How to review a strategy after launch

Use the same return convention, trade filters, costs, and exposure definition when comparing live outcomes with the research. A live result with different symbol, session, risk, or manual intervention is not a clean test of the candidate.

Record the live sample size and elapsed time before drawing a conclusion.

Check whether signals arrived when expected, orders filled, stops were placed, the terminal stayed connected, and risk limits behaved correctly. A negative period with correct behaviour may be variance; a positive period with broken controls is not a success.

A review should have pre-defined triggers: operational errors, material execution drift, a drawdown threshold, or a statistically unusual change in trade behaviour. Avoid changing parameters just because the latest trade was disappointing.

When you pause, preserve the logs and configuration so the diagnosis is based on evidence rather than a moving target.