The Proof Ledger

A permanent record of what the engine said at the time, which evidence it used, what could invalidate the thesis and what happened afterwards.

Read validation rules View thesis radar

Ledger protocolActiveFreeze-before-outcome enforced

Recorded outcomes0Populates from completed live windows

Model version32.0.0-attention-priorityVersion attached to every thesis

Predictive calibrationAwaiting sampleNo hit rate claimed prematurely

Point-in-time records

Every completed thesis remains auditable here.

The ledger rules are implemented. Outcome statistics appear only after qualified theses and their observation windows are actually completed.

Awaiting validated first record

✓

Ledger structure prepared · live worker not connected

No thesis enters the published ledger until its worker output, evidence checks and freeze are validated. No historical winners are fabricated.

01Thesis freeze gateEvidence, score, model version and invalidation are sealed.READY

02Observation windowThe declared horizon starts only after the freeze.GOVERNED

03Outcome measurementRelative return, MFE and MAE use consistent definitions.NOT LIVE

04Permanent recordSuccesses, failures and later corrections remain visible.IMMUTABLE

Why there is no sample winner: displaying a fabricated or backfilled result would undermine the evidence standard this ledger is designed to protect.

Freeze before outcome

Inputs, evidence, model version, counterarguments and invalidation are sealed before the observation window.

Measure consistently

Relative outcomes, MFE, MAE, regime context and horizon use governed definitions.

Keep failures visible

Wrong theses remain searchable. Corrections and methodology changes are appended, never rewritten.

Learn from the record

The ledger becomes the calibration dataset for better evidence weighting and false-positive control.