Understanding Centimorgans: The Unit Behind DNA Matching

If you have used a consumer DNA test, you have seen a number labeled “cM” attached to every match. That number, the centimorgan count, is the single most useful piece of information in your match list. It is also the most misunderstood. A centimorgan is not a measurement of physical DNA. It is a unit of genetic distance, and the difference matters when you are trying to work out how you and a match are related.

What a centimorgan actually measures

A centimorgan, abbreviated cM, measures the probability that two points on a chromosome will be separated by recombination during meiosis. One centimorgan is roughly a 1% chance of separation per generation. A 50 cM segment of chromosome has roughly a 50% chance of being broken up by recombination in one generation, on average.

This is a probabilistic, not a physical, unit. A centimorgan corresponds to roughly one million base pairs of DNA on average, but the ratio varies across the genome because recombination is more frequent in some regions than others. When you see “you share 850 cM with this match,” the company is reporting genetic distance, not a base pair count.

Why total shared cM predicts relationship

Closer relatives share more cM because fewer generations of recombination have broken up the DNA they inherited from a common ancestor. Full siblings share an enormous amount of DNA, around 2,500 cM on average, because they share two parents. First cousins share around 850 cM because they share two grandparents and DNA has had two extra generations to recombine. Second cousins share around 230 cM. Third cousins share around 75 cM. By fifth cousin, shared DNA falls below the noise floor of most tests.

The actual numbers, drawn from real measurements, come from the Shared cM Project, an ongoing crowdsourced study run by Blaine Bettinger and collaborators. It is the standard reference for how much shared DNA each relationship range typically produces, and we use its numbers throughout our shared DNA and relationship chart.

Total cM versus longest segment

A match list shows two cM numbers. Total cM is the sum across all your shared segments. Longest segment is the length of the single biggest one. Both matter, and they tell you slightly different things.

Total cM is the primary signal of how closely you are related. Longest segment helps distinguish recent from distant shared ancestry, because long segments survive fewer generations of recombination. Two people who share 85 cM in a single 85 cM segment are usually much closer related than two people who share 85 cM spread across eight 10-15 cM segments.

This is why the Shared cM Project tracks both numbers, and why advanced genealogists treat segment data as a meaningful signal rather than just total cM.

Where the noise floor is

Below roughly 7 cM for a single segment, shared DNA gets unreliable. Short segments occur by chance in any large database, and at small sizes you cannot distinguish a real shared ancestor from coincidence. Identical by state, where two people share a stretch by coincidence rather than common ancestry, becomes more common at small segment sizes.

The consumer companies set their own thresholds. AncestryDNA hides matches below 8 cM by default. 23andMe and MyHeritage use similar cutoffs. We do not recommend lowering them; the additional matches are mostly noise.

The same cM amount can mean different relationships

Shared cM does not give a single relationship answer. It usually narrows the options to a range. Around 1,750 cM total can mean a parent-child relationship, full siblings, or in unusual cases a grandparent-grandchild or aunt-niece relationship. The cM count rules out distant cousins; it does not pick which close relationship is correct.

This is why a relationship predictor tool that returns a range of plausible relationships is more useful than one that returns a single guess. The DNA narrows the possibilities; family context and triangulation with other matches narrows them further.

How recombination produces the variation

Two siblings share around 2,500 cM with each other on average, but the actual numbers in real sibling pairs range from around 2,209 to 3,384 cM, according to the Shared cM Project. That variation is the result of recombination being a random process. Each child inherits a random shuffle of each parent’s chromosomes, so siblings end up with overlapping but not identical inheritances.

The same is true for any relationship. The averages are clean. The distributions around the averages are wide, and they widen as you move out to more distant relatives. Third cousins share an average of about 75 cM, but the real-world range runs from 0 cM, meaning they share no detectable DNA at all, up to several hundred cM. About 10% of third cousins do not share enough DNA to show up as matches.

For how matching works end to end, see how DNA matching works. For the relationship table itself, see shared DNA and relationship chart, and for the pillar overview, see the ancestry and genealogy guide.