inspection paradox

Arises when the value of the variable being sampled is related to the probability of it being sampled. For example, consider bus waiting times. The longer the waiting time between buses, the more likely people are to encounter such an interval. So people can perceive that the bus waiting time is longer than the quantity that you would get if you averaged all the bus waiting times together.

There's another way to view this, which is you get different values depending on how the expectation is computed. The expectation across bus waiting times is one quantity. The expectation across people's experience is another quantity.