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

# Candidate ranking (Phase 4)

All 22 hypotheses from `research_gaps.md`, scored 1–10 per charter-§7 axis
(10 = best). Axes: **T** testability on hfmarketdata.io · **N** cleanliness
of the artifact null · **O** OOS feasibility · **M** multiple-testing
discipline · **S** survival odds after costs (5 = n/a, structural claim) ·
**V** novelty of the honest test · **R** reproducibility · **P** public-atlas
value · **I** implementation simplicity · **D** safety against
self-deception.

| H | Hypothesis (short) | T | N | O | M | S | V | R | P | I | D | Σ |
|---|---|--|--|--|--|--|--|--|--|--|--|---|
| H15 | Intraday momentum post-2018 decay | 9 | 7 | 9 | 9 | 6 | 9 | 9 | 10 | 7 | 8 | **83** |
| H11 | SPX→SPY lead = staleness artifact | 9 | 9 | 8 | 9 | 5 | 8 | 9 | 9 | 8 | 9 | **83** |
| H13 | Turn-of-month survived 2006–2026 | 9 | 8 | 9 | 8 | 6 | 8 | 9 | 9 | 8 | 8 | **82** |
| H03 | Index VR residual = Fisher share | 9 | 9 | 8 | 9 | 5 | 7 | 9 | 8 | 8 | 9 | **81** |
| H16 | Overnight premium convention share | 9 | 9 | 8 | 9 | 5 | 7 | 9 | 8 | 8 | 9 | **81** |
| H20 | Intraday artifact U-profile (input) | 9 | 9 | 8 | 9 | 5 | 6 | 9 | 8 | 8 | 10 | **81** |
| H14 | Monday effect stays dead (control) | 9 | 9 | 9 | 9 | 5 | 4 | 9 | 7 | 9 | 10 | **80** |
| H09 | ES→SPY ordering at the 1min floor | 8 | 8 | 8 | 9 | 5 | 8 | 9 | 8 | 7 | 8 | **78** |
| H10 | Crypto weekend → Monday equity gap | 8 | 7 | 7 | 9 | 5 | 9 | 8 | 9 | 7 | 7 | **76** |
| H21 | Vol persistence (positive control) | 9 | 9 | 9 | 9 | 5 | 3 | 9 | 6 | 7 | 10 | **76** |
| H01 | Liquid 1–60min reversion = 0 net | 9 | 9 | 8 | 8 | 2 | 4 | 9 | 7 | 8 | 9 | **73** |
| H08 | SPY→sector ETFs zero on fresh pairs | 9 | 8 | 8 | 8 | 3 | 5 | 9 | 6 | 8 | 9 | **73** |
| H07 | Large→small decay + artifact share | 8 | 8 | 8 | 7 | 3 | 6 | 8 | 8 | 6 | 8 | **70** |
| H05 | Crypto 1min reversion by maturity | 8 | 7 | 7 | 7 | 4 | 7 | 8 | 7 | 7 | 7 | **69** |
| H17 | Crypto hour-of-week calendar | 8 | 8 | 8 | 5 | 4 | 7 | 8 | 7 | 7 | 5 | **67** |
| H02 | Daily reversal decay curve | 8 | 7 | 8 | 7 | 2 | 5 | 8 | 7 | 6 | 7 | **65** |
| H18 | Holiday effects are dead | 7 | 8 | 7 | 7 | 4 | 4 | 8 | 5 | 8 | 7 | **65** |
| H12 | Options activity → next-day vol | 7 | 6 | 8 | 7 | 5 | 7 | 7 | 7 | 4 | 6 | **64** |
| H22 | OpEx-week volume/vol patterns | 7 | 6 | 7 | 6 | 4 | 7 | 8 | 7 | 6 | 6 | **64** |
| H19 | DST-transition distortions | 7 | 7 | 7 | 7 | 4 | 6 | 8 | 5 | 6 | 5 | **62** |
| H04 | Post-jump 1min overreaction | 7 | 5 | 7 | 6 | 3 | 7 | 7 | 7 | 5 | 5 | **59** |
| H06 | FX session-boundary reversion | 7 | 6 | 7 | 5 | 3 | 6 | 8 | 5 | 6 | 6 | **59** |

## Selected for prototyping (charter: 3–5)

**Candidate 01 — H15, intraday momentum post-publication decay** (Σ83).
The single highest-information test: a famous published effect (Gao et al.
2018) with a freezable specification, an untouched 2018–2026 out-of-sample
window that only time could create, a mechanism proxy available in our
options chains (Baltussen et al.), moderate cost exposure (2 trades/day in
SPY), and near-zero researcher degrees of freedom. Every outcome —
persistence, decay, or reversal — is a publishable McLean–Pontiff-style
data point.

**Candidate 02 — H13, turn-of-month 2006–2026** (Σ82). The last calendar
survivor per Marquering et al. (2006), untested since on open data;
low-frequency (12×/yr → real cost-survival chance), one pre-registered
window, permuted-calendar null + SPA. Highest chance in the whole list of a
genuine Level-3 finding — and a clean negative if dead.

**Candidate 03 — H11 + H03, the index-artifact decomposition** (Σ83/81).
One machinery, two published deliverables: prove the SPX→SPY 1min "lead" is
pure print-staleness (result-type C/D: artifact demonstration), and quantify
the Fisher share of index-level variance-ratio momentum. Cleanest nulls of
the entire list; already seeded by expB (+0.065 measured).

**Candidate 04 — H10, crypto-weekend → Monday equity gap** (Σ76, novelty
pick). The one genuinely NEW question this dataset is structurally
positioned to ask (24/7 crypto × equity session opens). Pre-registered
exposed set {COIN, MSTR, RIOT, MARA, HUT} + placebo set + SPY-gap control +
permuted-weekend null. Higher self-deception risk (selection, short joint
history ~400 weekends) — ranked last of the selected four accordingly.

**Reserve — H09** (ES→SPY at the 1min floor): promoted if a selected
candidate dies early at the data layer.

## Not selected (why, in one line each)

Controls/inputs H14, H20, H21 run *inside* the micro-experiment pipeline
(expE, expB/expG, framework) — they are mandatory, not candidates. H16
(strong score) folds into candidate 03's convention machinery as a shared
deliverable. H01/H08 are expected-negatives that expC/expD produce as scan
output without candidate-level investment. H07/H02 are decay re-measurements
scheduled after the scans (reuse candidate-03 machinery). H05/H17 (crypto)
wait for volume-semantics verification (expA follow-up noted in the data
profile). H12/H22 need the options ingestion layer (heavier implementation).
H04/H06/H19 have the weakest nulls or lowest power — revisit only if the
scans surface something.

**Multiple-testing accounting:** selection made on *priors and design
cleanliness*, not on any real-data effect sizes — no data peeking occurred
beyond expA/expB artifact levels. The 22-hypothesis budget stands for expF.