Family Tree DNA Updates Matching Thresholds

EDIT 11/12/2019 – Sometime in 2018 or 2019, Family Tree DNA made minor modifications to the matching threshold scheme described below. Although the current scheme is close to the one outlined below, it is proprietary and therefore difficult to predict exactly.


Family Tree DNA will be updating its matching thresholds in the very near future (QC is currently being run). Here’s a flowchart of the modified matching algorithm:



From FTDNA’s announcement (modified to indicate that the changes are coming very soon):

You asked for it – we listened!

For several years the genetic genealogy community has asked for adjustments to the matching thresholds in the Family Finder autosomal test.

Currently, the current matching thresholds – the minimum amount of shared DNA required for two people to show as a match are:

– Minimum longest block of at least 7.69 cM for 99% of testers, 5.5 cM for the other one percent

– Minimum 20 total shared centiMorgans

Some people believed those thresholds to be too restrictive, and through the years requested changes that would loosen those restrictions.

Very soon, the following changes will be implemented to the matching program.

– No minimum shared centiMorgans, but if the cM total is less than 20, at least one segment must be 9 cM or longer.

– If the longest block of shared DNA is greater than 9 cM, the match will show regardless of total shared cM or the number of matching segments.

The entire existing database has been rerun using the new matching criteria, and all new matches have been calculated with the new thresholds.

Most people will see only minor changes in their matches, mostly in the speculative range. They may lose some matches but gain others.


The announcement FINALLY! settles the 7.69 cM versus 5.5 cM debate! Both were technically correct, although the vast majority of people were at the 7.69 cM threshold. I’d never encountered a 5.5 cM threshold in any of my many test-takers.

I just downloaded data for me, my mother, and my father to take a “before” snapshot. I’d like to compare that to an “after” snapshot. It will be a useful comparison, although I’m usually more interested in matches that won’t be affected by this low threshold.

For more:

18 Responses

  1. Stuart Jordan 24 May 2016 / 9:56 am

    I’m not sure this is good news for Ashkenazi Jews and others from homogeneous groups. This just feels like it is going to throw even more untraceable matches into the mix. I wish they would give people a choice on the level as opposed to imposing on everyone.

    • Blaine Bettinger 24 May 2016 / 10:07 am

      Stuart – I agree, this will likely produce more matches for endogamous populations. However, there is always the ability to filter matches at FTDNA by total shared cM, cM segment size, and so on. I for one would rather have all the data, and then be able to filter and analyze it as I see fit.

  2. Gone Researching 24 May 2016 / 11:05 am

    What/How exactly are you snapshot-ing at FTDNA to have a comparison? Just the kits you manage, or some information from particular matches of those kits?

    • Blaine Bettinger 24 May 2016 / 11:21 am

      I downloaded my chromosome browser data (it’s a one-click CSV download on the chromosome browser page) for myself, my father, and my mother.

      • Patricia Shears 26 May 2016 / 11:17 pm

        Sorry new to this.,what is a. Chromosome browser. Where can I get it pls.?

  3. Adam Dunn 24 May 2016 / 8:16 pm

    When are they going to use the X for matching?

    • Irish in NYC 25 May 2016 / 11:26 am

      I agree, FTDNA needs to include X and X-only matches. 23andMe does this already.
      This would be more valuable than tinkering with the low end of atDNA. For me, FTDNA’s latest change is quantity over quality. FTDNA is the best of the big 3, but there is a heck of lot more they could be doing (how about built-in triangulation?, integrating Advanced Matching with Matches, integration of Parent-Child kits, just to name a few).

  4. Kalani 24 May 2016 / 10:33 pm

    I have a few project members whose LONGEST BLOCK is less than what my mother has (5.11cM). My aunt has 4.63cM, total is 133cM. While I know that these would not be a match within a genealogical time frame, I guess (according to your flow chart) that they would no longer be matches because it will all now be 7.69cM? Or will the 1% still apply to those of us who do fall out of that 7.69cM minimum? Janine told me last year that I’m the exception. 🙂

  5. Sarah L M Christiansen 24 May 2016 / 10:36 pm

    I would do something different. If the longest segment is at least 9 AND the total is less than 20 then they need to have surnames or a tree, otherwise it is just clutter. I realize that people test because someone else requested it, but put in what is known. And tiny matches will not help adoptees or others who do not know their family history.

  6. jim 25 May 2016 / 1:39 pm

    not convinced that its the correct response – I have a bunch of matches that would not be a match under the new format – Matches in the 1600- early 1700’s are most always less then 5 – usually around the 4 level – also some have no regular matches but match for more then 10 on the X – I have documented matches at this level and thanks to a tree link could trace the ancestory back – now they are not a match? Someone is not thinking of those of us who are searching in the early 1700’s

    • Blaine Bettinger 25 May 2016 / 2:10 pm

      Jim – unfortunately, research has shown that 33% of phased 5 cM segments are false, and the percentage increases drastically for 4 cM and smaller. It is currently impossible to decipher between real and false segments in the 5 cM and below range. Or, at least, there is no peer-reviewed scientific research to show a method for deciphering between 4 cM and 5 cM.

      For example, there is no evidence to suggest triangulation of a segment this small means it is real instead of false. There is no evidence to suggest that finding this segment in a parent means it is real instead of false. Both would suggest to me that they are very likely real, but with the 33% figure looming overhead, the amount of evidence needed to convince someone analyzing this with an objective scientific background is VERY high at the current time. I have high hopes for the future, but it’s a long way off.

      Let’s say that you detest white chocolate, it makes you sick. I give you a bowl of M&Ms filled with a combination of 66% milk chocolate M&Ms and 33% white chocolate, but you can’t visually tell the difference. Would you grab one and eat it? I wouldn’t.

      • Irish in NYC 29 May 2016 / 8:00 pm

        I agree that below 5 seems perilous although perhaps it can work within a triangulated group with documented trees that are hopefully more recent than the 1700s. For example, you have people who match well above 5 cM on one (or more) segments, but one member of the documented tree group matches at say 4-5 cM on one of the segments. This would still seem to support the common ancestry.

        Also do you know if below 5 cM rule still holds if the number of SNPs increases (I’m thinking like 750+ SNPs and 4-5 cM).

        Finally, I assume you are still talking about atDNA here. I wonder if this rule applies for X?

  7. Janis 25 May 2016 / 7:15 pm

    I have a 3rd cousin who has tested at Ancestry and is on GEDmatch who shares only a 5.5 cM largest segment with me. So I really do want to see and judge these seemingly weaker matches for myself. This 3rd cousin was in my Ancestry matches and hints until their updates set the minimum at 6.0 cM.
    I have 2 other 3rd cousins who do not appear in my matches at all. They have not tested at FTDNA but have tested at Ancestry and are on GEDmatch. One shares a segment over 9 cM and over 13 cM total but is still not in my matches.

  8. Sharon Brown 27 May 2016 / 4:08 pm

    Could I have a copy of your free book. Address is 31590 Champions Circle, Temecula, CA 92591

  9. GH BASS coupon 23 May 2017 / 11:07 am

    Scorpions signe astrologique tirages tarot gratuit

Comments are closed.