Backtesting: look beyond the equity curve.

What a backtest actually tells you

A backtest applies a set of rules to historical data. It describes how those rules behaved under the assumptions of that test. It cannot establish how the strategy will perform in a different market environment.

Before interpreting the equity curve, record the symbol, data source, timeframe, date range, spread model, commission, leverage and execution assumptions. These choices can materially change the result.

Start with data and execution costs

Use the most detailed historical data available for the strategy. Intrabar decisions require suitable tick data; a test based on bar prices may miss the path that price took. Account for variable spreads, commission, swaps and plausible slippage.

Review missing data and unusual price gaps. A strategy that enters frequently can be particularly sensitive to small execution costs. Testing a range of costs is more informative than choosing one optimistic number.

Separate development from evaluation

Choose an out-of-sample period before tuning parameters. Use the development period to explore an idea, then evaluate the frozen rules on untouched data. Repeatedly checking the same evaluation period gradually turns it into another development set.

Walk-forward evaluation can help explore how parameters behave over time, but it does not eliminate selection bias. Keep a research log, including unsuccessful experiments, to make your process easier to audit.

Examine the risks hidden by the curve

Look at drawdown duration as well as depth. Review losing streaks, trade concentration, exposure and the number of independent observations. Many trades generated from the same market episode can provide less evidence than the raw count suggests.

A useful review also asks whether one month, instrument or parameter setting accounts for most of the result. Fragile performance deserves further investigation, not a larger allocation.

Move to forward testing deliberately

Run the frozen configuration in a demo environment. Compare intended signals with actual orders and inspect terminal logs. Verify trading hours, broker symbol names, minimum volume and how the EA handles disconnections.

Read our position-sizing guide before choosing exposure, and our MetaTrader 5 setup checklist before installation. Historical results and demo execution do not guarantee future results. This article is educational and is not investment advice.