What is the typical range for AUC on an ROC curve?

What is the typical range for AUC on an ROC curve?

• A. 0.5 to 1.0
• B. 0 to 1.0
• C. -1 to 1
• D. 0 to 0.5

Answer: (A)

Discrimination across all thresholds is what the AUC captures: how well the model ranks positives above negatives. The AUC value always falls between 0 and 1. A value of 0.5 means no discriminative ability (like random guessing), while values closer to 1 indicate better performance. In practice, any useful model will have an AUC above 0.5, and if it’s below 0.5 you can flip the predictions to get above 0.5. So, the typical useful range is from 0.5 to 1.0.

Discrimination across all thresholds is what the AUC captures: how well the model ranks positives above negatives. The AUC value always falls between 0 and 1. A value of 0.5 means no discriminative ability (like random guessing), while values closer to 1 indicate better performance. In practice, any useful model will have an AUC above 0.5, and if it’s below 0.5 you can flip the predictions to get above 0.5. So, the typical useful range is from 0.5 to 1.0.