Find the Winner of an Array Game

You can simulate the game without physically rotating a queue after every comparison. Treat the first value as champion and scan challengers in order. If champion beats the challenger, increment its streak; otherwise make the challenger champion and reset streak to one. Return when the streak reaches k. The important optimization is that the global maximum cannot lose, so if k is very large the answer is simply the maximum once it becomes champion; at most n-1 challenger comparisons are needed to reach it. Explain the invariant: after processing each challenger, champion is the winner that would be at the front and streak is its consecutive wins. Test k one, already-largest first element, descending input, duplicates according to tie rules, and k much larger than n. The optimized scan is O(n) time and O(1) extra space, whereas literal queue simulation can be unnecessarily long for huge k.