{"abstract":"Momentum dampening scales the carried buffer.","category":"Optimizer update state","checks":6,"contract":"v_next=mu*v+(1-dampening)*g; return [p-lr*v_next,v_next].","contract_signature":"p, g, v, lr, mu, damp","evaluation_group":"optimizer-toy-momentum-dampening","failed_approach":"Removing dampening from the entire recurrence preserves the carried buffer but leaves the new gradient undampened.","family":"w3-optimizer-update-state-momentum-dampening","id":"FA-98951","implementations":{"attempt":{"sha256":"32280c960ca88e33e95cca727c80c36c124741ba6663aca984118e4b051c3bef","source":"\"\"\"Failure Map reference implementation. Python standard library only.\"\"\"\nimport json\n\nN = 1\nobservations = []\ndef solve(p, g, v, lr, mu, damp):\n    nv=mu*v+g\n    return [p-lr*nv,nv]\ndef check(label, actual, expected):\n    observations.append({\"check\": label, \"actual\": actual, \"expected\": expected, \"passed\": actual == expected})\ncheck('fresh buffer',solve(8,4,0,1,0.5,0.5),[6,2])\ncheck('retained velocity',solve(8,4,6,1,0.5,0.5),[3,5])\ncheck('full dampening',solve(3,7,2,1,0.8,1),[1.4,1.6])\ncheck('zero momentum',solve(4,-2,9,2,0,0.25),[7,-1.5])\ncheck('zero gradient',solve(5,0,4,1,0.5,0.25),[3,2])\ncheck('zero learning rate',solve(2,3,4,0,0.5,0.5),[2,3.5])\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":"79670437008ff5fae2fb5827cc4acffd3b1cc041a009ba1bc00810b63d42f753","source":"\"\"\"Failure Map reference implementation. Python standard library only.\"\"\"\nimport json\n\nN = 1\nobservations = []\ndef solve(p, g, v, lr, mu, damp):\n    nv=(1-damp)*(mu*v+g)\n    return [p-lr*nv,nv]\ndef check(label, actual, expected):\n    observations.append({\"check\": label, \"actual\": actual, \"expected\": expected, \"passed\": actual == expected})\ncheck('fresh buffer',solve(8,4,0,1,0.5,0.5),[6,2])\ncheck('retained velocity',solve(8,4,6,1,0.5,0.5),[3,5])\ncheck('full dampening',solve(3,7,2,1,0.8,1),[1.4,1.6])\ncheck('zero momentum',solve(4,-2,9,2,0,0.25),[7,-1.5])\ncheck('zero gradient',solve(5,0,4,1,0.5,0.25),[3,2])\ncheck('zero learning rate',solve(2,3,4,0,0.5,0.5),[2,3.5])\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":"These are deterministic scalar or short-vector toy contracts, not claims about production optimizer APIs or training 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-optimizer-update-state-momentum-dampening","generated_at":"2026-09-30T16:48:18.534176+00:00","license":"CC0-1.0","python":"3.12.14","seed":1,"split":"open-access"},"relevance":"Optimizer implementations must associate each parameter with its own mathematically specified update state.","root_cause":"Dampening is applied to the whole momentum recurrence.","sha256":"d9f3100b4f6b7b67ea98cc79813c55b27f36f8dc15a948aa5fbe51e2f7b8e036","title":"Momentum dampening scales the carried buffer · case 01","variant":1,"variant_policy":"The same explicitly authored fixtures repeat across all five numbered records; runtime N is unused. These records share one mechanism and must remain together in evaluation splits.","verified":true,"visibility":"public","verification":{"attempt":{"elapsed_ms":29.253,"exit_code":1,"observations":[{"actual":[4.0,4.0],"check":"fresh buffer","expected":[6,2],"passed":false},{"actual":[1.0,7.0],"check":"retained velocity","expected":[3,5],"passed":false},{"actual":[-5.6,8.6],"check":"full dampening","expected":[1.4,1.6],"passed":false},{"actual":[8,-2],"check":"zero momentum","expected":[7,-1.5],"passed":false},{"actual":[3.0,2.0],"check":"zero gradient","expected":[3,2],"passed":true},{"actual":[2.0,5.0],"check":"zero learning rate","expected":[2,3.5],"passed":false}],"passed":false,"stderr":"","stdout":"{\"observations\": [{\"check\": \"fresh buffer\", \"actual\": [4.0, 4.0], \"expected\": [6, 2], \"passed\": false}, {\"check\": \"retained velocity\", \"actual\": [1.0, 7.0], \"expected\": [3, 5], \"passed\": false}, {\"check\": \"full dampening\", \"actual\": [-5.6, 8.6], \"expected\": [1.4, 1.6], \"passed\": false}, {\"check\": \"zero momentum\", \"actual\": [8, -2], \"expected\": [7, -1.5], \"passed\": false}, {\"check\": \"zero gradient\", \"actual\": [3.0, 2.0], \"expected\": [3, 2], \"passed\": true}, {\"check\": \"zero learning rate\", \"actual\": [2.0, 5.0], \"expected\": [2, 3.5], \"passed\": false}], \"passed\": false}\n"},"broken":{"elapsed_ms":27.013,"exit_code":1,"observations":[{"actual":[6.0,2.0],"check":"fresh buffer","expected":[6,2],"passed":true},{"actual":[4.5,3.5],"check":"retained velocity","expected":[3,5],"passed":false},{"actual":[3.0,0.0],"check":"full dampening","expected":[1.4,1.6],"passed":false},{"actual":[7.0,-1.5],"check":"zero momentum","expected":[7,-1.5],"passed":true},{"actual":[3.5,1.5],"check":"zero gradient","expected":[3,2],"passed":false},{"actual":[2.0,2.5],"check":"zero learning rate","expected":[2,3.5],"passed":false}],"passed":false,"stderr":"","stdout":"{\"observations\": [{\"check\": \"fresh buffer\", \"actual\": [6.0, 2.0], \"expected\": [6, 2], \"passed\": true}, {\"check\": \"retained velocity\", \"actual\": [4.5, 3.5], \"expected\": [3, 5], \"passed\": false}, {\"check\": \"full dampening\", \"actual\": [3.0, 0.0], \"expected\": [1.4, 1.6], \"passed\": false}, {\"check\": \"zero momentum\", \"actual\": [7.0, -1.5], \"expected\": [7, -1.5], \"passed\": true}, {\"check\": \"zero gradient\", \"actual\": [3.5, 1.5], \"expected\": [3, 2], \"passed\": false}, {\"check\": \"zero learning rate\", \"actual\": [2.0, 2.5], \"expected\": [2, 3.5], \"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."}}