Risk Metrics¶

Comprehensive risk analysis metrics for strategy evaluation.

Overview¶

Risk metrics quantify the risk characteristics of trading strategies, helping traders understand downside exposure, volatility, and tail risks.

Core Risk Metrics¶

Volatility¶

Annualized standard deviation of returns:

from rustybt.analytics import RiskAnalytics

risk = RiskAnalytics(backtest_result)
metrics = risk.calculate_risk_metrics()

volatility = metrics['volatility']  # Annualized
print(f"Volatility: {volatility:.2%}")

Interpretation: - <10%: Low volatility - 10-20%: Moderate volatility - >20%: High volatility

Sharpe Ratio¶

Risk-adjusted return (excess return / volatility):

sharpe = metrics['sharpe_ratio']
print(f"Sharpe Ratio: {sharpe:.2f}")

Interpretation: - <1.0: Poor risk-adjusted returns - 1.0-2.0: Good - 2.0-3.0: Excellent - >3.0: Exceptional (verify not overfitted)

Sortino Ratio¶

Downside risk-adjusted return:

sortino = metrics['sortino_ratio']
print(f"Sortino Ratio: {sortino:.2f}")

Advantages over Sharpe: - Only penalizes downside volatility - Better for asymmetric return distributions

Maximum Drawdown¶

Largest peak-to-trough decline:

max_dd = metrics['max_drawdown']
print(f"Max Drawdown: {max_dd:.2%}")

Interpretation: - <10%: Excellent - 10-20%: Good - 20-30%: Acceptable - >30%: High risk

Value at Risk (VaR)¶

Potential loss at confidence level:

var_95 = risk.calculate_var(confidence_level=0.95)
print(f"VaR (95%): {var_95:.2%}")

Interpretation: Expected to lose more than VaR only 5% of the time.

Conditional VaR (CVaR)¶

Expected loss when VaR is exceeded:

cvar_95 = risk.calculate_cvar(confidence_level=0.95)
print(f"CVaR (95%): {cvar_95:.2%}")

Use case: Tail risk assessment.

Complete Example¶

from rustybt.analytics import RiskAnalytics
from rustybt.utils.run_algo import run_algorithm
import pandas as pd

# Run backtest
result = run_algorithm(
    algorithm_class=MyStrategy,
    start=pd.Timestamp('2020-01-01'),
    end=pd.Timestamp('2023-12-31'),
    capital_base=100000
)

# Create risk analytics
risk = RiskAnalytics(
    backtest_result=result,
    benchmark_returns=spy_returns,  # Optional benchmark
    confidence_level=0.95
)

# Calculate all metrics
metrics = risk.calculate_risk_metrics()

# Print comprehensive risk report
print("=== Risk Metrics ===\n")

print(f"Volatility: {metrics['volatility']:.2%}")
print(f"Downside Deviation: {metrics['downside_deviation']:.2%}")
print(f"\nRisk-Adjusted Returns:")
print(f"Sharpe Ratio: {metrics['sharpe_ratio']:.2f}")
print(f"Sortino Ratio: {metrics['sortino_ratio']:.2f}")
print(f"Calmar Ratio: {metrics['calmar_ratio']:.2f}")

print(f"\nDrawdown Analysis:")
print(f"Max Drawdown: {metrics['max_drawdown']:.2%}")
print(f"Max DD Duration: {metrics['max_drawdown_duration']} days")

print(f"\nTail Risk:")
print(f"VaR (95%): {metrics['value_at_risk']:.2%}")
print(f"CVaR (95%): {metrics['conditional_var']:.2%}")

if 'beta' in metrics:
    print(f"\nMarket Exposure:")
    print(f"Beta: {metrics['beta']:.2f}")
    print(f"Correlation: {metrics['correlation']:.2f}")