Gold robot backtest guide: validate an XAUUSD EA before going live

In the MT5 Strategy Tester, select "every tick based on real ticks" and use at least 12 months of data from your broker. Real ticks reproduce the actual spread and price movement, which is critical for a scalping or intraday gold EA.

If your broker does not provide enough tick history, the test is less reliable. Consider downloading additional data from a trusted source or switching to a broker with deeper history.

Gold spreads vary widely. Use the average spread you see during your main trading session, not the tightest spread of the day. Add commission if your account charges it per lot.

A strategy that looks profitable with a 15-point spread may fail with a 35-point spread. Always test a range of spread values to see how sensitive the result is.

Focus on profit factor, maximum drawdown, consecutive losses and total number of trades. A profit factor above 1.3 with a max drawdown under 15% and at least 200 trades is a reasonable starting point.

Ignore the final return until you understand the drawdown. A 300% return with a 60% drawdown is not the same as a 50% return with a 5% drawdown.

A smooth, steadily rising equity curve with controlled drawdowns is better than a jagged curve with sudden spikes. Sudden gains often come from one lucky period and do not repeat.

Look at monthly breakdowns. A strategy that only profits in one or two months may be overfitted to a specific market condition.

After a promising backtest, run the robot on a demo account for at least two to four weeks. Compare live fills, spread and slippage with the backtest assumptions.

Only then move to live with a small portion of your capital and a conservative risk setting. Increase size only after the live results confirm the backtest story.