We rank matches using a combined genetic-distance score. It blends PCA/UMAP proximity (where you sit in a multi-dimensional genetic space relative to each ancient individual) with IBS/allele-sharing metrics (the share of identical alleles you have in common).
Lower distance scores mean a stronger match. Individuals and site clusters with high genotyping quality and dense SNP coverage carry more weight in the ranking; sparse or low-coverage genomes get down-weighted so they don't produce misleading placements. The result is a list ordered from your closest ancient match to your most distant one within the dataset.