The Doom Loop Engine
Three interactive instruments for thinking about how an oil-supply shock could compound. They are reasoning scaffolds, not forecasts: every probability, edge and buffer figure is a subjective input you can change. The tools propagate your assumptions — they do not validate them. Their value is that a sceptic can sit down, move the sliders, and watch the catastrophe either build or dissolve under their own hand.
Three tools, three questions
The three instruments answer different questions and can legitimately disagree. The Fragility Monitor ranks the individual failure modes by expected harm — what to worry about most right now. The Doom Loop Engine lets those failure modes interact and draw down a finite buffer — whether they compound into allocation, and when. The Vulnerability Tiering ranks which countries a sustained shock pushes toward collapse — where, and in what order. A risk can top one and barely move another; that is the difference between a big standalone threat and a big systemic one.
1 · Systemic Fragility Monitor
A node-level risk register. Each failure mode scores P × Impact, re-priced by the ceasefire master-variable and ranked by expected harm. The on-ramp: scannable in seconds.
2 · The Doom Loop Engine
The dynamic model. The same failure modes become continuous stress levels that interact through a feedback graph and draw down the drawable cushion above the allocation floor. It runs thousands of trials on every change and reports the probability of allocation, the time-to-allocation, and — via the sensitivity sweep — which assumptions actually move the answer.
Two numbers do the heavy lifting
It is tempting to think a compound-cascade model lives in its web of interactions — eight failure modes, thirteen feedback edges, a master variable for the war. Run the sensitivity sweep in the engine above, though, and something humbling falls out: almost the entire spread in the probability of allocation comes from just two numbers — how fast a stressed system draws down the global buffer, and how large that buffer is.
Move any single failure mode's likelihood, or any edge between them, by half in either direction and the headline barely twitches. Move the draw rate or the cushion by the same amount and it swings by fifty points.
That isn't the cascade being irrelevant — it's the cascade telling you where its own fragility lives. The nodes and edges are the engine that creates the draw in the first place; the cushion and the draw rate are the gate the outcome is most sensitive to. And those two are the two we can actually measure: the rate the world is drawing on its spare barrels, and how many spare barrels remain. The vulnerability was never in the wiring — it's in the cushion. Pin those two numbers down with data, and stop agonising over edge weights you can't observe — until the scenario worsens, the network lights up, and the wiring starts to bite again.
The vulnerability was never in the wiring — it's in the cushion.
3 · Vulnerability Tiering
Where, and in what order. The same shock produces collapse in one country and a survivable recession in its neighbour — decided by three stacked multipliers: oil-import exposure, cereal-import dependence (the diesel→food channel), and FX reserves (coping), combined conjunctively and ranked. Reserves are a live World Bank pull; cereal and oil are sourced seeds. A susceptibility map, not a forecast.
The framework behind these tools: Compound Cascade methodology →