[GEDmatch] As a Spaniard, I usually get both north african and scandinavian in Oracle

[agil]

Hello.

Here is the thing. I am a spaniard with 4 grandparents from different regions and places. I always get spaniard as first option in Oracle - that's not a surprise.

The thing is, with 3 or 4 populations the results changes. I almost always get 50% spanish / 25% spanish + 25% spanish and then 25% norwegian/german/english/whatever + 25% sardinian/algerian/mozabite/jewish

Now I understand how GEDmatch works. I figured out that, since I am from different regions, the little differences might actually make a sardinian + norwegian more accurate than spaniard from region X + spaniard from region Y even if the latest is actually true. But most calculators use spanish regions and it makes no sense that it would show up such big disparity as first and most accurate option in 3 and 4 population instead of spanish + spanish.

These are my kits in case you want to see them by yourselves:

My GEDmatch kit from 23andme: M201977

My genesis gedmatch kit from LivingDNA: PT2976460

Some Gedmatch savvy guy can explain this?

[randwulf]

I believe you already have stated the correct answer in your post. When the Oracle search deals with your particular variation from the spreadsheet averages it is able to get closer by using out of area combinations. The three and four population Oracles aren't usually correct for me, either.

[JerryS.]

Spain it seems like Italy has several regions used because of the influences from other countries make them lean one way or the other. different calculator models will reflect different slants and pull/drift whatever the word is one way or another. they will use whichever population pool they are made for as replacements or substitutes, but they will still be close/neighboring areas.