Outcome distributions across randomized conditions
Each scenario was re-simulated many times under randomized exogenous conditions — an external shock, mediator pressure, overall mood, and which delegation enters with momentum — to see how robust each outcome is.
How to read this. These are Monte Carlo samples of an AI model's distribution over plausible outcomes, not forecasts. Each trial randomizes the conditions and simulates one realization; we aggregate. Wide spread = outcome is condition-sensitive; tight spread = robust. Satisfaction is the model's 0–1 rating per delegation. Exploratory and educational only — see the methodology.