Typically one would assume that it was not coincidental and then adjust the bounds for how often we expect the event occurs based on observations, or lack thereof.