Performance visibility

• Equity curve, cumulative R, drawdown, expectancy, plan adherence, outlier wins, outlier mistakes, and model trades.
• Trade-level review of entry, exit, R multiple, MFE, MAE, setup, mistake tags, and what the evidence showed.
• Strategy Memory links scanner evidence, plans, replay runs, journal entries, research notes, privacy labels, and audit checks.

Backtesting boundaries

• Historical replay must include source, fill, slippage, spread, look-ahead, and assumption manifests.
• AI review should explain process evidence and propose paper tests; it must not promise expected future returns.
• External engines can be used behind isolated adapters, but the app UI remains clean-room and broker-safe.