Rigor before conviction.
The methodology page exists to make the research defensible. It should explain what was tested, how it was tested, and where the conclusions stop.
1. Define the Event
Every study starts with a precise market event: session, time window, first-touch rule, break condition, and outcome window.
2. Remove Lookahead
Features must only use information available at the decision timestamp. Anything known only after the event belongs in outcome analysis.
3. Validate Samples
Sample sizes, missing days, duplicate bars, holidays, contract roll behavior, and abnormal sessions are documented.
4. Compute Results
Reports should include conditional probability, confidence intervals, expectancy, MAE/MFE, and failure distributions when available.
5. Segment Regimes
Studies become more useful when separated by volatility, range, open location, formation order, trend context, and prior-session structure.
6. Publish Limits
Each report should state what the result does not prove. A clean limitation section is part of the brand.
Logic. Math. Walkthroughs.
Version two of the site includes dedicated pages for the study pipeline, the mathematics behind the results, and a page that walks a single study from raw data to final probability.
Mathematical Framework
Explains conditional probability, expectancy, R-multiples, and confidence intervals.
Open Page →Function Walkthrough
Shows exactly how a single study is constructed from raw session data.
Open Page →