SPB Git

spb/anomaly-atlas Public License

Systematic discovery & rigorous validation of statistical anomalies in open HF market data (hfmarketdata.io) — pre-registered, artifact-null-driven, fully reproducible. Live atlas: www.anomaly-atlas.io

Python 61.4% JavaScript 28.7% CSS 8.6% Shell 0.7% Makefile 0.5%

# project: anomaly-atlas document: expE_calendar_scan/analysis author: Simon-Pierre Boucher contact: contact@spboucher.ai data_source: hfmarketdata.io created: 2026-08-12 modified: 2026-08-12 status: reviewed

# Analysis — expE_calendar_scan

Run: results/expE_calendar_scan/20260812T065222Z/results.json (protocol and 8-test budget pre-specified; single declared instrument SPY; permuted-calendar null n=2000; manifest embedded). Level 0 throughout.

# 1. The calendar is empty (and that is the finding)

class observed (bp) perm 95% band p family-wise
mon −1.4 [−7.9, +7.9] 1.00
tue +4.0 [−7.9, +8.0] 0.99
fri −3.8 [−8.0, +8.2] 0.99
turn_of_month +4.7 [−8.2, +8.3] 0.97
pre_holiday +5.7 [−20.8, +21.0] 0.93
post_holiday +2.9 [−21.2, +20.7] 1.00

Every dot inside its band; zero FDR rejections; the family-wise max-stat (mini reality check) never gets close. Turn-of-month — the last calendar survivor as of Marquering et al. (2006) — fails the permuted-calendar bar in train (2000–2016) and decays inside it (+7.9 bp in 2000–2007 → +1.6 bp in 2008–2015). This is exactly the Sullivan–Timmermann–White prediction, reproduced on open data with an honest budget.

# 2. Controls behaved — the negative is trustworthy

  • Monday (negative control): −1.4 bp, p 0.73. Had the pipeline "found" the canonical corpse, the pipeline would have been declared broken (the pre-registered criterion). It did not.
  • H20 U-shape (positive control): strongly present. Volatility 6.6 bp (09:30) → 2.4 (midday) → 2.9 (15:30); EDGE spread declines monotonically 2.8 → 1.2 bp — the spread profile is an L, not a U (the close's liquidity is cheap; the open's is expensive). Both numbers now live in the taxonomy: any first-30-minutes return claim fights 2–3× the midday artifact level, and open-heavy strategies pay double spread.

# 3. Honest caveats

Holiday classes have n=144 → bands of ±21 bp: low power, reported as such (absence of evidence here is weak evidence of absence). Single instrument by design — cross-sectional calendar claims were deliberately out of budget.

# 4. Hand-off

Nothing from expE enters the expF candidate pool. Candidate 02 (H13 turn-of-month) keeps its pre-registered validation protocol but now carries a strong dead-on-arrival prior — the honest expectation is a Level-1-fails negative finding for the atlas. expF (correction battery) is next.