Aggregating Cases

Two doctors. In every single department — cardiology, oncology, emergency — Doctor A has a higher survival rate than Doctor B. You promote Doctor A. Then the data splits. Doctor B treated ten times more critical patients in every department. Overall, Doctor B's survival rate is higher. You had all the numbers. You read them correctly. And you were completely wrong. This is Simpson's Paradox. It has reversed medical trial conclusions, decided court cases, and misled hiring decisions at major universities. It has never been a game. WITNESS A scenario loads. Two entities, one aggregate statistic each. You decide who performs better. Then the bar splits into subgroups — and the conclusion swings the other way. You watch it happen. Then again. Then again. By round five you stop trusting aggregate numbers. BUILD The subgroup rates are fixed. You control the volumes. Drag sliders until the…