Pancake — Verifiable Backtests for AI Agents

Use Pancake to run a verifiable backtest for this AI trading strategy: given the strategy logic, historical market data, and trading cost assumptions; output return, max drawdown, and Sharpe ratio, and state whether the result is reproducible.

Deterministic backtest results, key performance metrics, and a reproducibility note.

Re-run the strategy backtest with the same dataset version and parameters, verify whether the result matches exactly, and list any sources of discrepancy.

A line-by-line consistency check and analysis of discrepancy causes.

Backtest this AI trading strategy across market regimes, such as bull, bear, and range-bound markets, and summarize performance and risk exposure in each regime.

A regime-by-regime performance comparison, risk profile, and concise conclusion.