Is the number of failures directly proportional to the rate to success?