Case Study

Cascade Simulator — SupplyChainRisk-AllRungs.bayes

Set a shock on any upstream node and watch the disruption propagate layer by layer through the supply chain graph. Compares prior (baseline) vs shocked posterior at each node.

Why cascade simulation matters

A single-shot intervention answers one question: what does pushing this lever do, holding everything else equal? Real systems don't behave that way. An incident in a cyber control triggers a physical safety event, which triggers an economic disruption, which triggers a regulatory response — each step changes the probability landscape for the next. Treating each link as independent and then summing the consequences understates the joint probability of the whole chain by a margin that grows with the number of steps.

The cascade simulator runs the full causal chain end-to-end. It propagates evidence forward through every dependent variable, accumulates the joint distribution over the terminal nodes, and shows where the intervention's effect amplifies, attenuates, or reverses along the path. The output is a probability distribution over the entire downstream consequence, not a point estimate at a single node — which is the only basis on which mitigation investment can be sized correctly.

increase from prior
decrease from prior
unchanged
bar fill = posterior value as % of scale (0–100)