👪 Cousin Relationship Calculator

Enter how many generations back each person is from your shared common ancestor (parent = 1 generation, grandparent = 2, great-grandparent = 3, and so on) to find the precise cousin relationship label.

From two generation counts to a label

Count each person down from the couple they both descend from, a child of that couple being 1. The smaller count fixes the cousin degree and the gap fixes the removal:

degree  = min(A, B) - 1
removed = max(A, B) - min(A, B)

The defaults, 2 and 3, give degree 1 and removal 1: first cousins once removed. Degree 0 is not a cousin relationship at all, so the tool switches wording and reports siblings when the counts match, or an aunt or uncle to niece or nephew chain when they do not. Labels are printed mechanically, which is why you see "1 time(s) removed" rather than "once removed".

What that label predicts in DNA

Because each generation halves the DNA passed on, and the pair descends from two shared ancestors, the expected share follows from the same two counts:

expected share = 2^(1 - A - B)      cM = share × 6,800
A, BRelationshipExpectedObserved range
2, 21st cousins12.5%, ~850 cM575-1,330 cM
2, 31st cousins once removed6.25%, ~425 cM200-850 cM
3, 32nd cousins3.13%, ~212 cM90-480 cM
The observed ranges are far wider than the expectations, and they overlap: a measured total narrows the possible relationships rather than proving one. Half relationships descend from a single shared ancestor, so halve the expected figure. Endogamy pushes real totals above it.

Frequently asked questions

How much DNA do second cousins share?

About 3.13% is expected, near 212 cM, but real second cousins have been recorded anywhere from roughly 90 to 480 cM. Published averages sit a little above the theoretical figure because pairs sharing too little to detect never appear in the data.

What numbers do I enter for a first cousin once removed?

2 and 3. Your first cousin is a grandchild of your shared grandparents, and their child is a great-grandchild, so the counts differ by one.

Can I work backwards from shared cM to these numbers?

Only roughly. Several combinations of A and B produce overlapping totals, so start from documented trees and use the DNA to test the label, not to generate it.