Compare AVL trees and Red-Black trees. When would you prefer one over the other?

AVL trees are more strictly balanced (height difference <= 1) than Red-Black trees (longest path <= 2x shortest), making AVL faster for lookups. However, Red-Black trees require fewer rotations for insertions/deletions (max 2 vs potentially O(log n) for AVL). Choose AVL for lookup-heavy workloads (databases) and Red-Black for modification-heavy workloads. Red-Black trees are used in C++ STL map/set, Java TreeMap; AVL in some database indices.