I have participated in a few and done a bunch of modeling/simulation things that resemble scenario exercises structurally in but I haven’t actually designed/run scenario exercises myself. I’m more interested in the math/logic foundations of counterfactual/modal methods
Which makes sense because variational/perturbation methods are typically used for tractability. A high-dimensional and intractable primary problem can yield a tractable low-dimensional variational problem with almost no loss if the nominal reference trajectory (scenario) is good