Return-based risk computed in the open analytics core (quantlib.risk) from 675 daily returns (trailing 3y window), annualized at 252/yr. Click any metric for its methodology.
Signed daily quantiles (a 5% loss is −5%); CVaR ≤ VaR ≤ 0. The Cornish-Fisher row adjusts the Gaussian VaR for skewness and excess kurtosis (fat tails), so it has no historical counterpart.
Notes
skewness: skewness is undefined when the variance (second central moment m2) is zero
kurtosis: kurtosis is undefined when the variance (second central moment m2) is zero
benchmark_correlation: correlation_to_benchmark is undefined: a series has zero variance (flat/degenerate denominator)
sharpe_ratio: sharpe_ratio is undefined when the excess-return std is zero
sortino_ratio: sortino_ratio is undefined when the downside deviation is zero
var_cornish_fisher_95: insufficient data for a Cornish-Fisher estimate
var_cornish_fisher_99: insufficient data for a Cornish-Fisher estimate
omega_ratio: omega_ratio is undefined when the downside area (loss denominator) is zero
tail_ratio: tail_ratio is undefined when the lower-quantile value (denominator) is zero
gain_to_pain_ratio: gain_to_pain_ratio is undefined when there are no losing periods (zero pain)
win_loss_ratio: win_loss_ratio is undefined without at least one win and one loss
upside_potential_ratio: upside_potential_ratio is undefined when the downside deviation is zero