A backtest simulates defined rules on historical data. It describes a result under specific assumptions, not future tradability or returns.
Describe the experiment before the result
Record the hypothesis, period, markets and code version before evaluating the test. Keep track of variants already attempted. Archiving only the best run removes context needed to interpret it.
Execution is an assumption
Spreads, fees, slippage and order timing affect simulation. The MT5 tester offers different tick models and delay assumptions. Coarse data can support an initial check; it does not replace execution testing appropriate to the strategy.
Keep new data separate
A change made after seeing a poor result is a new development decision. The previously untouched period is no longer an independent test. Separate development, historical evaluation and later demo forward observation in the report.
A practical example
Documentation example: develop version A on period 1 and evaluate it on period 2. A change creates version B. Preserve results for both versions; do not retroactively describe version B as the original plan.
For your next test
- Document source, symbol, time zone and gaps.
- Include fees and execution settings.
- Disclose negative runs and limitations.
Common question
Is MT5 forward testing already live trading?
No. The tester’s forward feature splits historical periods. A later demo test observes new data and must be reported separately.
Sources & further reading
Educational content, not investment advice. Numerical examples are hypothetical, not results of a REVENQOR system.
Continue learning
Recognize overfitting before taking a system live →
Understanding drawdown: depth and recovery →
Explore Systems Lab ↗