Level 03Advanced

Metrics That Actually Matter

Net profit is the least informative number on a backtest report.

2 min readAdvanced

Net profit is the first number everyone looks at and the least informative one on the report. It tells you nothing about how much risk was taken to get it, or whether the result could have happened by chance.

A handful of other metrics tell you far more, and reading them together is what separates a serious evaluation from a screenshot.

Edge: expectancy and profit factor

Expectancy is the average currency amount made per trade. Profit factor is gross profit divided by gross loss. Both describe the same underlying thing from different angles, and both should be calculated net of costs.

A profit factor above 1.0 is profitable, but 1.2 is fragile once costs and slippage move slightly. Anything above 2.0 on a decent sample deserves scepticism before celebration — it is often a sign of a short sample or an unlucky-free period.

Pain: drawdown and its duration

Maximum drawdown is the largest peak-to-trough decline in equity. Its duration is how long the strategy spent underwater. Both matter, and duration is the more emotionally loaded of the two.

Combine them with return to get a sense of whether the strategy is tradeable by a human being. A strategy returning 40% a year with a fourteen-month drawdown is technically excellent and practically untradeable for most people.

Risk-adjusted return

The Sharpe ratio divides excess return by the standard deviation of returns. It rewards consistency and penalises volatility. Sortino refines this by penalising only downside deviation, which most traders find more meaningful.

These are comparative tools, not absolute targets. Use them to compare strategies or parameter sets against each other, not to decide whether something is good in isolation.

VaR Calculator

Calculate Value at Risk to understand potential portfolio drawdowns in extreme scenarios.

Tool

Sample size is the gate

Every metric above is meaningless below a certain number of trades. Thirty trades is anecdote. Two hundred is a beginning. Below that threshold, your metrics are dominated by variance and will not reproduce.

This is the single most common reason a trader's live results diverge from their test. The test looked excellent because it contained forty trades, one of which was enormous. Nothing was wrong except the sample.

Worked example

What to look at on a backtest report, and in what order

MetricWhat it tells youWarning sign
Trade countWhether anything else is meaningfulUnder 100 trades
ExpectancyAverage profit per trade after costsNegative or near zero
Profit factorGross profit relative to gross lossBelow 1.3
Max drawdownThe pain you must sit throughLarger than you can tolerate
Drawdown durationHow long you sit through itMonths longer than your patience
Sharpe / SortinoReturn per unit of riskCompared across strategies only

Key takeaways

  • Net profit without a risk measure tells you almost nothing
  • Expectancy and profit factor describe the edge; drawdown describes the cost of holding it
  • Sharpe and Sortino are for comparing strategies, not for absolute judgement
  • Below roughly 100 trades, no metric on the report is statistically meaningful
At a glance
  • Profit factor and expectancy describe the edge
  • Maximum drawdown and its duration describe the pain
  • Sharpe and Sortino adjust return for the risk taken
  • Trade count determines whether any of it is statistically meaningful
Common mistake

Celebrating an excellent profit factor built on fewer than thirty trades.

Try it on Strategist

Reading is not verification

Take the concept you just read and turn it into explicit rules, then test it. That is the only way to know whether it actually works.

Build a strategy