{ "schema_version": "alliance-calibration-comparison.v1", "generated_at": "2026-08-07 07:57:08.499299+00:00", "baseline": { "run": { "run_id": "g1-candidate-prompt-1.2.0-2026-08-06", "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.2.0", "generated_at": "2026-08-06 12:48:10.139793+00:00" }, "expected_count": 9, "ready_count": 9, "missing_or_failed_count": 0, "error_count": 0, "recovered_after_retry_count": 2, "direction_accuracy": 1.0, "evidence_precision": 0.7272727272727273, "evidence_recall": 0.8888888888888888, "dimension_means": { "bond": 0.3, "goal": 0.175, "task": 0.43333333333333335 }, "rows": [ { "case_id": "oas-g0-001", "perspective": "independent_observer", "dimension": "goal", "expected_direction": "low", "status": "ready", "error_code": null, "direction_match": true, "value": 0.15, "confidence": 0.85, "evidence_precision": 1.0, "evidence_recall": 1.0 }, { "case_id": "oas-g0-002", "perspective": "independent_observer", "dimension": "task", "expected_direction": "low", "status": "ready", "error_code": null, "direction_match": true, "value": 0.1, "confidence": 0.6, "evidence_precision": 1.0, "evidence_recall": 1.0 }, { "case_id": "oas-g0-003", "perspective": "independent_observer", "dimension": "bond", "expected_direction": "drop", "status": "ready", "error_code": null, "direction_match": true, "value": 0.3, "confidence": 0.2, "evidence_precision": 0.5, "evidence_recall": 0.5 }, { "case_id": "oas-g0-005", "perspective": "client_agent_report", "dimension": "bond", "expected_direction": "low", "status": "ready", "error_code": null, "direction_match": true, "value": 0.1, "confidence": 0.75, "evidence_precision": 0.5, "evidence_recall": 1.0 }, { "case_id": "oas-g0-006", "perspective": "client_agent_report", "dimension": "task", "expected_direction": "rise", "status": "ready", "error_code": null, "direction_match": true, "value": 0.8, "confidence": 0.75, "evidence_precision": 1.0, "evidence_recall": 1.0 }, { "case_id": "oas-g0-007", "perspective": "client_agent_report", "dimension": "bond", "expected_direction": "low", "status": "ready", "error_code": null, "direction_match": true, "value": 0.1, "confidence": 0.75, "evidence_precision": 0.3333333333333333, "evidence_recall": 1.0 }, { "case_id": "oas-g0-008", "perspective": "client_agent_report", "dimension": "bond", "expected_direction": "high", "status": "ready", "error_code": null, "direction_match": true, "value": 0.7, "confidence": 0.45, "evidence_precision": 1.0, "evidence_recall": 1.0 }, { "case_id": "oas-g0-008", "perspective": "independent_observer", "dimension": "goal", "expected_direction": "low", "status": "ready", "error_code": null, "direction_match": true, "value": 0.2, "confidence": 0.45, "evidence_precision": 1.0, "evidence_recall": 1.0 }, { "case_id": "oas-g0-008", "perspective": "independent_observer", "dimension": "task", "expected_direction": "low", "status": "ready", "error_code": null, "direction_match": true, "value": 0.4, "confidence": 0.4, "evidence_precision": 0.3333333333333333, "evidence_recall": 0.5 } ] }, "candidate": { "run": { "run_id": "g1-candidate-prompt-1.2.0-readskewfix-2026-08-07", "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.2.0", "generated_at": "2026-08-07 07:56:52.784424+00:00" }, "expected_count": 9, "ready_count": 9, "missing_or_failed_count": 0, "error_count": 0, "recovered_after_retry_count": 0, "direction_accuracy": 1.0, "evidence_precision": 0.7391304347826086, "evidence_recall": 0.9444444444444444, "dimension_means": { "bond": 0.32499999999999996, "goal": 0.15000000000000002, "task": 0.4166666666666667 }, "rows": [ { "case_id": "oas-g0-001", "perspective": "independent_observer", "dimension": "goal", "expected_direction": "low", "status": "ready", "error_code": null, "direction_match": true, "value": 0.1, "confidence": 0.75, "evidence_precision": 1.0, "evidence_recall": 1.0 }, { "case_id": "oas-g0-002", "perspective": "independent_observer", "dimension": "task", "expected_direction": "low", "status": "ready", "error_code": null, "direction_match": true, "value": 0.15, "confidence": 0.6, "evidence_precision": 1.0, "evidence_recall": 1.0 }, { "case_id": "oas-g0-003", "perspective": "independent_observer", "dimension": "bond", "expected_direction": "drop", "status": "ready", "error_code": null, "direction_match": true, "value": 0.4, "confidence": 0.3, "evidence_precision": 1.0, "evidence_recall": 1.0 }, { "case_id": "oas-g0-005", "perspective": "client_agent_report", "dimension": "bond", "expected_direction": "low", "status": "ready", "error_code": null, "direction_match": true, "value": 0.1, "confidence": 0.65, "evidence_precision": 0.3333333333333333, "evidence_recall": 1.0 }, { "case_id": "oas-g0-006", "perspective": "client_agent_report", "dimension": "task", "expected_direction": "rise", "status": "ready", "error_code": null, "direction_match": true, "value": 0.8, "confidence": 0.8, "evidence_precision": 1.0, "evidence_recall": 1.0 }, { "case_id": "oas-g0-007", "perspective": "client_agent_report", "dimension": "bond", "expected_direction": "low", "status": "ready", "error_code": null, "direction_match": true, "value": 0.1, "confidence": 0.65, "evidence_precision": 0.3333333333333333, "evidence_recall": 1.0 }, { "case_id": "oas-g0-008", "perspective": "client_agent_report", "dimension": "bond", "expected_direction": "high", "status": "ready", "error_code": null, "direction_match": true, "value": 0.7, "confidence": 0.55, "evidence_precision": 1.0, "evidence_recall": 1.0 }, { "case_id": "oas-g0-008", "perspective": "independent_observer", "dimension": "goal", "expected_direction": "low", "status": "ready", "error_code": null, "direction_match": true, "value": 0.2, "confidence": 0.45, "evidence_precision": 1.0, "evidence_recall": 1.0 }, { "case_id": "oas-g0-008", "perspective": "independent_observer", "dimension": "task", "expected_direction": "low", "status": "ready", "error_code": null, "direction_match": true, "value": 0.3, "confidence": 0.4, "evidence_precision": 0.3333333333333333, "evidence_recall": 0.5 } ] }, "drift": { "comparable_predictions": 24, "mean_absolute_score_delta": 0.05208333333333334, "dimension_mean_delta": { "bond": 0.006249999999999999, "goal": -0.012499999999999987, "task": -0.01250000000000001 }, "direction_accuracy_delta": 0.0, "evidence_recall_delta": 0.05555555555555558, "evidence_precision_delta": 0.011857707509881354 } }