The fact that we don't keep repeating tests in the same arena is not because the probability of the hypothesis showing its falsity in other arenas goes up after it has passed tests in one arena.