Something broke in random Ruby tests. Some tests say that I have wrong answers, but when I check them through puts, it gives the correct values, those - my solution returns the correct answers, but the tests perceive them with different values

For example

def champernowe_digit(n)

...

p res #=> 6

res

end

#=> Random digit tested: 8563414405108046 - Expected: 6, instead got: 5

Can one get it to run in constant time?

Everybody is doing it in O(log n), which is plenty good enough really.