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%
1.2 KB · 23 lines json
Raw Blame History
1{2  "anomaly_id": "F002-spx-spy-staleness-artifact",3  "title": "The SPX->SPY minute-scale 'lead' is index content staleness, not price discovery",4  "summary": "SPY leads the SPX index print by one minute (+0.132 cross-correlation in train 2014-2015, +0.080 on the 2016-2021 validation split) at ~0.95 contemporaneous correlation, and the lead SURVIVES both-fresh print synchronization because the staleness lives inside the index computation (Fisher 1966, measured live). It also survives Hansen SPA - demonstrating that statistical correction cannot detect mechanism. Any 'ETF leads the index' claim on bar data is presumptively this artifact.",5  "negative_result": true,6  "key_numbers": {7    "train_lead_plus1min": 0.132,8    "validation_lead_plus1min": 0.08,9    "contemporaneous_corr": 0.965,10    "artifact_share_from_print_gaps": 0.011  },12  "evidence": [13    "results/expB_artifact_baselines/",14    "results/expD_leadlag_scan/",15    "results/expH_oos_stability/"16  ],17  "reproduce": "python experiments/micro/expD_leadlag_scan/benchmark.py; expH xcorr section",18  "author": "Simon-Pierre Boucher",19  "contact": "contact@spboucher.ai",20  "data_source": "hfmarketdata.io",21  "confidence_level": 222}23