A precision tool for measuring whether your classifier is actually better than guessing.

Most ML dashboards chase complexity with endless ROC curves and glossy metrics that obscure what actually matters. The Accidental Prediction Rate Calculator strips the noise down to a single, confrontational question: is your model genuinely better than chance? Its compact layout pairs Cohen's kappa, proportional baselines, and binomial significance into one screen that a data scientist can read in seconds. Anyone who's been burned by a model that scored 83% accuracy on a heavily imbalanced set will instantly understand why this tool exists.