sigmaquant.performance.metrics.calmar_ratio#

sigmaquant.performance.metrics.calmar_ratio(returns, frequency, kind='simple')#

Compute the Calmar ratio.

Parameters:

  • returns (ArrayLike) –

    Input time series:

    • kind="simple": simple returns

    • kind="log": log-returns

    • kind="pnl": additive PnL

  • frequency (Frequency) –

    Sampling frequency of the input series:

    • "D": daily

    • "W": weekly

    • "M": monthly

    • "Y": yearly

  • kind (str, default="simple") – Type of input values.

Returns:

Calmar ratio.

Return type:

float

Notes

The Calmar ratio is defined as:

\[\text{Calmar} = \frac{\mathrm{CAGR}}{\left| \mathrm{Max Drawdown} \right|}\]

where:

  • \(\mathrm{CAGR}\) is the compound annual growth rate.

  • \(\mathrm{Max Drawdown}\) is the maximum peak-to-trough drawdown over the sample period.

If kind="pnl", a Calmar-like metric is computed:

\[\frac{\text{Average Annual Return}} {\left| \text{Monetary Max Drawdown} \right|}\]

In this case the numerator is the annualized mean PnL and the denominator is the absolute monetary maximum drawdown.