The reliability/availability assessment of a periodically tested component is dependent on its standby failure rates. This is not a directly measurable quantity and must be inferred from test data. The probability of failure to come "on-line" when need arises can be determined using the binomial distribution. This probability can then be used to determine s.