0.0009116000000000001,"unknown_costs":0,"median_ms":139.87163217583048,"p95_ms":171.0393742308952,"missing_latency":0,"cost_basis":"single-call mean, three calls per item for cloud; local hardware/energy unknown"},{"cohort":"routing","task":"feedback","variant":"responses","historical":false,"n":48,"correct":48,"oracle_correct":48,"failures":0,"posterior":{"mean":0.98,"low":0.9274807363357226,"high":0.9994834435024067,"prob_above_bar":0.9998681382170594,"jeffreys_mean":0.9897959183673469,"a":49,"b":1},"oracle_posterior":{"mean":0.98,"low":0.9274807363357226,"high":0.9994834435024067,"prob_above_bar":0.9998681382170594,"jeffreys_mean":0.9897959183673469,"a":49,"b":1},"cost_usd":0.0008575999999999996,"known_cost_usd":0.0008575999999999996,"unknown_costs":0,"median_ms":864.2193886335008,"p95_ms":1092.965225848214,"missing_latency":0,"cost_basis":"single-call mean, three calls per item for cloud; local hardware/energy unknown"},{"cohort":"routing","task":"triage","variant":"qwen3:0.6b","historical":false,"n":48,"correct":27,"oracle_correct":26,"failures":0,"posterior":{"mean":0.56,"low":0.4221040520359127,"high":0.6933462477210417,"prob_above_bar":3.5680982335791605e-06,"jeffreys_mean":0.5612244897959183,"a":28,"b":22},"oracle_posterior":{"mean":0.54,"low":0.40230543910987826,"high":0.6746612748834462,"prob_above_bar":8.193539891071419e-07,"jeffreys_mean":0.5408163265306123,"a":27,"b":23},"cost_usd":0,"known_cost_usd":0,"unknown_costs":0,"median_ms":162.86441654665396,"p95_ms":172.77543530799448,"missing_latency":0,"cost_basis":"single-call mean, three calls per item for cloud; local hardware/energy unknown"},{"cohort":"routing","task":"triage","variant":"archive-qwen3:0.6b","historical":true,"n":48,"correct":27,"oracle_correct":26,"failures":0,"posterior":{"mean":0.56,"low":0.4221040520359127,"high":0.6933462477210417,"prob_above_bar":3.5680982335791605e-06,"jeffreys_mean":0.5612244897959183,"a":28,"b":22},"oracle_posterior":{"mean":0.54,"low":0.40230543910987826,"high":0.6746612748834462,"prob_above_bar":8.193539891071419e-07,"jeffreys_mean":0.5408163265306123,"a":27,"b":23},"cost_usd":0,"known_cost_usd":0,"unknown_costs":0,"median_ms":null,"p95_ms":null,"missing_latency":48,"cost_basis":"single-call mean, three calls per item for cloud; local hardware/energy unknown"},{"cohort":"routing","task":"triage","variant":"qwen3:1.7b","historical":false,"n":48,"correct":27,"oracle_correct":27,"failures":0,"posterior":{"mean":0.56,"low":0.4221040520359127,"high":0.6933462477210417,"prob_above_bar":3.5680982335791605e-06,"jeffreys_mean":0.5612244897959183,"a":28,"b":22},"oracle_posterior":{"mean":0.56,"low":0.4221040520359127,"high":0.6933462477210417,"prob_above_bar":3.5680982335791605e-06,"jeffreys_mean":0.5612244897959183,"a":28,"b":22},"cost_usd":0,"known_cost_usd":0,"unknown_costs":0,"median_ms":183.5464795003645,"p95_ms":200.04193127388135,"missing_latency":0,"cost_basis":"single-call mean, three calls per item for cloud; local hardware/energy unknown"},{"cohort":"routing","task":"triage","variant":"archive-qwen3:1.7b","historical":true,"n":48,"correct":27,"oracle_correct":27,"failures":0,"posterior":{"mean":0.56,"low":0.4221040520359127,"high":0.6933462477210417,"prob_above_bar":3.5680982335791605e-06,"jeffreys_mean":0.5612244897959183,"a":28,"b":22},"oracle_posterior":{"mean":0.56,"low":0.4221040520359127,"high":0.6933462477210417,"prob_above_bar":3.5680982335791605e-06,"jeffreys_mean":0.5612244897959183,"a":28,"b":22},"cost_usd":0,"known_cost_usd":0,"unknown_costs":0,"median_ms":null,"p95_ms":null,"missing_latency":48,"cost_basis":"single-call mean, three calls per item for cloud; local hardware/energy unknown"},{"cohort":"routing","task":"triage","variant":"qwen3:4b","historical":false,"n":48,"correct":37,"oracle_correct":36,"failures":0,"posterior":{"mean":0.76,"low":0.6337569811407238,"high":0.8665712891577062,"prob_above_bar":0.10419408908927974,"jeffreys_mean":0.7653061224489796,"a":38,"b":12},"oracle_posterior":{"mean":0.74,"low":0.6113040853115662,"high":0.8505292712792318,"prob_above_bar":0.054361904969024016,"jeffreys_mean":0.7448979591836735,"a":37,"b":13},"cost_usd":0,"known_cost_usd":0,"unknown_costs":0,"median_ms":307.7106249984354,"p95_ms":331.91507103620097,"missing_latency":0,"cost_basis":"single-call mean, three calls per item for cloud; local hardware/energy unknown"},{"cohort":"routing","task":"triage","variant":"archive-qwen3:4b","historical":true,"n":48,"correct":37,"oracle_correct":36,"failures":0,"posterior":{"mean":0.76,"low":0.6337569811407238,"high":0.8665712891577062,"prob_above_bar":0.10419408908927974,"jeffreys_mean":0.7653061224489796,"a":38,"b":12},"oracle_posterior":{"mean":0.74,"low":0.6113040853115662,"high":0.8505292712792318,"prob_above_bar":0.054361904969024016,"jeffreys_mean":0.7448979591836735,"a":37,"b":13},"cost_usd":0,"known_cost_usd":0,"unknown_costs":0,"median_ms":null,"p95_ms":null,"missing_latency":48,"cost_basis":"single-call mean, three calls per item for cloud; local hardware/energy unknown"},{"cohort":"routing","task":"triage","variant":"qwen3:8b","historical":false,"n":48,"correct":43,"oracle_correct":43,"failures":0,"posterior":{"mean":0.88,"low":0.7777233037173732,"high":0.9537109689181043,"prob_above_bar":0.8477739474555218,"jeffreys_mean":0.8877551020408163,"a":44,"b":6},"oracle_posterior":{"mean":0.88,"low":0.7777233037173732,"high":0.9537109689181043,"prob_above_bar":0.8477739474555218,"jeffreys_mean":0.8877551020408163,"a":44,"b":6},"cost_usd":0,"known_cost_usd":0,"unknown_costs":0,"median_ms":334.2553124530241,"p95_ms":381.5186982101295,"missing_latency":0,"cost_basis":"single-call mean, three calls per item for cloud; local hardware/energy unknown"},{"cohort":"routing","task":"triage","variant":"archive-qwen3:8b","historical":true,"n":48,"correct":43,"oracle_correct":43,"failures":0,"posterior":{"mean":0.88,"low":0.7777233037173732,"high":0.9537109689181043,"prob_above_bar":0.8477739474555218,"jeffreys_mean":0.8877551020408163,"a":44,"b":6},"oracle_posterior":{"mean":0.88,"low":0.7777233037173732,"high":0.9537109689181043,"prob_above_bar":0.8477739474555218,"jeffreys_mean":0.8877551020408163,"a":44,"b":6},"cost_usd":0,"known_cost_usd":0,"unknown_costs":0,"median_ms":null,"p95_ms":null,"missing_latency":48,"cost_basis":"single-call mean, three calls per item for cloud; local hardware/energy unknown"},{"cohort":"routing","task":"triage","variant":"qwen3-coder:30b","historical":false,"n":48,"correct":45,"oracle_correct":44,"failures":0,"posterior":{"mean":0.92,"low":0.8313410992976988,"high":0.9773091193569268,"prob_above_bar":0.9729377425635334,"jeffreys_mean":0.9285714285714286,"a":46,"b":4},"oracle_posterior":{"mean":0.9,"low":0.80398581652111,"high":0.9660282950160443,"prob_above_bar":0.928236387167815,"jeffreys_mean":0.9081632653061225,"a":45,"b":5},"cost_usd":0,"known_cost_usd":0,"unknown_costs":0,"median_ms":333.74314644606784,"p95_ms":376.85899821226485,"missing_latency":0,"cost_basis":"single-call mean, three calls per item for cloud; local hardware/energy unknown"},{"cohort":"routing","task":"triage","variant":"archive-qwen3-coder:30b","historical":true,"n":48,"correct":45,"oracle_correct":44,"failures":0,"posterior":{"mean":0.92,"low":0.8313410992976988,"high":0.9773091193569268,"prob_above_bar":0.9729377425635334,"jeffreys_mean":0.9285714285714286,"a":46,"b":4},"oracle_posterior":{"mean":0.9,"low":0.80398581652111,"high":0.9660282950160443,"prob_above_bar":0.928236387167815,"jeffreys_mean":0.9081632653061225,"a":45,"b":5},"cost_usd":0,"known_cost_usd":0,"unknown_costs":0,"median_ms":null,"p95_ms":null,"missing_latency":48,"cost_basis":"single-call mean, three calls per item for cloud; local hardware/energy unknown"},{"cohort":"routing","task":"triage","variant":"decisions","historical":false,"n":48,"correct":44,"oracle_correct":43,"failures":0,"posterior":{"mean":0.9,"low":0.80398581652111,"high":0.9660282950160443,"prob_above_bar":0.928236387167815,"jeffreys_mean":0.9081632653061225,"a":45,"b":5},"oracle_posterior":{"mean":0.88,"low":0.7777233037173732,"high":0.9537109689181043,"prob_above_bar":0.8477739474555218,"jeffreys_mean":0.8877551020408163,"a":44,"b":6},"cost_usd":0.0008598000000000002,"known_cost_usd":0.0008598000000000002,"unknown_costs":0,"median_ms":140.52379148779437,"p95_ms":175.6352087618628,"missing_latency":0,"cost_basis":"single-call mean, three calls per item for cloud; local hardware/energy unknown"},{"cohort":"routing","task":"triage","variant":"responses","historical":false,"n":48,"correct":46,"oracle_correct":45,"failures":0,"posterior":{"mean":0.94,"low":0.8602128129775362,"high":0.9871915786817211,"prob_above_bar":0.9923731144747153,"jeffreys_mean":0.9489795918367347,"a":47,"b":3},"oracle_posterior":{"mean":0.92,"low":0.8313410992976988,"high":0.9773091193569268,"prob_above_bar":0.9729377425635334,"jeffreys_mean":0.9285714285714286,"a":46,"b":4},"cost_usd":0.0008138333333333336,"known_cost_usd":0.0008138333333333336,"unknown_costs":0,"median_ms":857.3272501428921,"p95_ms":1162.8374408794723,"missing_latency":0,"cost_basis":"single-call mean, three calls per item for cloud; local hardware/energy unknown"}],"paired":[{"cohort":"transfer","a":"decisions-zero","b":"responses-zero","difference":-0.3375,"low":-0.39620686121215803,"high":-0.28322554646836257,"prob_positive":0.0,"groups":22,"draws":4000,"method":"task-stratified family Bayesian bootstrap; empirical support only"},{"cohort":"transfer","a":"decisions-taught","b":"responses-taught","difference":-0.53125,"low":-0.5751369399448403,"high":-0.4703656730738328,"prob_positive":0.0,"groups":22,"draws":4000,"method":"task-stratified family Bayesian bootstrap; empirical support only"},{"cohort":"transfer","a":"decisions-taught","b":"decisions-zero","difference":-0.1875,"low":-0.23502767219260895,"high":-0.1408794371689231,"prob_positive":0.0,"groups":22,"draws":4000,"method":"task-stratified family Bayesian bootstrap; empirical support only"},{"cohort":"applicability","a":"decisions-zero","b":"responses-zero","difference":-0.3291666666666666,"low":-0.3784553071974812,"high":-0.27699995997758425,"prob_positive":0.0,"groups":48,"draws":4000,"method":"task-stratified family Bayesian bootstrap; empirical support only"}],"cascades":[{"router":"text-only","n":240,"correct":165,"oracle_correct":166,"failures":10,"posterior":{"mean":0.6859504132231405,"low":0.6262028090827387,"high":0.7427867850177939,"prob_above_bar":1.0375049156855433e-08,"jeffreys_mean":0.6867219917012448,"a":166,"b":76},"oracle_posterior":{"mean":0.6900826446280992,"low":0.63051402544846,"high":0.7466753677138176,"prob_above_bar":2.408142970416412e-08,"jeffreys_mean":0.6908713692946058,"a":167,"b":75},"cost_usd":0.078786556,"known_cost_usd":0.078786556,"unknown_costs":0,"median_ms":283.46829336661096,"p95_ms":395.54238136457195,"missing_latency":0,"sent_to_jev":120,"timing":"conditional sequential replay"},{"router":"logistic","n":240,"correct":164,"oracle_correct":165,"failures":10,"posterior":{"mean":0.6818181818181818,"low":0.6218969765468292,"high":0.7388928208509506,"prob_above_bar":4.387525023031258e-09,"jeffreys_mean":0.6825726141078838,"a":165,"b":77},"oracle_posterior":{"mean":0.6859504132231405,"low":0.6262028090827387,"high":0.7427867850177939,"prob_above_bar":1.0375049156855433e-08,"jeffreys_mean":0.6867219917012448,"a":166,"b":76},"cost_usd":0.08015000199999998,"known_cost_usd":0.08015000199999998,"unknown_costs":0,"median_ms":185.90883200522512,"p95_ms":402.4504609299983,"missing_latency":0,"sent_to_jev":97,"timing":"conditional sequential replay"},{"router":"asset-union","n":240,"correct":170,"oracle_correct":171,"failures":10,"posterior":{"mean":0.7066115702479339,"low":0.647813946590047,"high":0.7621746687107869,"prob_above_bar":5.790620598980322e-07,"jeffreys_mean":0.7074688796680498,"a":171,"b":71},"oracle_posterior":{"mean":0.7107438016528925,"low":0.6521530174435923,"high":0.7660354104949378,"prob_above_bar":1.222670908491291e-06,"jeffreys_mean":0.7116182572614108,"a":172,"b":70},"cost_usd":0.07469533999999997,"known_cost_usd":0.07469533999999997,"unknown_costs":0,"median_ms":296.21846142947675,"p95_ms":405.930574951595,"missing_latency":0,"sent_to_jev":132,"timing":"conditional sequential replay"},{"router":"qwen-06b","n":240,"correct":157,"oracle_correct":158,"failures":11,"posterior":{"mean":0.6528925619834711,"low":0.5919024035250231,"high":0.7114888747534849,"prob_above_bar":6.357167496034258e-12,"jeffreys_mean":0.6535269709543569,"a":158,"b":84},"oracle_posterior":{"mean":0.6570247933884298,"low":0.5961720178351235,"high":0.7154190433220925,"prob_above_bar":1.7076093454253663e-11,"jeffreys_mean":0.6576763485477178,"a":159,"b":83},"cost_usd":0.09482120000000001,"known_cost_usd":0.09482120000000001,"unknown_costs":0,"median_ms":626.4846669452041,"p95_ms":771.9472485952533,"missing_latency":0,"sent_to_jev":0,"timing":"conditional sequential replay"},{"router":"qwen-4b","n":240,"correct":163,"oracle_correct":164,"failures":10,"posterior":{"mean":0.6776859504132231,"low":0.6175964711396726,"high":0.7349935318090581,"prob_above_bar":1.821443251392282e-09,"jeffreys_mean":0.6784232365145229,"a":164,"b":78},"oracle_posterior":{"mean":0.6818181818181818,"low":0.6218969765468292,"high":0.7388928208509506,"prob_above_bar":4.387525023031258e-09,"jeffreys_mean":0.6825726141078838,"a":165,"b":77},"cost_usd":0.07806576200000001,"known_cost_usd":0.07806576200000001,"unknown_costs":0,"median_ms":1504.2816250290198,"p95_ms":2125.7016770493506,"missing_latency":0,"sent_to_jev":116,"timing":"conditional sequential replay"},{"router":"luna-router","n":240,"correct":170,"oracle_correct":171,"failures":9,"posterior":{"mean":0.7066115702479339,"low":0.647813946590047,"high":0.7621746687107869,"prob_above_bar":5.790620598980322e-07,"jeffreys_mean":0.7074688796680498,"a":171,"b":71},"oracle_posterior":{"mean":0.7107438016528925,"low":0.6521530174435923,"high":0.7660354104949378,"prob_above_bar":1.222670908491291e-06,"jeffreys_mean":0.7116182572614108,"a":172,"b":70},"cost_usd":0.18726682399999992,"known_cost_usd":0.18726682399999992,"unknown_costs":0,"median_ms":3176.4815,"p95_ms":4972.170407631686,"missing_latency":0,"sent_to_jev":142,"timing":"conditional sequential replay"},{"router":"jev-confidence","n":240,"correct":177,"oracle_correct":178,"failures":8,"posterior":{"mean":0.7355371900826446,"low":0.6783114955560496,"high":0.7890758793259661,"prob_above_bar":7.1969089956017e-05,"jeffreys_mean":0.7365145228215768,"a":178,"b":64},"oracle_posterior":{"mean":0.7396694214876033,"low":0.6826926979089023,"high":0.7928945149611109,"prob_above_bar":0.00013243524669737656,"jeffreys_mean":0.7406639004149378,"a":179,"b":63},"cost_usd":0.09476327199999998,"known_cost_usd":0.09476327199999998,"unknown_costs":0,"median_ms":374.2335,"p95_ms":619.3184291208296,"missing_latency":0,"sent_to_jev":153,"timing":"conditional sequential replay"},{"router":"always-jev","n":240,"correct":180,"oracle_correct":181,"failures":0,"posterior":{"mean":0.7479338842975206,"low":0.6914743294014548,"high":0.8005125728538232,"prob_above_bar":0.0004220985394667399,"jeffreys_mean":0.7489626556016598,"a":181,"b":61},"oracle_posterior":{"mean":0.7520661157024794,"low":0.6958749649973335,"high":0.8043117891311331,"prob_above_bar":0.0007309007080465509,"jeffreys_mean":0.7531120331950207,"a":182,"b":60},"cost_usd":0.043413972000000016,"known_cost_usd":0.043413972000000016,"unknown_costs":0,"median_ms":349.9155,"p95_ms":445.89439999999996,"missing_latency":0,"sent_to_jev":240,"timing":"conditional sequential replay"},{"router":"always-luna","n":240,"correct":157,"oracle_correct":158,"failures":11,"posterior":{"mean":0.6528925619834711,"low":0.5919024035250231,"high":0.7114888747534849,"prob_above_bar":6.357167496034258e-12,"jeffreys_mean":0.6535269709543569,"a":158,"b":84},"oracle_posterior":{"mean":0.6570247933884298,"low":0.5961720178351235,"high":0.7154190433220925,"prob_above_bar":1.7076093454253663e-11,"jeffreys_mean":0.6576763485477178,"a":159,"b":83},"cost_usd":0.09482120000000001,"known_cost_usd":0.09482120000000001,"unknown_costs":0,"median_ms":155.86422896012664,"p95_ms":216.34925599209961,"missing_latency":0,"sent_to_jev":0,"timing":"conditional sequential replay"}],"failures":[{"cohort":"transfer","case_id":"project_selection-test-20260920-0035","variant":"decisions-zero","status":"invalid","reason":"Invalid choice"},{"cohort":"transfer","case_id":"project_selection-dev-20260920-0007","variant":"decisions-zero","status":"invalid","reason":"Invalid choice"},{"cohort":"transfer","case_id":"project_selection-test-20260920-0035","variant":"decisions-taught","status":"invalid","reason":"Invalid choice"},{"cohort":"applicability","case_id":"fresh-project_selection-test-202609201-0047","variant":"decisions-zero","status":"invalid","reason":"Invalid choice"},{"cohort":"applicability","case_id":"fresh-project_selection-test-202609201-0051","variant":"decisions-zero","status":"invalid","reason":"Invalid choice"},{"cohort":"applicability","case_id":"fresh-project_selection-test-202609201-0046","variant":"decisions-zero","status":"invalid","reason":"Invalid choice"},{"cohort":"applicability","case_id":"fresh-project_selection-calibration-202609201-0021","variant":"decisions-zero","status":"invalid","reason":"Invalid choice"},{"cohort":"applicability","case_id":"fresh-project_selection-test-202609201-0057","variant":"decisions-zero","status":"invalid","reason":"Invalid choice"},{"cohort":"transfer","case_id":"project_selection-dev-20260920-0008","variant":"decisions-taught","status":"invalid","reason":"Invalid choice"},{"cohort":"applicability","case_id":"fresh-project_selection-calibration-202609201-0007","variant":"decisions-zero","status":"invalid","reason":"Invalid choice"},{"cohort":"transfer","case_id":"project_selection-test-20260920-0008","variant":"decisions-zero","status":"invalid","reason":"Invalid choice"},{"cohort":"applicability","case_id":"fresh-stateful_config-test-202609201-0036","variant":"decisions-zero","status":"invalid","reason":"Invalid choice"},{"cohort":"transfer","case_id":"project_selection-test-20260920-0024","variant":"decisions-zero","status":"invalid","reason":"Invalid choice"},{"cohort":"transfer","case_id":"project_selection-test-20260920-0012","variant":"decisions-zero","status":"invalid","reason":"Invalid choice"},{"cohort":"applicability","case_id":"fresh-project_selection-test-202609201-0040","variant":"decisions-zero","status":"invalid","reason":"Invalid choice"},{"cohort":"transfer","case_id":"project_selection-test-20260920-0028","variant":"decisions-zero","status":"invalid","reason":"Invalid choice"},{"cohort":"applicability","case_id":"fresh-project_selection-test-202609201-0043","variant":"decisions-zero","status":"invalid","reason":"Invalid choice"},{"cohort":"transfer","case_id":"project_selection-test-20260920-0007","variant":"decisions-zero","status":"invalid","reason":"Invalid choice"},{"cohort":"applicability","case_id":"fresh-project_selection-test-202609201-0038","variant":"decisions-zero","status":"invalid","reason":"Invalid choice"},{"cohort":"applicability","case_id":"fresh-project_selection-calibration-202609201-0024","variant":"decisions-zero","status":"invalid","reason":"Invalid choice"},{"cohort":"transfer","case_id":"project_selection-test-20260920-0007","variant":"decisions-taught","status":"invalid","reason":"Invalid choice"},{"cohort":"applicability","case_id":"fresh-project_selection-test-202609201-0023","variant":"decisions-zero","status":"invalid","reason":"Invalid choice"},{"cohort":"applicability","case_id":"fresh-project_selection-test-202609201-0042","variant":"decisions-zero","status":"invalid","reason":"Invalid choice"},{"cohort":"transfer","case_id":"project_selection-dev-20260920-0007","variant":"decisions-taught","status":"invalid","reason":"Invalid choice"},{"cohort":"transfer","case_id":"project_selection-test-20260920-0021","variant":"decisions-zero","status":"invalid","reason":"Invalid choice"},{"cohort":"applicability","case_id":"fresh-project_selection-calibration-202609201-0022","variant":"decisions-zero","status":"invalid","reason":"Invalid choice"},{"cohort":"transfer","case_id":"project_selection-test-20260920-0036","variant":"decisions-taught","status":"invalid","reason":"Invalid choice"},{"cohort":"applicability","case_id":"fresh-project_selection-test-202609201-0044","variant":"decisions-zero","status":"invalid","reason":"Invalid choice"},{"cohort":"transfer","case_id":"project_selection-test-20260920-0036","variant":"decisions-zero","status":"invalid","reason":"Invalid choice"},{"cohort":"transfer","case_id":"project_selection-test-20260920-0020","variant":"decisions-zero","status":"invalid","reason":"Invalid choice"}],"freeze_sha256":"30097f107e722a0bc0d6bb0900a075e1fa40a29e127851ae33f66f8b3dd6da6a"}
