##### ###### ##### ### # # ### # # ###### ## ## ## ## ## ## ## # # # # # ## ##### #### ##### # # # # # # # #### ## # ## ## ## ## # # # # # ## ## # ###### ## ### # ### # ######
##### ###### ##### ### # # ### # # ###### ## ## ## ## ## ## ## # # # # # ## ##### #### ##### # # # # # # # #### ## # ## ## ## ## # # # # # ## ## # ###### ## ### # ### # ######
(Divide and conquer)
Each comparison eliminates half the array. After 1 step, n/2 remain. After 2 steps, n/4 remain. After k steps, n/2^k remain.
When n/2^k = 1, you have found the target or determined it does not exist. Solving gives k = log₂(n).
This is O(log n). No other comparison-based search can beat this on sorted data.
