I suspect the missing bit of context here is that we have a lot more potential sites than that we have archaeologists with the time and budget to investigate all of them.
So in programmer terms, most potential sites never make it out of the priority queue