Two traders can test the same Expert Advisor over the same period and still receive different results.
This does not always mean that the EA or the Strategy Tester is wrong. In many cases, the difference comes from the broker environment used for the test.
The practical point
An EA does not trade an abstract chart. It trades the prices, symbol settings and execution conditions supplied by a specific broker.
The following four areas can materially affect a backtest:
1. Spread and trading costs
A wider spread changes both entry and exit prices. This is especially important for short-term systems and XAUUSD strategies, where a relatively small change in cost can turn a marginal trade from a win into a loss. Commission and swap should also be included when they apply to the account.
2. Historical tick data
Different brokers can have different tick sequences, missing periods and price spikes. A strategy that reacts to intrabar movement may therefore enter at a different moment—or may not enter at all.
For this reason, “Every tick based on real ticks” is generally the most informative MT5 testing mode when suitable broker data is available.
3. Symbol specifications
Check the symbol’s digits, tick size, tick value, contract size, minimum lot, lot step and minimum stop distance.
These values affect position sizing, margin calculations and whether an order can be placed as intended. Even XAUUSD and GOLD symbols can differ considerably between brokers.
4. Execution and server conditions
Zero-delay testing represents ideal execution. Live trading can include latency, slippage, requotes and rejected modifications.
MT5 can emulate fixed or random delay, which helps show whether a strategy is highly sensitive to execution speed.
What traders should check
Before judging an EA from one attractive report, I recommend checking:
-
the correct symbol and recommended timeframe;
-
the quality and coverage of the historical data;
-
realistic spread, commission and swap assumptions;
-
tick size, tick value, contract size and volume step;
-
results with a fixed or random execution delay;
-
more than one market period, including an out-of-sample period;
-
performance on a demo account using the intended broker.
The goal is not to force every broker to produce an identical equity curve. The goal is to understand how much the strategy changes when realistic conditions change.
This is also why I recommend testing the free demo of Ninja Trinity XAU in the MT5 Strategy Tester with your own broker’s XAUUSD or equivalent GOLD symbol before considering live use.
Ninja Trinity XAU:
https://www.mql5.com/en/market/product/185737
Closing note
A strong backtest is useful, but it is only one part of EA evaluation. Broker-specific testing, delay testing and demo forward testing provide a more complete picture of how an automated strategy may behave.
Risk notice
Backtest results are hypothetical and do not guarantee future performance. Live results can differ because of spread, slippage, execution delay, commissions, swaps, liquidity and market conditions. Use only risk capital and choose a lot size appropriate for your account.
