Shared Roots: Ancient Oceania by DNAGENICS

[szin]

"Oceania Ancient - Vanuatu - Malakula MAL002"
255 total shared SNPs. 4 cm on chr11.

Is this really reliable ? It's seems unlikely to me that I could match such samples.

It could indicate ancient East or South Asian ancestry- not entirely out of the realm of possibility given the Huns and other Steppe populations crossing into Europe.
As an example: I often get tiny amounts of Oceanian/Aboriginal/Andamanese on G25 runs. In my case, I know this to be most likely a signal of an Austronesian population like the Santali who contributed to the Romani gene pool. You're half-Serbian so more likely to have Hungarian ancestors in genetic time, potentially with not insignificant East Asian admix. In my case, you have to go back to the thirteenth century to find a Hungarian ancestor. Despite this ancestor being significantly East Asian admixed (her mother being of the Cumans), I score next to no East Asian on calculators.

Anyway, they updated the algorithm and added new populations. I don't match with as many as I do on the Ancient African Shared Roots, which is surprising: