anomaly-atlas
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.

Code / experiments/micro/expH_oos_stability/README.md

experiments/micro/expH_oos_stability/README.md 16 lines
---
project: anomaly-atlas
document: expH_oos_stability/README
author: Simon-Pierre Boucher
contact: contact@spboucher.ai
data_source: hfmarketdata.io
created: 2026-08-12
status: draft
---

# expH_oos_stability

Out-of-sample stability: once-touched holdout and sub-period evaluation of survivors

Status: **completed 2026-08-12** — validation split opened once: negative finding replicates (0 net survivors at kappa=1), CKX fails robustness, ES->SPY does not replicate, daily reversal was crisis-regime. First atlas entries issued.