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/expC_reversion_scan/README.md

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

# expC_reversion_scan

Mean-reversion scan: variance ratios, half-life, AR tests net of the bounce null

Status: **completed 2026-08-12** — 127 cells, corrected triage: 14 multi-lag-reversion cells + daily 2008-15 reversal family -> expF. Key output: the bounce-null variance-consistency lesson (taxonomy T1).