What is evaluation of retrieval systems using Recall@K and MRR tradeoffs?

Recall@K measures coverage of relevant items, while MRR measures ranking position quality.

Recall@K evaluates whether relevant items appear in top-K results, focusing on completeness. MRR (Mean Reciprocal Rank) focuses on how early the first relevant result appears. These metrics often trade off: optimizing for recall may reduce ranking precision, while optimizing MRR may ignore broader coverage.