Replay is useful when the exercise has a defined scope. Choose a question before opening a historical sequence: recognising context, following an entry rule or reviewing how an invalidation was handled.
Define the exercise before revealing candles
Record the instrument, starting date, context timeframe and observation timeframe. Write the conditions that qualify an example and the conditions that invalidate it. Decide how you will record a decision to take no trade.
Choose a date that is available in the product. The supplied history may differ across instruments and timeframes. If a period cannot be loaded, record that limitation instead of replacing it silently with another sample.
Observe only what was available at the time
Advance through the sequence without using future candles to justify an earlier decision. If you have already inspected the outcome, label that example as a review rather than an independent test.
When switching timeframes, check the displayed date and cursor. A candle on a higher timeframe can include time beyond your decision point; distinguish the information already formed from the completed historical candle.
Keep a simple session log
- Starting date and instrument.
- Rule being examined and information used at the decision point.
- Entry, invalidation and planned exit, when applicable.
- Reason for skipping an otherwise eligible example.
- Outcome and deviations from the written process.
Review the exercise
Separate execution mistakes from cases where your rules were followed but the result was unfavourable. Do not rewrite the rule after every outcome. If you change it, start a new version and test it on a separate period.
Continue with preparing and reviewing a backtest for a broader sample. Replay observations and simulated results do not guarantee live results.