116 active prediction markets are priced based on stale debate outcomes. Two major new signals have emerged: (1) 129 hypotheses now have lifecycle='validated' (avg score 0.760); (2) 716 hypotheses now have evidence_validation_score > 0 from the mechanistic claim verifier (merged 2026-04-28).
NOTE: The prediction market resolution engine (task 314cc632) handles RESOLVED markets. This task handles ACTIVE markets — updating their prices, not resolving them.
WHAT TO DO:
1. SELECT pm.id, pm.current_price, h.composite_score, h.lifecycle, h.evidence_validation_score, pm.updated_at FROM prediction_markets pm JOIN hypotheses h ON pm.hypothesis_id = h.id WHERE pm.status = 'active'
2. For each market, compute target_price:
- lifecycle='validated': add +0.1 to price prior
- evidence_validation_score delta: positive evidence → price up
- composite_score vs market price divergence > 0.15: recalibrate
3. Skip if |target_price - current_price| < 0.05 (no meaningful movement)
4. UPDATE prediction_markets SET current_price=target_price, updated_at=NOW()
5. UPDATE hypotheses SET market_price=target_price WHERE id=...
6. INSERT INTO price_history (market_id, price, event_type='validation_signal_recalibration', metadata_json)
7. Cap: no market moves > 0.3 in single recalibration
8. Log: market_id, old_price, new_price, signal_type for each updated market
TARGET: >= 30 markets recalibrated with measurable price movement.
DO NOT: resolve markets; create new markets; modify markets with status != 'active'; write to scidex.db.
Spec: docs/planning/specs/exchange_market_recalibration_validation_signals_spec.md (merged after task creation)