100%

03 — Surrogate Calibration Instrument

Goal

Make the refuter EARN its veto: track refuter-prediction vs measured-outcome agreement and down-weight an uncalibrated refuter. The 2-of-2 prediction failures say the surrogate is unvalidated.

Implementation Sketch

  • calibration_record appended per measured bet: {refuter_majority_verdict, measured_direction, agreed: bool}.
  • veto weight = f(rolling agreement); a refuter set with low agreement cannot unilaterally withdraw a bet — it can only de-prioritize.
  • Append-only; calibration history is read via the §13 read API.

Spec References

  • scripts/ledger_evolution_runtime/verification.py (refuter records)
  • learning-ledger.md §11.2 (verification_record), §3 teeth
  • Evidence: 2-of-2 measured bets contradicted the a-priori refuter prediction

Work Items

  • Implement calibration tracking: for every empirically-measured bet, record refuter-prediction vs measured-outcome agreement into the calibration_record (the 2-of-2-predictions-wrong evidence motivates this).
  • Feed calibration back into verification: an under-calibrated refuter loses veto weight (calibration gates the gate). Plan-agnostic; the calibration history lives in the plan’s own evolutions[], never in generic tooling.
  • Cover with scripts/ledger_evolution_runtime/tests/test_calibration.py (agreement recorded + veto-weight feedback); update learning-ledger.md §11.

Fresh intel (regenerated)

{ “schema_version”: 1, “target”: { “kind”: “plan-section”, “ref”: “evolution-empirical-redesign/s-e40e5a1c” }, “generated_at”: “2026-06-19T22:48:33.001094+00:00”, “graph_state”: { “head_sha”: “e4b7c39a”, “last_code_import_at”: “2026-06-19T22:46:52.846Z”, “embedding_stale”: false, “insights_stale”: false, “cpg_stale”: false }, “surfaces”: {}, “summary”: “intel-package[plan-section:evolution-empirical-redesign/s-e40e5a1c] agent-authored dossier”, “degraded”: false, “dossier”: { “objective_symbols”: [], “tiers”: {}, “agent_authored”: true, “difficulty”: { “class”: “routine”, “research_online”: false, “research_mode”: “auto”, “signals”: [], “plan_dir”: “/home/eric/projects/ori_lang/plans/evolution-empirical-redesign” } } }

(full dossier: surrogate-calibration—s-e40e5a1c.intel.json)