Common MT5 Strategy Tester Mistakes That Skew Results

When a backtest doesn’t translate to live performance, most traders assume the market changed or the strategy doesn’t work. Sometimes that’s true. Often the backtest itself was structurally broken from the start, producing numbers that looked valid but were never measuring what the trader thought they were measuring.

The Modelling Method Nobody Checks

Open Prices Only and Control Points are the two modelling methods that most traders leave the Strategy Tester on without thinking about it. Both evaluate the EA’s logic at specific points within each bar, either only at the open or at a handful of internal points, rather than across the full tick sequence.

If the strategy makes any intrabar decisions, entries triggered by price hitting a level mid-candle, stops activated by an intrabar spike, exits based on indicator crosses within the bar, then these modelling methods will either miss those events entirely or trigger them at the wrong time. The test looks like it’s running the strategy. It isn’t. It’s running a simplified approximation that doesn’t replicate how the EA would actually behave on live price data.

Every Tick or Every Tick Based on Real Ticks is the correct setting for any EA with intrabar logic. Leaving the default in place and wondering why results don’t match live trading is one of the most common MT5 strategy tester mistakes that skew results.

The Time Zone Problem That Corrupts Session-Based Tests

MT5 Strategy Tester runs on the server time of whatever data source it’s using. If that GMT offset differs from the live broker’s server time, a session-based strategy coded to trade specific hours will fire during the wrong window in the backtest.

A strategy built to trade the London session by checking whether the hour is between 8 and 12 on a data source running UTC+0 will test completely differently when the live broker runs UTC+3. The backtest appears to test the London session. It’s actually testing a different three-hour window with different liquidity and volatility characteristics. The equity curve reflects performance in the wrong session, not the one the strategy was designed for.

Verify the GMT offset of both the test data and the live broker server before drawing any conclusions from session-dependent backtests.

Look-Ahead Bias: Knowing the Future Without Realising It

This one is subtle and consistently produces the most inflated backtest results. It happens when the EA references an indicator value that includes data from bars that would not have been complete at the time of the trade decision.

The most common version: using the previous bar’s close value for a signal, then executing on the same bar rather than the next. The previous bar’s close is only known after that bar closes. In a live environment the EA acts on the opening of the next bar. In a backtest without proper handling of this sequence the EA appears to know the close before it happens. The result is a strategy that looks far more profitable than it can ever be in live conditions because it’s making decisions with information that didn’t exist at the time.

Position Sizing That Makes Drawdown Statistics Useless

Running a backtest at a fixed lot size and then comparing the drawdown percentage to a prop firm’s maximum drawdown limit is a common MT5 strategy tester mistake that skew results in a specific and dangerous way. Fixed lot drawdown statistics don’t reflect percentage-based risk the way a live funded account operates.

An EA running at a fixed 0.1 lots on a $10,000 account shows a different risk profile than the same EA running at 1% risk per trade on the same account, particularly as the account balance changes during the test. If the live funded account will use percentage-based sizing, the backtest needs to match that model or the drawdown statistics in the report are not measuring the right thing.

Conclusion – Common MT5 Strategy Tester Mistakes That Skew Results

common MT5 strategy tester mistakes that skew results all share the same outcome: a backtest that looks like validation but isn’t. Check the modelling method before running anything with intrabar logic. Verify the time zone alignment for session-based strategies. Audit the EA’s bar referencing logic for look-ahead bias. And match the position sizing model in the test to what will actually run on the funded account.

FAQ – Common MT5 Strategy Tester Mistakes That Skew Results

1. How do I check whether my EA has look-ahead bias?

Review the MQL5 code specifically for any indicator call that uses shift 0 on a bar that hasn’t closed yet, combined with an execution that fires on the same bar rather than waiting for confirmation. If you’re using a pre-built EA and can’t review the code directly, run the test in visual mode and compare the entry signals to what you’d expect from a manual chart read at those points. Entries that seem to anticipate the close of a bar are a signal of look-ahead bias.

2. Does the time zone issue affect all MT5 brokers the same way?

No. Each broker’s MT5 server runs on a specific GMT offset and that offset varies between brokers. Some run UTC+2, some UTC+3, some adjust for daylight saving and some don’t. Always check the current server time displayed in the bottom right of the MT5 terminal and compare it to the data source time used in the Strategy Tester.

3. Is Open Prices Only ever acceptable for backtesting?

For daily or weekly timeframe strategies where the EA logic only fires at bar open and holds positions for multiple bars without intrabar exit conditions, Open Prices Only can be acceptable and runs significantly faster. For anything that checks intrabar conditions, it’s not appropriate and the results should be discarded.

We have helped thousands of traders reach funding at TTT Markets from account sizes of $5k upwards to $500k. Check out our programs. 

Additional resources:

Wrong Strategy Tester results – Strategy Tester – General – MQL5 programming forum 

MT5 Strategy Tester: Complete Backtesting Guide — Forex Robot Easy