Define the scope→02
Inspect the source→03
Record the limits
Write the question before the test
Define one question and the observations that would answer it. Keep the entry rules, exit rules, instruments and evaluation window fixed for that experiment. Changing rules after seeing results makes the record harder to interpret.
Separate development and evaluation
Data used to develop an idea should be distinguished from data used to evaluate it. Record every assumption and any later change. A favourable historical result does not establish a future outcome.
Understand the demo environment
Demo execution and balances can differ from a funded account. Note the platform, account type and price source. Do not mix demo observations with the measured funded-account tests in our audit register.
Include operational costs and failures
If you model an experiment, document its assumptions about costs, data gaps and failed orders. Treat these as stated assumptions unless you have corresponding observations. Avoid presenting a clean chart as proof of a robust process.
Do not turn a broker test into a strategy claim
A measured spread or support response says nothing about whether an approach will make money. The comparison tool helps inspect account evidence; it does not recommend a strategy or automate trading.