I'm not sure I saw what you were referring to.
I thought this was interesting "In general, we sought to make the populations as small as possible, subject to the constraint of being able to distinguish between them. Populations may be inherently difficult to distinguish because of historical mixing, or we might not have had enough data to tell them apart. As we obtain more data, that populations will become easier to distinguish."
Do you think this this is true for ancestry as well?
But how would they distinguish individuals out of a group or determine which portion of a person's ancestry would be passed on by a couple.
For example, I have an ancestor who was from Wetheral England. I actually saved each person from the 1841 census and did a quick study of them. I was surprised at the number of men there who listed their birthplace as local but their wives were from Scotland. If you went by the men's surnames you would have a different impression than by what their wives were. And that could change again another generation back but perhaps only for a few of those. So their wives may have moved in location. Also, their surnames would change (usually) each generation so even if they had a Scottish maiden name their mother may not have, etc.
Ancestry hasn't tested this group so no reflection or comment directed at ancestry here. I am concerned about an overgeneralization of individuals for location, surname or cluster and particularly where it becomes applied science.