A genetic test to see whether someone has ancestry from Melenasia or Sudan or Scotland is likely to give a different result than casual visual classification. The latter is what we commonly call "race".

I can tell the difference between German and Irish people at better than chance and you think people can’t tell the difference between Melanesians, Caucasians and Sub Saharan Africans? Bone structure, skin tone, hair colour and hair texture don’t give as much information as a whole genome sequence but telling the difference between an East Asian and a West Eurasian isn’t rocket surgery. By the same token it’s not hard to tell the difference between an Ethiopian and a Nigerian, or a Swede and a Spaniard, on average.

> You think people can’t tell the difference between Melanesians, Caucasians and Sub Saharan Africans?

Get someone to show you pictures and have you identify them. I bet you'll make a lot of mistakes.

Historically, not only can people not tell the difference objectively, they decide what the difference is subjectively. In the United States, the child of a white slave owner and a black slave, if their parentage was known, was ... black. People of "mixed race" might be deemed a different race than their own full siblings, based on which visible characteristics they got from which ancestor.

Even today, people migrating from one country to another are often surprised that their racial classification is different in the new country than at home.

>> You think people can’t tell the difference between Melanesians, Caucasians and Sub Saharan Africans?

> Get someone to show you pictures and have you identify them. I bet you'll make a lot of mistakes.

I’m going to sleep now but if you find ten pictures of each I’ll be happy to do your survey. I’m sure many others would be too if you put it on surveymonkey.

The overlap in skin tone alone between Caucasians and the other two groups is ~0 so I doubt there would be any mistakes there. I’m sure there would be many more mistakes between Melanesians and Sub Saharan Africans but even just going on whether they have a broad nose or not you’d get pretty far distinguishing SSA from Melanesians. I’m not claiming perfect distinction. I’m claiming that doing better than chance is unremarkable. These groups are visibly different on average. Some groups the population values don’t even overlap. The shortest adult Maasai is taller than the tallest adult Pygmy.

> In the United States, the child of a white slave owner and a black slave was... black.

Who cares about the historical American caste system, with its one drop rule? The social construction of race is a real thing and so is the fact that different ancestry groups are different, and can be distinguished, on average and subject to error. Pretending ancestry is entirely socially constructed is crazy. If you get thirty pictures, 10 East Asian, 10 West Eurasian, and ten people who are first generation crosses from both populations, there is very unlikely to be any mislabelling of the base population individuals.

All human ancestors came from the continent of Africa if you go back far enough. Which specific ancestors are you referring to?

On the coarsest basis the groups that existed 3-5,000 years ago. The major continental ancestry groups of today were pretty much done by then, whether you divide them at K=2 and have SSA and everyone else, SSA, West Eurasian and Asian at 3, add the Americas at 4, and Melanesians, Papuans and the various ethnoi that resemble them from the Aboriginals in Australia to the Andamanese or Negritos at 5. 5,000 years ago those groups all existed in a somewhat recognisable form. Plenty of the modern variety in them didn’t. South Asians didn’t, Northern Europeans, didn’t. Both emerged as a result of population movements and adaptation after that. 5,000 years ago Europeans were blue eyed but either brown or black skinned and the population of the Indian subcontinent were more like modern Andamanese than any other group. Ancestral South Indians are a mix of that group and Iranian plateau agriculturalists and Ancestral North Indians were horse riding milk drinking central Asians.

If you go back far enough all humans come from Africa, and most (at least 90%) comes from the latest Out of Africa expansion, ~200K years ago. But all non SSA humans are 2% Neanderthal, on average, and there’s other archaic hominid introversion that varies by geography. The Tibetan high altitude adaptations are probably 2 million years old.

Which ancestors one refers to will depend on the question you’re asking.

I thought the out of africa theory was recently challenged dye to older human findings elsewhere.