An e-commerce platform handles product search queries. Each search filters on category, price range, and brand. Currently, the database has a separate index on each filter column. The database retrieves matching row IDs from each index and intersects the results, examining approximately 25,000 rows per query. A proposed change creates a single combined index on (category, price_range, brand), narrowing to approximately 800 rows per query in a single traversal. Why does the combined index produce fewer rows examined?