DATA COLLECTION
Collect & preserve context
Market structure, price behavior, options, liquidity, market-state context, and research outcomes are organized for analysis.
METHODOLOGY
Evidence before confidence.
The public methodology page explains the philosophy of the research workflow without exposing proprietary code, thresholds, or model internals.
RESEARCH CANDIDATE
Turn an observation into a testable hypothesis
Interesting patterns are treated as candidates rather than automatically assumed to be useful.
HISTORICAL RESEARCH
Measure behavior across relevant slices
Ticker, regime, time of day, direction, volatility, and related context can all matter.
VALIDATION
Test against data the idea did not optimize on
Holdout and later-unseen evidence help separate a repeatable finding from an overfit result.
PROSPECTIVE MONITORING
Observe what happens next
Research can remain in shadow or prospective collection while additional evidence accumulates.
REVIEW
Promote, continue watching, or reject
Passing one test does not imply automatic live effect. Research and production decisions remain separate.
◎
Objective
Measure data before telling a story about it.
◇
Transparent
Explain status, evidence, limitations, and next steps.
✓
Validated
Later-unseen evidence matters more than in-sample excitement.
↻
Continuous
Research remains open to new evidence and changing market behavior.