Read a backtest the way a sceptic would
A strategy tester report is a claim about the past made under assumptions you did not choose. This reads the report MetaTrader wrote, recalculates the statistics from the individual trades, and tells you which of those assumptions would invalidate the result.
Diagnosis
Ordered by what would invalidate the result first. A backtest can be profitable and still be worthless if the conditions it was produced under do not exist.
What this report actually is
The conditions the test ran under. These decide how much the numbers below are worth.
Equity from the closed trades
Built by accumulating each closed trade onto the starting balance. The dashed line is where you started.
Statistics, recalculated
Every figure here comes from the trade list in the report, not from the summary MetaTrader printed. That is why they can be compared across MT4 and MT5 reports.
| Metric | Value |
|---|
Drawdown here is measured on closed trades. MetaTrader's own drawdown figure includes unrealised loss, so it is usually the larger of the two — the gap between them is the pain the equity curve hides.
Month by month
Where the profit came from. A long test with only a few active months is a short test wearing a long one's dates.
| Month | Profit | Trades |
|---|
How this is calculated
Statistics are recalculated from the trade list rather than read from the report's own summary. MetaTrader prints its labels in the language of the terminal that produced the file, but the deal type column stays in English in every language — so reading the trades is the only way the same code can handle a report from any terminal. It also means MT4 and MT5 reports end up in the same table.
Read the diagnosis before the numbers. Modelling quality, spread assumptions and sample size decide whether the profit factor above means anything at all; a profit factor of 3 over 40 trades at 25% modelling quality is not a better strategy than a profit factor of 1.3 over 2,000 trades at 99%.
Two figures will not match the report: drawdown, because this measures closed trades while MetaTrader includes unrealised loss, and recovery factor, which MetaTrader derives from that larger drawdown. Both differences are intentional and are shown rather than hidden. Nothing here can tell you whether the strategy was fitted to the data — for that, run the same settings on symbols and periods it was not built on.
Nothing is uploaded. Every number on this page is computed in your browser, and the link you copy carries only the values you typed.
Related tools
A backtest is the best case that a set of assumptions can produce. The forward result is what happens when the assumptions meet a broker.