MT5 Strategy Tester vs live trade history: why they differ

Where tester results live, why they are not in the History tab, and how to compare them with real trades.

The MetaTrader 5 Strategy Tester can replay years of price data in minutes and produce an equity curve, a profit factor and a drawdown figure. It is a useful way to check an idea. It is also the source of a very common mistake: mixing tester numbers with live account numbers, or expecting the two to agree. This page explains where tester results live, why they are not in your account history, what makes them differ from live trading, and how to keep both in a journal without confusing them.

Two different records

When you run a test, the terminal builds a report of simulated trades inside the Strategy Tester panel, with tabs such as results, graph and report. Those trades never touched the broker’s server. They do not appear in the History tab of the Toolbox, because that tab shows the real trades of the account you are logged in to. The tester has its own settings, its own virtual deposit and its own report. If you ever see backtest trades in your account history, something unusual is going on; check you did not log in to a demo account used for testing.

Why results differ from live

Several well-known gaps explain most differences. None of them mean the tester is broken; they mean a model is not the market.

• Modelling mode. The tester can run on every tick, on real ticks if the broker provides them, on one-minute bars or on opening prices only. Cheaper modes are faster and less detailed, and they can misplace stop loss or take profit hits inside a bar.
• Data quality. History data may have gaps, or may not match your broker’s feed exactly.
• Spread. A test may use a fixed or the recorded spread, while live spread widens at the rollover and around news.
• Slippage and requotes. These are not part of a basic test. The live side is covered in spread and slippage in the trade history.
• Execution rules. Filling policies and execution modes differ by broker; see execution modes and filling policies.
• Overfitting. Settings tuned until the past looks perfect rarely repeat.

Check the tester’s settings screen for the modelling option you used; names vary slightly by build. Writing that option next to the result is a good habit, because a figure without its mode is hard to compare.

What a fair comparison looks like

Comparing a backtest to live results is most useful when both use the same period, the same symbol and the same position sizing rule. Compare shape and behaviour, not exact profit: the number of trades per week, the average holding time, the share of winners, the size of the biggest losing streak and the maximum drawdown. If the live win rate is far below the test, the first suspects are costs and fills, the next are rule changes you made during live trading. The statistics themselves are explained on MT5 trading statistics, and the demo side of the same question is on demo vs live account journal.

• Fix a window, for example the last three months, and cut both records to it.
• Use the same sizing. Percent-of-equity sizing and fixed lots are not comparable.
• Compare process measures first and profit last.
• Note every rule you changed since the test began.

Where a journal fits

A journal records what happened on a real account. AI Trading Journal, a Windows app for MetaTrader 5 only, reads the closed positions of the account you connect, one entry per closed position with partial closes combined and the result calculated as profit plus commission plus swap, dated by the last exit on the broker’s server clock. It does not import Strategy Tester reports and it does not record intended prices or slippage, so it cannot do the backtest side for you. What it can do is give you the live half of the comparison in a file on your own PC, with your notes and tags, so the comparison is made against recorded facts and not memory.

For the test side, keep a short note or sheet per run: the idea, the symbol, the period, the modelling mode, the settings and the headline results. A free structure for that is in the Notion backtesting template page. The app is read-only: it never places, modifies or closes a trade.

A worked example (hypothetical)

A trader tests a simple breakout idea on one symbol over two years using one-minute bars and records a profit factor of 1.6 and a maximum drawdown of nine percent. After ten weeks live, the journal shows a profit factor of 1.1 and a drawdown of twelve percent. Rather than concluding the idea is dead, the trader compares the pieces. Average spread paid live was larger than in the test, three trades were skipped because of a news filter that was not in the test, and two entries were taken late. Rerunning the test with realistic spread and the news filter lowers its profit factor to 1.3. The remaining gap is attributed to late entries, which become the next thing to track in notes. All numbers are invented.

Limits and cautions

• A good backtest does not predict live profit, and nothing on this page is a promise or advice.
• Modelling options and report names differ between builds; check your terminal.
• The app records closed live positions only; it does not read tester reports.
• Keep tester and live notes separate so the two are never blended into one statistic.

Frequently asked questions

Related guides

• A local trading journal for MT5
• P&L calendar for MT5
• Trade notes, ratings and screenshots
• Trading progress and goals
• Back up your journal and move it to a new PC
• An offline trading journal for Windows
• How to set monthly trading goals
• Import your whole MT5 history
• How to tag trades
• Trade screenshots for your journal
• All trading journal comparisons