Testing & evaluation
How to Read an EA Backtest Report Beyond Net Profit
Read an EA backtest by checking the test setup, drawdown, trade distribution and costs before focusing on net profit. A high win rate or profit factor alone cannot tell you whether the strategy is robust or appropriate for live use.
Start above the performance numbers
Confirm the program version, settings, dates, broker symbol, deposit and modeling assumptions. A screenshot without that context is hard to reproduce. When comparing two runs, write down everything that changed before attributing the difference to a code improvement.
For example, a larger final balance might simply reflect a larger starting deposit or larger positions. Compare like with like and preserve the complete report, including unfavorable periods.
Four measures to read together
The MetaTrader report reference defines the platform's metrics. Profit factor compares gross profits with the magnitude of gross losses. Drawdown describes declines from earlier values; balance and equity drawdowns describe different account paths. Trade counts provide context for all of them.
| Figure | Ask next |
|---|---|
| Net profit | Over which period, with which exposure and costs? |
| Profit factor | How many losses occurred, and how concentrated were gains? |
| Equity drawdown | How deep and how prolonged were floating losses? |
| Win rate | How large were losses compared with wins? |
Why an 80% win rate can still lose money
Consider an illustrative set of ten trades: eight gains of $10 and two losses of $50. The win rate is 80%, but the total is $80 − $100 = −$20 before additional costs. Profit factor is $80 ÷ $100 = 0.8.
Now change the two losses to $20 each. The win rate stays at 80%, but the total becomes $40. This arithmetic is not a backtest or a forecast; it shows why counting winners alone leaves out essential information.
Read the path, not just the ending
In another hypothetical account, equity rises to $1,200 and then falls to $900 before recovering. That particular decline is $300, or 25% of the preceding peak. Ending above the starting balance would not erase the intervening loss.
A balance curve can look calm while open positions carry large floating losses. Inspect equity as well, and check the duration of the recovery. Also keep drawdown amounts and percentages distinct: the largest money decline and largest percentage decline need not occur at the same point.
Look for concentration and missing evidence
Check whether a small number of trades or one short period supplies most of the gains. Look at losing streaks and changes in position size. If the results depend on increasing exposure after losses, understand the resulting worst-case exposure rather than relying on the average trade.
Finish your review with three notes: what looks reproducible, which assumptions are uncertain, and which next test could challenge the result. Use an untouched test period to investigate generalization. No report metric turns past performance into a guarantee.
Back to top