trade-hypothesis-ideator

Testable trade strategy hypotheses are generated by providing a normalized JSON input bundle with trade logs and market context. The system outputs 1-5 ranked, falsifiable hypothesis cards containing proposed rule changes, kill criteria, and minimum viable experiments for A/B backtesting.

A falsifiable trade hypothesis is a structured thesis with a defined mechanism, evidence basis, and kill criteria that can be proven invalid through experimentation. You need one to run efficient minimum viable experiments, reduce false-positive entries, and rapidly triage candidate strategy ideas.

The best way to structure trade analytics data for hypothesis testing is using a normalized JSON input bundle. This bundle must contain specific keys: objective, strategy_context, constraints, market_context, trade_log_summary, and journal_snippets to successfully generate ranked hypothesis cards.

Yes, the generated trade hypothesis output works with backtesting setups by validating outputs against included JSON schemas and exporting pursue recommendations. It generates strategy.yaml and metadata.json files compatible with edge-finder-candidate/v1 for downstream A/B backtesting integration.

Yes, you can use journal snippets and market context for strategy ideation by including them in your normalized JSON input bundle. The system processes these inputs alongside trade logs to apply critique prompts, score evidence strength, and surface ranked trade hypothesis cards.

You should not use automated hypothesis generation when you lack a normalized JSON input bundle containing structured trade logs and market context. The system requires objective, strategy_context, constraints, trade_log_summary, and journal_snippets to produce valid, falsifiable trade hypothesis cards.