Basically FuturesUnderstand it
Version your strategy definition, dates, datasets, costs, observed metrics and failed experiments. A result without assumptions cannot be audited. Include sample size, regime coverage and uncertainty alongside win rate or profit factor. The research notebook should distinguish observed evidence from an interpretation. When a rule changes, create a new version and retest rather than quietly rewriting history.
Version 1 uses a 10-bar entry filter. Version 2 switches to 20 bars after an exploratory test. Keep both results and the reason for the change. Your next validation period should not be reused repeatedly to select versions.
Watch for this misconception
Deleting underperforming versions until the notebook tells a neat success story.
Draft a one-page research entry with hypothesis, dataset, costs, result, limitations and next step.
Expectancy calculator ↗What should happen when you materially change a strategy rule?
Check the sources
Hypothetical educational examples. Verify current product specifications and rules before making decisions.