The receipts behind every signal
Methodology, evidence pedigree, and calibration accuracy for every detector in the product. If a signal can't survive Hou-Xue-Zhang replication or McLean-Pontiff post-publication decay, it shouldn't earn a high grade — and you should be able to see why here.
Signal Graveyard
Every detector ships with a research profile: anchor paper, sample size, period, reported edge, decay risk, microcap exposure, and capacity warnings. Read this before you trust any conviction.
Open ▸
Calibration Lab
Brier score, log loss, expected calibration error, and conformal coverage per feature × horizon. When a probability says 70%, does it actually hit 70%? This is where you check.
Open ▸
Research Chat
Free-form market research chat with history persisted across sessions.
Open ▸
Methodology + Audit Trail
How every score is built, what we measure, and where we publish our forward-proof ledger. Read this before you ship money behind any signal.
Open ▸
Social-Arb Scout
Camillo-style trend detector. Tracks a curated tag library of brands, products, people, and cultural movements across public Reddit feeds and ships ATR-sized theses when two sources corroborate a multi-σ mention spike. Ideas, not advice.
Open ▸