Why does HashMap use a red-black tree instead of a regular binary tree?

In Java, HashMap uses a red-black tree instead of a regular binary tree to store elements in a single bucket during collisions, ensuring balance and, consequently, logarithmic time for search, insertion, and deletion.

A regular binary tree can degrade into a linear structure (list) in unfavorable insertion orders, which worsens performance to O(n). A red-black tree is self-balancing, providing stable performance even with a large number of collisions.