System Performance: Sharpe, Sortino, Calmar, Profit Factor
Performance metrics like Sharpe, Sortino, Calmar, and profit factor summarize a trading system's risk-adjusted returns, and this guide explains each metric with formulas and interpretation for beginners.
Interaktive Tools funktionieren in der übersetzten Ansicht möglicherweise nicht.
System Performance: Sharpe, Sortino, Calmar, Profit Factor
Two systems can both return 20% a year. The metrics below tell you which one took sane risk to get there.
Why metrics matter
Return alone is meaningless. A 30% return earned with 50% drawdowns is psychologically untradeable and mathematically inferior to a 25% return earned with 8% drawdowns. Risk-adjusted metrics let you compare systems honestly and pick the one most likely to survive.
1. Profit factor
The simplest and most intuitive metric.
Profit Factor = Gross Profit / Gross Loss
| Value | Interpretation |
|---|---|
| < 1.0 | Losing system |
| 1.0–1.2 | Marginal |
| 1.2–1.5 | Decent |
| 1.5–2.0 | Strong |
| > 2.0 | Exceptional (or overfit) |
Profit factor is intuitive but ignores timing — it doesn't care whether the profits came in steady increments or one lucky trade.
2. Sharpe ratio
Measures return per unit of total volatility.
Sharpe = (Mean Return − Risk-Free Rate) / Standard Deviation of Returns
For intraday/daily systems, the risk-free rate is often set to 0.
| Annualized Sharpe | Interpretation |
|---|---|
| < 0.5 | Poor |
| 0.5–1.0 | Acceptable |
| 1.0–2.0 | Good |
| 2.0–3.0 | Excellent |
| > 3.0 | Suspicious (overfit or arb) |
Limitations: Sharpe penalizes upside volatility equally with downside. A system with occasional large gains looks worse than it should.
3. Sortino ratio
Like Sharpe, but only penalizes downside volatility.
Sortino = (Mean Return − Risk-Free Rate) / Downside Deviation
Downside deviation considers only returns below a target (often 0 or MAR).
A trend-following system might have:
- Sharpe = 1.1 (penalized for upside bursts)
- Sortino = 1.8 (only downside matters)
For asymmetric strategies, Sortino is more honest than Sharpe. A Sortino above 2.0 is solid; above 3.0 is exceptional.
4. Calmar ratio
Measures return per unit of worst drawdown.
Calmar = Annualized Return / Maximum Drawdown
| Calmar | Interpretation |
|---|---|
| < 0.5 | Poor |
| 0.5–1.0 | Acceptable |
| 1.0–3.0 | Good |
| > 3.0 | Exceptional |
Calmar answers: "How much pain did I endure for each unit of return?" A system with 30% annual return and 10% max drawdown has Calmar = 3 — strong.
5. Maximum drawdown
Not a ratio but a critical absolute number. The largest peak-to-trough decline in the equity curve.
- Express in both % of equity and R multiples
- The psychological limit for most traders is 20–25%
- Above 30% most traders abandon the system mid-drawdown
6. Recovery factor
Recovery Factor = Net Profit / Maximum Drawdown
Similar in spirit to Calmar, but uses net profit instead of annualized return. Above 5 is excellent.
7. Expectancy per trade
Expectancy = (Win% × Avg Win) − (Loss% × Avg Loss)
Measured in R-multiples. Anything above 0.2R per trade is solid for a discretionary system; 0.1R is acceptable for high-frequency.
Which metrics to use together
No single metric tells the full story. A minimal set:
- Profit factor — profitability in raw terms
- Sharpe — risk-adjusted return, penalizes volatility
- Sortino — risk-adjusted return, only penalizes downside
- Calmar — return per unit of worst drawdown
- Maximum drawdown — psychological survivability
- Expectancy — quality per trade
A system is "good" if it scores well on most of these simultaneously.
Worked comparison
| Metric | System A | System B |
|---|---|---|
| Annual return | 30% | 30% |
| Max drawdown | 25% | 8% |
| Profit factor | 1.4 | 1.6 |
| Sharpe | 1.2 | 1.9 |
| Sortino | 1.5 | 2.4 |
| Calmar | 1.2 | 3.75 |
Same headline return, very different systems. System B is the clear winner — more profit per unit of risk taken and a drawdown most traders can survive.
Common pitfalls in metric use
- Annualizing short samples: a 1-week Sharpe of 5.0 means nothing — extrapolation lies
- Ignoring sample size: 30 trades can't produce stable metrics
- Reporting only the best metric: choose the metric before testing, not after
- Comparing across frequencies: a daily system's Sharpe can't be directly compared to a monthly one without annualization
- Using backtested drawdown as the live expectation: live drawdown is typically 1.5–2× backtested
Practical thresholds for retail systems
A retail-discretionary or simple systematic system is "good" when:
- Profit factor ≥ 1.3
- Sharpe ≥ 1.0
- Sortino ≥ 1.5
- Calmar ≥ 1.0
- Max drawdown ≤ 20%
- Expectancy ≥ 0.15R per trade
Above all these thresholds simultaneously is rare; below all is unacceptable.
Next: zoom in on the most important absolute number — maximum drawdown and recovery.
Live Chart
Open full chart →Related market data, powered by TradingView.