SIMULATION · EXPERIMENT · Evidence

· Co-GM evidence, including information private during play

broken-measure-blind-01

BLIND_REVIEWS_COMPLETED — AI_ONLY_PLAYTEST

Completed turns: 0. Calls started: 4. Rejected attempts: 0.

Measured provider cost: unavailable for one or more calls. Incomplete calls: 0; reserved for unknown charges: $0.0000.

Role Calls Summed call seconds Input tokens Output tokens
blind 4 112.9 303638 23991

Simulation decisions are not human approvals. Fixture runs verify orchestration, not game quality. Model review does not prove narrative privacy or human enjoyment. Token totals are measured at completed-call boundaries; in-flight calls may exceed the stopping threshold. Inspect turn records for decisions, state changes, private reports, and delivery hashes.

← Return to archive