What is MRR and where is it used?

MRR (Mean Reciprocal Rank) is a quality metric for ranking, often used in information retrieval tasks and ranking systems (Learning to Rank). It measures the average of the reciprocal ranks of the first relevant result in a list of answers.

Example: if the relevant document for a query is at position 1, the reciprocal rank = 1/1 = 1; if at position 3 — 1/3 ≈ 0.33. MRR is the average of such values across all queries.

MRR helps evaluate how quickly a user receives a relevant result and is used in search engines, chatbots, recommendation systems, and other applications where the position of the relevant answer is important.