⚖️ Fair Scoring Lab: Calibration vs Equalized Odds vs the 80% Rule
Two groups, one score and a threshold: audit selection rates, TPR/FPR, precision and calibration with closed-form maths checked by Monte Carlo.
⚙ Under the hood
Two groups, one score, one threshold: audit selection rate, TPR/FPR, precision, calibration and the four-fifths rule with closed-form Gaussian maths, one-click threshold strategies and Monte Carlo checks.
ML FairnessAlgorithmic BiasEqualized OddsCalibrationROC
2D · HTML5 Canvas 2D · 60 FPS target · runs fully client-side, no install