Skip to main content

Metrics

Quantalytics ships a curated set of risk/return statistics. All metric functions accept iterables or pandas Series and return floats. The performance_summary helper collects the most common measures into a single dataclass.

Performance Summary​

import quantalytics as qa

summary = qa.metrics.performance_summary(returns, risk_free_rate=0.02)
summary.as_dict()

Fields available on PerformanceMetrics:

  • annualized_return
  • annualized_volatility
  • sharpe
  • sortino
  • calmar
  • max_drawdown
  • cumulative_return

Individual Metrics​

import quantalytics as qa

sharpe = qa.metrics.sharpe(returns, risk_free_rate=0.015)
sortino = qa.metrics.sortino(returns, risk_free_rate=0.015, target_return=0.0)
calmar = qa.metrics.calmar(returns)
drawdown = qa.metrics.max_drawdown(returns)
romad = qa.metrics.romad(returns)
omega = qa.metrics.omega(returns)
max_dd_pct = qa.metrics.max_drawdown_percent(returns)
longest_dd_days = qa.metrics.longest_drawdown_days(returns)
underwater = qa.metrics.underwater_percent(returns)
avg_dd = qa.metrics.average_drawdown(returns)
avg_dd_days = qa.metrics.average_drawdown_days(returns)
recovery = qa.metrics.recovery_factor(returns)
ulcer = qa.metrics.ulcer_index(returns)
serenity = qa.metrics.serenity_index(returns)

All metrics support customizing the periods_per_year argument to align with daily, weekly, or monthly data.

Tail Risk Metrics​

var = qa.analytics.value_at_risk(returns, confidence=0.95)
cvar = qa.analytics.conditional_value_at_risk(returns, confidence=0.95)
  • value_at_risk returns the historical loss magnitude you should not exceed with the specified confidence.
  • conditional_value_at_risk averages all losses beyond that VaR threshold (expected shortfall).

Advanced Risk Metrics​

prob = qa.analytics.prob_sharpe(returns, risk_free_rate=0.0, target_sharpe=0.5)
smart_sharpe = qa.analytics.smart_sharpe(returns)
smart_sortino = qa.analytics.smart_sortino(returns)
smart_sortino_half = qa.analytics.smart_sortino_over_sqrt_two(returns)
  • prob_sharpe_ratio returns the probability that your sample Sharpe exceeds a target.
  • smart_sharpe_ratio and smart_sortino_ratio adjust the classical measures for skew/kurtosis.
  • /sqrt(2) variants rescale Sortino-style ratios to account for half-normal downside distributions.
  • omega compares upside vs downside partial moments around a target threshold.
  • Use max_drawdown_percent, longest_drawdown_days, and underwater_percent to translate drawdown paths into presentation-ready KPIs.
  • average_drawdown, average_drawdown_days, and recovery_factor capture typical drawdown depth/duration and how quickly returns reclaim new highs.
  • ulcer_index and serenity_index summarize drawdown variability and annualized-growth efficiency in a single score.