The core idea

Measure retrieval and generation separately.

A RAG answer can fail two ways: wrong context retrieved, or good context used badly. Evaluate each stage on its own to know which broke.