Performance Metrics
Raw returns only tell part of the story. A strategy with 50% return and violent swings can be harder to trade than one with 30% return and a smooth equity curve. These metrics give you a complete picture.
Cosa Imparerai
- ▸Why risk-adjusted metrics matter more than raw return
- ▸How each metric is calculated and what drives it
- ▸Practical benchmarks for evaluating strategy quality
- ▸How to compare two strategies side-by-side using metrics
- ▸The Win Rate definition used by the backtesting engine
Why metrics beat raw returns#
Two strategies both return 60% over a year. Strategy A does it with a 45% max drawdown and wild daily swings. Strategy B does it with a 15% max drawdown and a steady upward curve. Most traders would lose money on Strategy A because they'd exit during the painful drawdown — before the recovery. The metrics below help you identify which strategy you can actually hold through rough patches.
Return metrics#
Total Return (%)
Formula: (Final Equity − Initial Equity) / Initial Equity × 100
The overall percentage gain or loss over the entire backtest period. The most basic measure of profitability.
Excellent: > 50%
Good: 20–50%
Needs work: < 20%
Total return depends heavily on period length. A 30% return over 2 years is weaker than a 30% return over 6 months. Use Annualized Return to compare across different periods.
Annualized Return (%)
Formula: (1 + Total Return) ^ (1 / Years) − 1
Total return normalized to a one-year horizon. Allows fair comparison between strategies tested on different time ranges.
Example: A strategy with 20% return over 6 months has roughly 44% annualized return. A strategy with 30% return over 12 months has 30% annualized. The 6-month strategy wins on an annualized basis.
Risk metrics#
Annualized Volatility (%)
Formula: StdDev(Returns per candle) × √(periods per year)
Measures how much returns fluctuate period-to-period. Higher volatility means less predictable outcomes and more psychological stress. The engine infers the correct annualization factor from the actual candle frequency — a 1h backtest uses 8766 periods per year, not 252.
Low: < 20% — stable, predictable
Medium: 20–40% — moderate swings
High: > 40% — wild, stressful
Max Drawdown (%)
Formula: (Peak − Trough) / Peak × 100
The largest peak-to-trough decline in portfolio value. Represents the worst-case loss an investor would have experienced at any point in the backtest. This is the metric most traders feel in their stomach.
Excellent: < 15% — very safe
Acceptable: 15–30% — manageable
Dangerous: > 30% — hard to hold
Why it matters: A 50% drawdown requires a 100% recovery to break even. Psychologically, most traders abandon strategies well before that — typically around 30–40%. This is why controlling drawdown matters as much as generating returns.
Risk-adjusted performance#
Sharpe Ratio
Formula: (Annualized Return − Risk-Free Rate) / Annualized Volatility
The most widely used risk-adjusted metric. Measures how much excess return you earn per unit of total risk. Higher is better — it means more return for the same volatility.
< 0 — losing money
0–1 — below expectations
1–2 — good
> 2 — excellent
Context: The S&P 500 historically has a Sharpe around 0.5–0.8. Hedge funds target above 1.0. A backtest Sharpe above 2.0 is exceptional — but verify it isn't the result of over-fitting.
Sortino Ratio
Formula: (Annualized Return − Risk-Free Rate) / Downside Deviation
Similar to Sharpe but only penalizes downside volatility — returns below zero. Upside volatility (unexpectedly good days) doesn't count against you. This makes Sortino more accurate for strategies with asymmetric return distributions.
When to prioritize: Use Sortino when comparing strategies with different upside/downside profiles. A strategy that spikes up occasionally but rarely falls will have a better Sortino than Sharpe.
Calmar Ratio
Formula: Annualized Return / |Max Drawdown|
Compares annual returns to the worst-case drawdown experienced. A Calmar of 2.0 means you earn 2x your worst loss per year — your pain is justified by the gains.
< 1.0 — weak
1.0–3.0 — good
> 3.0 — excellent
Trading activity#
Win Rate (%)
Formula: Profitable Closed Trades / Total Closed Trades × 100
The percentage of round-trip trades (a BUY entry + SELL exit pair) that closed with positive P&L. A **closed trade** is a SELL order that exits a previously opened long position.
Note: Win Rate measures per-trade profitability, not per-day. A strategy that trades once per week and wins 6 out of 10 trades has a 60% Win Rate regardless of how many days those trades spanned.
Interpretation: 50–60% is typical for solid strategies. Win rates above 70% can indicate over-fitting. Rates below 45% suggest the strategy loses more often than it wins — this can still be profitable if winning trades are significantly larger than losing ones.
Number of Days
Calendar length of the backtest period. More days means more statistically significant results.
- Minimum: Minimum: 60 days (~2 months) for initial testing
- Recommended: Recommended: 180+ days (6 months) for reliable metrics
- Ideal: Ideal: 365+ days (1 year) to capture different market conditions including volatile and quiet periods
Comparing strategies#
Use metrics to compare strategies objectively. Total Return alone misleads — a strategy with lower returns but better risk-adjusted metrics is often the right choice for live trading.
Example comparison
| Metric | Strategy A | Strategy B | Winner |
|---|---|---|---|
| Total Return | 87% | 52% | Strategy A |
| Max Drawdown | -38% | -18% | Strategy B |
| Sharpe Ratio | 1.2 | 1.8 | Strategy B |
| Volatility | 42% | 24% | Strategy B |
| Overall | Strategy B |
Strategy A has higher total returns, but Strategy B wins on every risk-adjusted metric. Most traders would earn more with Strategy B because they could hold it through drawdowns without panicking.
Quick reference#
For returns
- Total Return
- Annualized Return
For risk
- Max Drawdown
- Annualized Volatility
For risk-adjusted quality
- Sharpe Ratio (most important)
- Sortino Ratio
- Calmar Ratio
For consistency
- Win Rate
- Equity curve smoothness
Target: A well-rounded strategy has Sharpe above 1.5, max drawdown below 25%, and a smooth equity curve. No single metric tells the whole story — balance them.