Forecasting Council — 20260723T095613198Z-order-20260716

Exploratory model pilot. Not preregistered, not confirmatory evidence. Historical questions are replayed after a selected-model, zero-evidence leakage audit; model outputs are objectively JSON-parsed and scored against frozen outcomes. The pre-existing primary endpoint remains corrupted aggregation under controlled registry drift; Brier score is reported as the registered utility metric with mandatory market-prior and independent-agent baselines.

Mode: model-pilot
Provider: llm.chutes.ai
Model: unsloth/Mistral-Nemo-Instruct-2407-TEE
Semantic backend: semahash-python-api
Sema version: 0.3.0
Canonicalization: v2
Order seed: 20260716

Condition	Trials	Drift trials	Detected	Corrupted aggregations	Correct exclusions	False exclusions
`baseline`	100	50	0/50	49/50	0/50	0/100
`addressed-voluntary`	100	50	47/50	0/50	0/50	0/100
`addressed-enforced`	100	50	48/50	0/50	48/50	0/100

Drift-scoped columns use drift-injected trials as the denominator; the rest use all trials in the condition. Every count is recomputed from trials.public.jsonl at build time.

Forecast utility and resource channels

Descriptive utility result: enforcement did not improve mean aggregate Brier in this run (0.2552 enforced versus 0.2435 baseline; lower is better). This is compatible with the semantic gate doing mechanism work without improving model forecasting performance.

Condition	Mean Brier aggregate	Mean Brier market prior	Mean Brier independent	Malformed / failed model outputs
`baseline`	0.2435	0.2506	0.2474	31
`addressed-voluntary`	0.2448	0.2506	0.2559	32
`addressed-enforced`	0.2552	0.2506	0.2594	42

Historical source: SimpleFunctions Settled Prediction Markets at revision a27e3e9307266481d51e087fffd5bf934410e01c, licensed CC-BY-4.0; attribution: SimpleFunctions (simplefunctions.dev).
Wire payload: 2,505,168 bytes.
Registry hydration/context: 1,180,320 bytes.
Tokenizer-derived coordination hydration: 596,820 context tokens using mistralai/Mistral-Nemo-Instruct-2407@04d8a90549d23fc6bd7f642064003592df51e9b3 (tokenizer SHA-256 e11c71726323d33da7b8d6f6f269f1988931c0a52b7122bcdd8c05042974e0db).
Provider model tokens: 1,299,880 input, 727,154 cached-input reads (a subset of input), 269,159 output, 0 separately reported reasoning, 1,569,039 total input + output + reasoning.
Provider retries/errors: 0/0; provider-reported cost: not reported.

Lower Brier is better. Model calls were independently sampled per condition, so small cross-condition Brier differences are exploratory and must be read alongside the clean controls. Cached-input reads are observational, not an additional token charge. Wire, hydration, and model-token channels are reported separately; a short reference is not treated as a context-token saving. Hydration context tokens are the sum of upstream-tokenizer tokens in the compact JSON serialization of the coordination object included in each model request. This excludes chat framing, question text, peer forecasts, and output. It is a context-channel measurement, not a claimed token or cost saving.