**Reliability:**

Reliability characterises the capability of a device, unit, procedure to perform without fault. Reliability is quantified in terms of probability. This probability is related either to an elementary act or to an interval of time or another continuous variable. Because the probability of failure is normally a small fraction, the reverse ratio rounded to integers is usually used.

For example, reliability of computer hard drives may be quantified as the probability of failure in one start/stop cycle. But, in reality, the reverse value is used - the average number of start/stop cycles per one failure (in usual hard drives this value is about 30,000...50,000). Another example is reliability of commercial jet-flights, which is normally related to the number of takeoffs/landings. In situations of this sort, the probability of failure is related to one elementary act (e.g. one hard drive failure, one plane crash).

On the other hand, in many real-world situations, the probability of failure is related to an interval of time or another continuous variable. For example, the reliability of electric bulbs (and many other devices) my be expressed as the probability of failure per one hour of operation. Reliability of the transmission in a car may be expressed as probability of failure per one mile. Normally such probability is presented in reverse units - time (or other quantity) of faultless operation per one failure, e.g 1,000 hours/failure for a bulb, 10,000 miles/failure for car transmission.

In statistical reliability theory, an important role is played by the Poisson process , which is a good model for failure events in time (or other continuous quantity, like distance). In such cases, the parameter characterises the incidence of failures, and the reverse value characterises the average operation time per one failure.

In surveys and tests, e.g. in psychometrics , the term "reliability" has a somewhat different meaning. See Reliability (in Survey Analysis) .