Since it divides the array into two parts, although not equal, its time complexity is O(logn), it is better than binary search in case of large arrays.