best setup pattern miner

The problem

Traders often struggle to identify which specific behaviors lead to their biggest wins versus their consistent losses. Reviewing thousands of rows of trade data manually makes it nearly impossible to spot the statistical confluence of market regimes, sessions, and indicators that create a high-probability edge.

What it does

• Analyzes historical trade data from CSVs, journals, or narrative logs to isolate the top 20-30% of performers by R-multiple.
• Extracts distinguishing features across market regimes, time-of-day sessions, and technical indicators to find your statistical edge.
• Calculates expected value (EV) and win rates for specific setup conditions to filter out lucky outliers from repeatable patterns.
• Generates a structured Markdown playbook including entry triggers, stop-loss rules, and invalidation criteria based on your actual history.

Why this beats prompting it yourself

General LLM prompts often hallucinate generic trading advice or "best practices" that don't apply to your specific style. This skill forces an objective, data-driven analysis of your unique trade history, ensuring the resulting playbook is grounded in your actual performance metrics rather than theory.

Use cases

• Convert a year of messy trading journal notes into a standardized strategy document.
• Identify which specific tickers or sessions are draining your capital despite overall profitability.
• Build a "Kill Criteria" list to prevent taking low-probability setups that historically fail.
• Formalize a backtested strategy into a concrete execution plan for a prop firm evaluation.

Known limitations

Requires a minimum of 30 to 50 trades to provide meaningful statistical patterns. Results are highly dependent on the quality of input data, such as entry/exit prices and risk-per-trade metrics.