Erfurt paper revisited in G25

[nureb]

Israel Iron Age, see: https://drive.google.com/file/d/1F2rKEVt...mA-sl/view

That is an aggregated sample, it is a two way mix between Anatolian and Levantine.. that does you no good. You should probably separate between Iron Age Anatolia and Iron Age Levantine.

"it is a two way mix between Anatolian and Levantine" no it's average Israel IA not aggregated. And not at all a mix between Anatolian and Levantine IA samples that would be false.

Average (s) and aggregate (collection of samples under one label) are semantically the same thing. 

I ran the sample. It had Iron Age Armenian and Iron Age Levantine 50/50 split.

[Rodoorn]

That is an aggregated sample, it is a two way mix between Anatolian and Levantine.. that does you no good. You should probably separate between Iron Age Anatolia and Iron Age Levantine.

"it is a two way mix between Anatolian and Levantine" no it's average Israel IA not aggregated. And not at all a mix between Anatolian and Levantine IA samples that would be false.

Average (s) and aggregate (collection of samples under one label) are semantically the same thing. 

I ran the sample. It had Iron Age Armenian and Iron Age Levantine 50/50 split.

I don't want to get into a semantic discussion, but one thing is clear Israel IA is not an admix of way mix between Anatolian and Levantine, or Armenian and Levantine. I'm North Dutch and you can make a model in which I'm kind of in the middle between other European population but that's only a kind of numerical average. No relationship with a real admix. Israel IA was Levantine face it.

[nureb]

Now I see where the problem really is. You are mediating the population yield with a northernmost groups (Germanic/Slavic) with southernmost groups which are Levantine and not much in between, that's where the flaw is. It is not a historically coherent model. Some may even use Italian and Levantine but not much in between that on the pretext of avoiding "overfit" at least that is somewhat makes, historically, more sense.

[nureb]

Now I see where the problem really is. You are mediating the population yield with a northernmost groups (Germanic/Slavic) with southernmost groups which are Levantine and not much in between, that's where the flaw is. It is not a historically coherent model. Some may even use Italian and Levantine but not much in between that on the pretext of avoiding "overfit" at least that is somewhat makes, historically, more sense.

Atleast that somewhat makes, historically, more sense *

Levantine and Italian would actually be underfitted, but it does meet it's 'anti-overfit' quota.