Traces a single classifier's ROC curve with every point tagged by its decision threshold, showing how lowering the cutoff trades a higher true-positive rate for more false positives. Filled markers and data labels make each operating point selectable by name.