What is the difference between discrimination and calibration?
PICTURE THIS: RAG CHATBOT
Simple meaning
Discrimination is ranking: do positives score higher than negatives?
WHY — Metrics instead of guessing?
Why interviewers care about Metrics:
contrast on Metrics, not two memorised paragraphs.
the developer, then one case where picking wrong hurts.
Name the idea, why it exists, then one short example.
End with when you use it and one common pitfall.
STEPS — What happens step by step?
Before you speak the answer, walk the interviewer through these steps:
- 1Discrimination is ranking: do
positives score higher than negatives?
- 2Calibration is whether a
predicted 0.8 is right about 80 percent of the time.
- 3How it works
ROC-AUC measures ranking
- 4reliability diagrams and Brier
score speak to calibration, which matters for expected-value decisions.
- 5Common mistake
What juniors usually get wrong.
- 6Close
When you pick this over the alternative.
EXAMPLE — See it in action
Here's a short line you can speak, broken into clear beats:
Note: Adapt this scaffold to your own project — keep it under 60–90 seconds.
Key takeaway
Discrimination is ranking: do positives score higher than negatives? Calibration is whether a predicted 0.8 is right about 80 percent of the time.