Quick answer: Open the Strategy Tester in MT5 (Ctrl+R), pick the EA, symbol, timeframe and date range, choose 'Every tick based on real ticks' for accuracy, set your deposit and leverage, run it and then study drawdown, profit factor and the equity curve rather than net profit alone.
Key takeaways
- Use real tick data for the most honest results.
- Test a few years, including different market conditions.
- Drawdown and consistency matter more than the final number.
- A backtest ignores some live costs like slippage.
- Over-optimized settings usually fail forward.
Why backtest at all
A backtest runs an EA on historical data to show how its rules would have behaved. It won't predict the future, but it can show how deep losing streaks got and whether the rules made sense in different markets.
Step-by-step: running a backtest
- Open View → Strategy Tester or press Ctrl+R.
- Select the Expert Advisor.
- Choose the symbol and timeframe the EA was built for.
- Set the date range. A few recent years is a good start.
- Choose the modeling: Every tick based on real ticks is the most accurate.
- Set deposit, currency and leverage similar to your real account.
- Load the EA's inputs, then click Start.
How to read the results
- Net profit: the final result. Least useful on its own.
- Maximal drawdown: the biggest fall from a peak. Ask yourself if you could live through it.
- Profit factor: gross profit divided by gross loss. Above 1 means winners outweighed losers in the test.
- Total trades: too few trades means the result may be luck.
- Consecutive losses: how long the worst losing streak lasted.
- Equity curve: a smooth line with shallow dips is easier to trade than a jagged one with the same profit.
What a backtest leaves out
- Live slippage and changing spreads.
- Broker execution delays and requotes.
- Your own reaction to a losing month.
That's why you still need a demo period before going live.
Avoid overfitting
Optimizing inputs until the past looks perfect usually produces settings that fail on new data. Keep optimization light, test on a period the settings never saw, and prefer settings that work reasonably across many periods over settings that are perfect on one.
Backtests and marketing
A seller who only shows a single perfect backtest, without drawdown or dates, is showing you marketing, not evidence. Read the red flags.
Frequently asked questions
How many years should I backtest an EA?
Several recent years with different market conditions is a sensible start. Very old data may not reflect current spreads and behavior.
What is a good profit factor?
Above 1 means the test made more than it lost. Treat very high numbers with suspicion, especially with few trades.
Why is my live result different from the backtest?
Live trading includes slippage, spread changes and execution delays that a backtest cannot fully model.
Which modeling mode is best?
'Every tick based on real ticks' is the most realistic in MT5, though it takes longer to run.
Want to see an EA built around limits you control? Remix Ai is an Expert Advisor for MetaTrader 5 with lot size, session hours, drawdown and equity limits. Launch price $199, one-time payment.
Risk disclosure: this article is for education only and is not financial advice. Trading with automated software carries a high risk of loss, and past or simulated results do not predict future results. Remix Ai sells software. We do not manage accounts, offer copy trading or sell trading signals.