Why nonfair lock is faster than fair lock (java.util.concurrent.locks.Lock) in Java Concurrency

This is the nonfair scenario:
When a thread B asks to hold a lock, if the lock is already taken by another thread A. Then the thread B will be suspended. After thread A finished his job, it releases the lock, then thread B is resumed. However, B needs a period of time to be able to actually run it's task from where it was suspended. And this period of time could be relatively long compared to another thread C who is asking for that exact same lock at the same time. If C could acquire the lock, does C's job and release the lock before B even "wakes up" (the time between B's resumed and B's able to actually run), then it's a win-win situation for B and C. Because C gets the lock and has it's job done and releases the lock, while B's "waking up" and B gets the lock no later than it otherwise would have.

In the fair scenario:
The C will be queued after B.

To set the fairness of a Lock, one can use the parametered constructor:
Lock lock = new ReentrantLock(true); //true for fair, false for nonfair