Statistical predictions (as for rainy days, accidents) are typically better for how many of a group will experience something than for which members of the group will experience it- -and better for how often something will happen than for exactly when.
See the general discussion in Unifying Concepts and Processes: Systems,
order, and organization.