capital markets · crypto · macro · technicals · validation
Research
Quant research for funds, desks and allocators. Every figure sourced, every signal tested against luck.
- desks
- 5
- measures, each cited
- 27
- asset classes
- 2
desks
measures, each cited
asset classes
Π01
The desks
Five desks, one standard.
Listed equities, US filers firstdesk EQ
Equity fundamentals
Every figure traced to the filing it came from.
What it measures
- Earnings qualityEDGAR financial statements
- Non-GAAP gapEDGAR filings and exhibits
- Beneish M-ScoreBeneish, 1999
- Insider purchasesCohen, Malloy and Pomorski, 2012
- Catalyst calendarEDGAR full text and event data
Deliverables
- Company deep dive
- Earnings quality screen
- Sector primer
- Market entry study
- Thirty-day company brief
Π02
The measurement, in figures
What the method does to a number that looks like skill.
A headline Sharpe, measured
20 variants · 730 days · seeded
Probability of skill, one test
94.5%
After twenty trials
35.0%
Twenty strategies with no edge. The best looks like skill until it is measured. Run it yourself.
- 0.9004Deflated Sharpe on the published worked example
- 12,870Overfitting splits measured in about one second
- 1.34Best Sharpe among 1,000 strategies with no edge
- 48,000Seeded seasons behind every scoring formula
- 2Independent protocols behind every on-chain close
- 9Invariants the ranking holds under attack
Π03
Commission research
Bring the market and the question. Scope and price are agreed in writing first.
A person replies within two working days
Research is analysis published for professional readers, never a personal recommendation to buy or sell an instrument.