You’re doing a random sampling of the earth to find and date various bones. No matter how old the oldest one we find is, there is almost certainly an older one out there somewhere. Eventually they’ll stop being pushed as far back or found as frequently.
The GTP applies as well in wildlife census counts. Paleoarcheology is a bit like a wilddeath census -- the subjects have long since been pining for the fjords, but encounters are randomly distributed.
Online ranking and collaborative review systems function similarly as well, with the ratings being the wildlife, and the "true" value being the total production estimate. At least if no gaming or other incentives exist (and yes, they do).
