"Yet there seemed to be some truth in the law of probability, according to which the chance of success is directly proportionate to the number of repetitions."