01Compose a scenario
Choose stressors and when they land. Combination effects are real interactions, not sums: they fire only when stressors overlap in time, so timing is part of the scenario. Three or more stressors make the interaction table meaningful.
02Simulate
600 deterministic trials over 18 months. Same seed and inputs always give the same result. After the first run, every change to stressors, mitigations, metric or assumptions recomputes automatically.
Contested assumptions — change them and argue about the result
No simulation yet. Results will show distributions for customer harm, adoption, control load, financial outcome and time-to-value, with the unmitigated version of the same scenario as a dashed outline.
03Failure-path replay
Run the simulation to replay the most representative failure path: the modal sequence of failure mechanisms among adverse trials, shown through its median member. Arrow keys step through it.
04Mitigation portfolio
Spend a $1.8M reserve and three change units. Every option has an owner, a delay before it works, and side effects. Some create new risks or slow value — the leverage figures show the net effect of each option given everything else you've chosen.
05Metric corruption lab
The steering committee's primary KPI is an input to the model: it adds behavioural pressure on specific failure mechanisms. Switch it and watch what the dashboard reports versus what is happening.
06Adversary lenses
Six critics with different incentives read the same run. Each cites different evidence, wants something different, and is wrong in a characteristic way.
07Pre-mortem & dissent
08Decision package
Downloads a local JSON file with scenario, assumptions, mitigations, budget and capacity use, outcome ranges, representative path, dissent, metric choice and remaining unknowns. Nothing leaves this page.
Preview package
Export to generate.