At depth zero Expected Shortfall is the worst case

At depth $0$ the sets competing for the worst total are exactly the singletons, so the worst total is the single largest loss and Expected Shortfall coincides with the worst-case measure. Since the divisor is the tail size $m+1$, at $m=0$ it is $1$ and no averaging occurs.

This identification is asserted in the docstring of the ExpectedShortfall definition but is not proved there, so it is recorded here as a theorem. It also connects the two measures: worstCase was separately shown to be coherent, and this exhibits it as the depth-zero member of the Expected Shortfall family.

As throughout CoherentRisk, the development counts states and refers to no probability measure, so $m$ is an integer depth rather than a confidence level.