{"abstract":"Arrival order wins capacity over stronger routes.","category":"Mixture of experts dispatch","checks":6,"contract":"Choose up to capacity rows per expert by score descending then token ID; return selected rows in original order.","contract_signature":"routes,capacity","evaluation_group":"moe-score-capacity","failed_approach":"Global score sorting admits the right rows but disrupts token order.","family":"w3-mixture-of-experts-dispatch-score-capacity","id":"FA-100601","implementations":{"attempt":{"sha256":"590ebd86c3ff6980b92194f6122796861aa53d342fe689d4c1da8b451415f2cb","source":"\"\"\"Failure Map reference implementation. Python standard library only.\"\"\"\nimport json\n\nN = 1\nobservations = []\ndef solve(routes,capacity):\n    return sorted(routes,key=lambda r:(-r[2],r[0]))[:capacity]\ndef check(label, actual, expected):\n    observations.append({\"check\": label, \"actual\": actual, \"expected\": expected, \"passed\": actual == expected})\ne='e'+str(N); rows=[[0,e,1],[1,e,9],[2,e,5]]\ncheck('best two in arrival order',solve(rows,2),[rows[1],rows[2]])\ncheck('best one',solve(rows,1),[rows[1]])\ncheck('all fit',solve(rows,5),rows)\ncheck('zero',solve(rows,0),[])\ncheck('empty',solve([],N),[])\ncheck('per expert quota',solve([[0,e,1],[1,'x',2],[2,e,3]],1),[[1,'x',2],[2,e,3]])\nprint(json.dumps({\"observations\": observations, \"passed\": all(x[\"passed\"] for x in observations)}, ensure_ascii=False))\nraise SystemExit(0 if all(x[\"passed\"] for x in observations) else 1)\n"},"broken":{"sha256":"94b3c5fe6085da96117dbd77dd1b9c160a2c81caed418703056530a53abb8e84","source":"\"\"\"Failure Map reference implementation. Python standard library only.\"\"\"\nimport json\n\nN = 1\nobservations = []\ndef solve(routes,capacity):\n    return routes[:capacity]\ndef check(label, actual, expected):\n    observations.append({\"check\": label, \"actual\": actual, \"expected\": expected, \"passed\": actual == expected})\ne='e'+str(N); rows=[[0,e,1],[1,e,9],[2,e,5]]\ncheck('best two in arrival order',solve(rows,2),[rows[1],rows[2]])\ncheck('best one',solve(rows,1),[rows[1]])\ncheck('all fit',solve(rows,5),rows)\ncheck('zero',solve(rows,0),[])\ncheck('empty',solve([],N),[])\ncheck('per expert quota',solve([[0,e,1],[1,'x',2],[2,e,3]],1),[[1,'x',2],[2,e,3]])\nprint(json.dumps({\"observations\": observations, \"passed\": all(x[\"passed\"] for x in observations)}, ensure_ascii=False))\nraise SystemExit(0 if all(x[\"passed\"] for x in observations) else 1)\n"}},"limitations":"Inputs are finite toy scores and routes; the contract does not model a production model, accelerator, or learned routing quality. This reproducer isolates one failure mechanism. Results cover the supplied fixtures. Variants within a family share a test contract and should remain grouped when constructing evaluation splits. Related mechanisms with a shared evaluation_group must also remain together; these controlled models are not independent production incidents.","method":"Deterministic executable model with adversarial boundary fixtures.","provenance":{"created_by":"Failure Map","dependencies":"Python standard library","family":"w3-mixture-of-experts-dispatch-score-capacity","generated_at":"2026-09-30T16:48:37.615583+00:00","license":"CC0-1.0","python":"3.12.14","seed":1,"split":"open-access"},"relevance":"A deterministic toy router makes this dispatch decision independently inspectable.","root_cause":"Capacity admission ignores the contract’s score priority.","sha256":"0fa74b37942024df63ac904b058810b5db3cf41a632d1bd28c858b29cc3bd9eb","title":"Arrival order wins capacity over stronger routes · case 01","variant":1,"variant_policy":"The fixture code references N, supplied as 1 through 5. Fixed boundary checks may repeat. These records share one mechanism and must remain together in evaluation splits.","verified":true,"visibility":"public","verification":{"attempt":{"elapsed_ms":29.955,"exit_code":1,"observations":[{"actual":[[1,"e1",9],[2,"e1",5]],"check":"best two in arrival order","expected":[[1,"e1",9],[2,"e1",5]],"passed":true},{"actual":[[1,"e1",9]],"check":"best one","expected":[[1,"e1",9]],"passed":true},{"actual":[[1,"e1",9],[2,"e1",5],[0,"e1",1]],"check":"all fit","expected":[[0,"e1",1],[1,"e1",9],[2,"e1",5]],"passed":false},{"actual":[],"check":"zero","expected":[],"passed":true},{"actual":[],"check":"empty","expected":[],"passed":true},{"actual":[[2,"e1",3]],"check":"per expert quota","expected":[[1,"x",2],[2,"e1",3]],"passed":false}],"passed":false,"stderr":"","stdout":"{\"observations\": [{\"check\": \"best two in arrival order\", \"actual\": [[1, \"e1\", 9], [2, \"e1\", 5]], \"expected\": [[1, \"e1\", 9], [2, \"e1\", 5]], \"passed\": true}, {\"check\": \"best one\", \"actual\": [[1, \"e1\", 9]], \"expected\": [[1, \"e1\", 9]], \"passed\": true}, {\"check\": \"all fit\", \"actual\": [[1, \"e1\", 9], [2, \"e1\", 5], [0, \"e1\", 1]], \"expected\": [[0, \"e1\", 1], [1, \"e1\", 9], [2, \"e1\", 5]], \"passed\": false}, {\"check\": \"zero\", \"actual\": [], \"expected\": [], \"passed\": true}, {\"check\": \"empty\", \"actual\": [], \"expected\": [], \"passed\": true}, {\"check\": \"per expert quota\", \"actual\": [[2, \"e1\", 3]], \"expected\": [[1, \"x\", 2], [2, \"e1\", 3]], \"passed\": false}], \"passed\": false}\n"},"broken":{"elapsed_ms":27.081,"exit_code":1,"observations":[{"actual":[[0,"e1",1],[1,"e1",9]],"check":"best two in arrival order","expected":[[1,"e1",9],[2,"e1",5]],"passed":false},{"actual":[[0,"e1",1]],"check":"best one","expected":[[1,"e1",9]],"passed":false},{"actual":[[0,"e1",1],[1,"e1",9],[2,"e1",5]],"check":"all fit","expected":[[0,"e1",1],[1,"e1",9],[2,"e1",5]],"passed":true},{"actual":[],"check":"zero","expected":[],"passed":true},{"actual":[],"check":"empty","expected":[],"passed":true},{"actual":[[0,"e1",1]],"check":"per expert quota","expected":[[1,"x",2],[2,"e1",3]],"passed":false}],"passed":false,"stderr":"","stdout":"{\"observations\": [{\"check\": \"best two in arrival order\", \"actual\": [[0, \"e1\", 1], [1, \"e1\", 9]], \"expected\": [[1, \"e1\", 9], [2, \"e1\", 5]], \"passed\": false}, {\"check\": \"best one\", \"actual\": [[0, \"e1\", 1]], \"expected\": [[1, \"e1\", 9]], \"passed\": false}, {\"check\": \"all fit\", \"actual\": [[0, \"e1\", 1], [1, \"e1\", 9], [2, \"e1\", 5]], \"expected\": [[0, \"e1\", 1], [1, \"e1\", 9], [2, \"e1\", 5]], \"passed\": true}, {\"check\": \"zero\", \"actual\": [], \"expected\": [], \"passed\": true}, {\"check\": \"empty\", \"actual\": [], \"expected\": [], \"passed\": true}, {\"check\": \"per expert quota\", \"actual\": [[0, \"e1\", 1]], \"expected\": [[1, \"x\", 2], [2, \"e1\", 3]], \"passed\": false}], \"passed\": false}\n"}},"member_only":{"stages":["fixed"],"fields":["implementations.fixed","verification.fixed","harness","repair"],"note":"The verified repair, its recorded checks, the repair description, and the scoring harness are available to members."}}