Why Your XAUUSD Backtest Looks Perfect and Your Live Account Doesn't
You found a gold EA, ran a quick backtest, saw a smooth equity curve — and the first week live looked nothing like it. Before blaming the EA, check how you tested it. On XAUUSD the tester mode can change the result more than any input.
The three tester traps on gold
- "Open prices only" or "1 minute OHLC": the tester guesses what happened inside the candle. Gold moves fast inside a single M1 bar, so stop-losses, take-profits and basket exits get filled in an order that never happened.
- "Every tick" (generated): better, but ticks are synthetic. Spread spikes around news and rollover are smoothed away.
- "Every tick based on real ticks": uses your broker's actual tick history, including real spread widening. This is the only mode worth trusting for a scalping or basket EA on gold.
A 5-minute honesty check
- Run the same EA, same dates, in generated ticks and in real ticks.
- If the profit or drawdown changes a lot, the strategy depends on the tester's assumptions, not on the market.
- Repeat on a second period that you did not use for choosing settings.
- Look at the worst floating loss, not only the final balance.
I build HyperVirtual (an MT5 gold EA) with this in mind: native MQL5, no DLLs, and every order is validated against the broker's rules before it is sent, so a failed or rejected order is handled instead of silently breaking the cycle. You can check this yourself, you do not need to trust a screenshot. Product page: https://www.mql5.com/en/market/product/198552
Run it on your own data
Click Free Demo on the product page, open Strategy Tester (Ctrl+R), choose XAUUSD, M1 and "Every tick based on real ticks", then compare the result with any other gold EA you own. Same dates, same deposit, same mode.
If you need a quick lot-size helper for your own tests, my free CodeBase tool is here: RiskLotCalculator.
Download the free demo and test it yourself: https://www.mql5.com/en/market/product/198552
