Historical expectations

Daily returns

Distribution, each as a share of its own sample

Quantile–quantile, live against backtest

Returns by horizon

Skew by horizon

Excess kurtosis by horizon

Calendar years

Value at Risk and expected shortfall

VaR as the confidence level moves

Expected shortfall as the confidence level moves

Normal reads the quantile off a Gaussian curve. Cornish–Fisher corrects that quantile for the observed skew and kurtosis. Historical sorts the realised returns.

Breaches in the live window

How does it look against benchmarks?

Ex-post vol normalized 10%

Drawdown

Backtest drawdown depth, share of sessions at each depth

The six deepest, dated