Randomization Controls for All Factors, Known and Unknown

The hope is that you control for all possible factors, known and unknown. Half of them you give the treatment to, half of them don't get the treatment. That's supposed to show whether or not the treatment works. Because we can see, if one side gets better and te other side doesn't get better, it looks like it's because of the treatment. It looks like whether it is is another it's another question.