The coefficient of inbreeding (COI) estimates how closely related an animal's parents are. It's the probability that both copies of a gene were inherited from the same ancestor - higher values mean a more line-bred pedigree.

Best In Show computes COI from the pedigree using the standard kinship method. That means the number is only as good as the family tree behind it: a COI computed from 5 known generations tells you far more than one computed from 2. Missing ancestors are treated as unrelated, which underestimates the true COI.

COI tools are available on the Pro and Team plans.

Every animal's Pedigree tab includes a COI panel showing the computed coefficient along with the common ancestors that contribute to it - the animals that appear on both the sire's and the dam's side of the tree.

Under the percentage we translate it into the pairing it corresponds to - for example, 8.4% sits between a first-cousin pairing (6.25%) and a half-sibling pairing (12.5%). Both the panel and the test-mating calculator show this reading.

We describe the figure rather than score it. What counts as an acceptable COI depends on your breed's own baseline, and in a closed studbook the breed average can already sit above a first-cousin pairing, so a single good-or-bad scale would flag ordinary, well-planned matings as a problem. The interpretation stays with you.

To improve the calculation, keep building the pedigree: ancestor profiles are free on every plan and don't count toward your animal limit.

The test-mating calculator lets you evaluate a pairing before it happens. On a Pro or Team plan you'll find it as COI & test mating at the top of My Animals; the COI panel on any animal's pedigree also opens it with that animal already filled in.

1. Search for and select a sire.
2. Search for and select a dam.
3. Best In Show computes the projected COI of the hypothetical offspring and shows a preview of what the litter's pedigree would look like, including the shared ancestors driving the number.

You can test any combination of animals on the site - your own, another breeder's stud, or ancestor profiles - as long as the genders and animal types make a valid pairing.

We take the math seriously, so here's what's happening under the hood. Best In Show uses the kinship method (also called the tabular or coancestry method) - the same approach used by animal-genetics researchers and professional pedigree databases. It's more accurate than the simple "path-counting" shortcut many online calculators use, especially on the real, partly-filled family trees breeders actually work with.

The kinship method rests on one elegant idea:

An animal's COI equals how closely related its two parents are. So to score a pairing, we measure the kinship between the sire and the dam by walking up both sides of the tree together.

That single rule, applied across the whole pedigree, means our numbers correctly handle the two things simpler tools get wrong:

• Ancestors who appear many times. A dog that shows up in five different places in the tree is counted through every one of those paths, not just once.
• Ancestors who are themselves inbred. If a shared ancestor was line-bred, that extra relatedness is carried forward into the final number, exactly as the genetics require.

As a sanity check, the classic textbook pairings come out to their known values: a parent bred to offspring or two full siblings both yield 25%, half-siblings give 12.5%, and first cousins give 6.25%.

Alongside the percentage, we run a second calculation - Wright's path method - over the same pedigree to produce the list of common ancestors and show how much each one contributes. On a fully recorded pedigree those contributions add up to the percentage itself.

Two things we're deliberately careful about:

• We never guess. Unknown ancestors are treated as unrelated rather than assumed, so an incomplete pedigree can only ever under-state the true COI - it will never invent inbreeding that isn't documented. That's why we also show how complete the underlying tree is.
• Bad data can't break the number. If a pedigree accidentally lists an animal as its own ancestor (a loop), we detect it and tell you there's a data problem instead of returning a wrong or nonsensical figure.

Want to verify the approach for yourself? The method is well documented: Malécot's coancestry (kinship) method and Wright's coefficient of relationship.

• Fill in at least 4-5 generations on both sides before leaning on the number.
• Avoid duplicate ancestor profiles. If the same ancestor exists twice under slightly different names, link the same profile in both places - duplicates hide common ancestry and lower the computed COI.
• COI is one input, not a verdict. Use it alongside health testing and your own knowledge of the lines.