How to read the MT5 Strategy Tester report
An MT5 backtest report has dozens of figures. Almost everything important is in a handful of them. What each one means, what values are reasonable and which ones should make you suspicious.
Net profit and profit factor
Net profit is everything won minus everything lost. Profit factor divides gross profit by gross loss: 1 means it makes nothing, 1.5 means it wins 1.5 for every unit it loses. Around 1.3–1.5 is often considered reasonable; values of 3 or more over many trades are rare and deserve suspicion, especially when they come out of an optimization.
Drawdown: absolute, maximal and relative
MT5 gives several versions. Absolute is how far it fell below the initial deposit. Maximal is the largest drop from a peak to the next low, in money; relative is that drop as a percentage. Each one appears for balance (closed trades) and for equity (including open ones). Look at equity: balance can hide trades that went deep against you before closing. Ask yourself whether you would sit through that drop with real money.
Recovery factor
Net profit divided by maximal drawdown. It sums up in one figure whether the profit made up for the risk you had to sit through: at 1, you earned as much as you lost at the worst moment. Higher is better; compare it between configurations tested over the same period, because it grows with the length of the test.
Number of trades
It is the most ignored figure and the one that decides whether the others mean anything. With 30 trades, a couple of lucky wins change the profit factor completely; with 300, much less. There is no universal minimum, but below about 100 over the whole period it pays to be very cautious, even more so if the optimization tried thousands of combinations.
Expected payoff and Sharpe ratio
Expected payoff is the average profit per trade. Compare it with your cost per trade (spread, commission, slippage): if it is only a few points, a slightly worse broker eats it. The Sharpe ratio measures return relative to how irregular it is; in the tester, below 0 means it loses and above 1 means the risk pays off. Use it to compare configurations with each other rather than as an absolute grade.
Traps when reading the report
- Looking only at net profit. Without drawdown and number of trades, it says nothing about risk or about whether it is chance.
- Comparing reports with different periods or deposits: profit and drawdown in money depend on both.
- Trusting a backtest with unrealistic modelling. "Every tick based on real ticks" is the most faithful; "Open prices only" only makes sense if the EA acts at the open of the bar.
- Reading the report of the best pass of an optimization as if it were a clean test: that pass was chosen precisely because it did well in that period.
Next step
A single report cannot tell you whether a configuration is robust or the lucky one among thousands; for that, Orometra analyzes the whole optimization. And when you test the chosen one on a new period, drop that report into Orometra: it tells you whether the result is normal for your EA, with a Monte Carlo over your own trades.