Honesty doctrine. Every candidate anomaly is an artifact until proven otherwise;
in-sample results are never findings; past statistical regularity does not imply future returns.
This is research on statistical properties of market data — not investment advice, not a trading system.