I thought I was pretty clear. I guess not.
The thing for which these tests and calculators are best, imo, is for comparison with people of your area. In my case, for example, it was clear from the first 23andme test and the dienekes calculator that I fit exactly with the people of my two ancestral areas. In fact I plot almost exactly right in between Bergamo and TSI. When the sample from Piemonte(actually a border area between Liguria/Piemonte/Lombardia and Emilia) became available I plotted extremely close to that.
So, I'm exactly who I thought I was.
If, however, you come from a "mixed" background, as most Americans do, it becomes very muddy. Historical borders from the time of the nation states don't correspond to ancient "population" enclaves, who were themselves combinations of prior groups in different proportions.
If you're a mixed American who wants to know what percent German, English, Scottish, Irish, French Huguenot and Scandinavian you are, these tests can't do it. Either the populations are too similar, or they're going to be differently computed based on where centrums are located and named etc.
So, a strange situation has developed where if you know where all your ancestors lived for the last 500 years, the tests accurately confirm it. If you don't know, they won't tell you with any great degree of accuracy.
Now, as Davef said, it's interesting to know what ancient populations contributed to that. How much EEF, WHG, CHG etc. How close am I to MN people, or steppe people, or, if the genomes are ever released, Etruscans, or Romans, or the ancient Ligures or the Bell Beakers. I'm interested in knowing what ancient populations created us.
I'm not going to find that out from 23andme or any of these other companies.
I'll give you an example of what I mean by confirmation bias in all of this. A lot of Americans and Canadians have family traditions of having an American Indian ancestor and are very invested in the idea. I knew one of them. All the early tests he took showed no such ancestry. He took test after test after test until he found one that gave him a tiny percentage. Now, is it real or just noise, an artifact of that particular algorithm? I have no idea, but he's now happy. The other side of that story is that when people like this take a dna test it more often than not turns out that the "Indian" ancestor was actually a person with some SSA ancestry who said that in order to "pass" into white society.
Or, for another example, I knew a man obsessed with the idea that he was descended from the Norman aristocracy of England and was therefore convinced that he had Scandinavian ancestry. He took all these tests, looking for the one that would give him the highest possible "Scandinavian" score. Far be it from me to burst his bubble and tell him that I doubt that at this stage anyone, particularly anyone at a commercial testing site, can tell the difference between a Danish Viking and an Anglo-Saxon peasant farmer. For no doubt typically masculine reasons, he much preferred being descended from Danish Vikings.
See, different strokes for different folks. Maybe I'm doing the same thing coming at this from a totally different angle. My mother's family tree is littered with the surname of our local robber baron aristocrats. I nevertheless imbibed a hatred for them with my mother's milk, probably literally.
(My birth area has always been a hot bed of anarchism, socialism, communism, and before that of any populist uprising that developed.) I have no desire to be connected to them. I much prefer to think I'm descended from retainers who took their surname. I also think that's objectively the most likely. On the other hand, I wouldn't take a test to see if we're related to the last descendants of that family even if they were willing to do it. See what I mean?
