# Observation validity under changing dynamics — invalidated study

Execution completed; scientific status: INVALIDATED. Legacy public_dynamics.hidden_parameter says the coefficient is the same in historical and target experiments; retained includes hidden coefficients, and changed says only initial geometry/clocks vary. These inherited statements conflict with new between-task continuity contracts. Although the new contracts are explicit in the system assignment, the delivered scientific metadata is internally inconsistent. Equal exposure does not remove this ambiguity, and the rule policy explicitly follows the new contract.

All 64 scheduled executions completed. Under registered scoring, the LLM produced 15/16 correct labels and 15/16 supported-correct classifications; the change-aware rule produced 16/16 with zero inference. The model used one uncertified old observation on a changed apparatus and gave the wrong answer. Blind caching produced 12/16 correct labels but only 10/16 supported-correct classifications. The intended formal contract admits a sufficient deterministic rule. However the comparison is scientifically invalidated by contradictory inherited public metadata, so this run cannot cleanly attribute the model error to validity management under coherent information. All observed scores and traces are preserved.

| Policy | Correct labels | Supported-correct | Unsupported | Calibrations | Model calls | Provider minutes |
|---|---:|---:|---:|---:|---:|---:|
| LLM_PERSISTENT | 15/16 | 15/16 | 1 | 14 | 85 | 72.02 |
| RULE_CHANGE_AWARE | 16/16 | 16/16 | 0 | 14 | 0 | 0.00 |
| RULE_ALWAYS_FRESH | 16/16 | 16/16 | 0 | 16 | 0 | 0.00 |
| RULE_BLIND_CACHE | 12/16 | 10/16 | 6 | 8 | 0 | 0.00 |

These are eight constructed apparatuses in two blocks. Changes were selected diagnostic cases; the study estimates no natural drift rate or precise population effect. Every policy used the same public tools and information. Private unchanged truth never certifies observation freshness. The unsafe blind cache is a diagnostic, not the competent rule baseline.

Including all attempts, the LLM used 288.081 provider seconds and 288.945 total task seconds per supported-correct classification. The change-aware rule used zero inference and 0.139 total task seconds per supported-correct classification, including guards. These are realized averages, not marginal causal costs. Loading is nested in provider time; complete timing and token records are retained.

Independent static audit rechecked all terminal citations, exact interval/target links, source payload hashes, observable applicability, source selection, calibration redundancy, decision caps, task context isolation, label scores, raw usage and provider-log token counts. No model or controller imports. Earlier seals and reservations verified unchanged. Separate provider-disabled reporting replay verified all 85 reconstructed requests, accepted response replay, scientific facts and observations after qualifying an overbroad changed-field denylist. This audit repair did not change executing source or inference. Exact fixtures and all rejected candidates were reproduced against the registered preparation-base history. The legacy continuity conflict is retained as scientific invalidation, not repaired in the run.

Full private source, fixtures, model conversations and journals are withheld. The public contracts, factual task statuses and hashes support descriptive checks, not full public replay. Prior memory-mechanism-v1 revision 2 and its qualifications remain intact.

Next proposal only: Test whether the observed stale-reuse failure persists under an internally consistent contract: preregister a new study after replacing legacy transfer claims with one scope-correct continuity description and auditing every emitted metadata field against it. Keep the competent change-aware rule and the same bounded design. Not launched.

Execution and publication are recorded separately. No next study or consciousness inference is authorized by these results.