World trade is more concentrated than the map suggests. A small set of passages, Suez, Panama, Hormuz, Malacca, Gibraltar, Bab el-Mandeb, carries a share of global flow wildly disproportionate to their number, and the last few years have demonstrated, repeatedly, that each of them can be impaired with little warning: a grounding, a drought, a security crisis, a policy decision. Chokepoint risk has migrated from tail scenario to operating assumption.
What separates the networks that absorb these events from the networks that are absorbed by them is not forecasting. Nobody reliably predicts the date a strait closes. The separation is in preparation of the response surface: knowing, in advance and quantitatively, what the alternatives are, what they cost, how long they take to activate, and which customers, contracts, and inventories are exposed at each hour of a disruption's life.
Rehearsal as infrastructure
This is what a stochastic network model is for. Not predicting the crisis, but having already simulated it a thousand times: rerouting via the cape versus waiting out the closure, air substitution for the flows that justify it, drawdown sequencing for buffer stocks, repricing for the capacity that suddenly matters. When the event arrives, the organization is not solving a novel problem under pressure. It is executing a policy it has already stress-tested, updating it with live data as the situation reveals itself. Hours matter in these windows, and rehearsed hours are cheaper than improvised ones by an order of magnitude.
The map reroutes itself, at a price.
Close a chokepoint and watch the network answer.
Chokepoints are priced at zero until the day they are priced at everything. Adaptive networks pay the option premium in advance.
The adaptive ideal
The end state worth building toward is a network with reflexes: continuous monitoring of the leading indicators around each chokepoint, standing contingency plans re-optimized as conditions drift, and pre-negotiated optionality, capacity, routings, contracts, that converts crisis response from procurement into activation. Disruption then stops being an existential event and becomes a priced, managed operating condition. The chokepoints are not going away. The scramble can.