Yesterday, Ancestry announced that they are going to remove smaller matches from their customer’s DNA match list around the first of August. I was not on this conference call myself. However, it’s a small community and others have reported consistent information.
I’m going to report what was said, then lay out a strategy for you to preserve your most useful smaller matches.
Update 7-18-2020: I have received clarification on several questions. I’m putting the updated information here:
Only segments below 8 will be deleted. However, Ancestry “rounds up,” so a segment between 7.5 and 7.99999(repeating) will be rounded up to 8. The only way to assure that you save all of the segments between 7.5 and 8 that you wish to preserve is to add all 8cM segments to groups or make notes, as described in the instructions. I’m referring to these segments as 6-8 cM segments.
ONLY segments to be salvaged will be ones in groups, with notes or matches whom you have messaged. Ancestry has confirmed that matches without these things, meaning matches in ThruLines or that you have placed in your tree will NOT BE PRESERVED unless they are grouped, have notes or you’ve messaged.
The determining factor is total cM, not smallest cM. So total cM between 6 and 7.9999, which rounds up to 8 will be removed. Multiple 6 cM segments where the total is 12 will be fine, for example. Again, it’s the total cMs, so no math needed.
Also, in August, Ancestry is adding decimal points to the amount of cMs. After that occurs, you won’t need to save 8 cM matches to salvage 7 cM matches that have been rounded up. Focus on 6 and 7 cM first.
7-26-2020 Update – Ancestry has posted that they have delayed the purge until the beginning of September, allowing another month to save 6-8 cM matches. They also confirmed that starred matches (in the groups) will also be saved.
A Little Background
For a bit of history, this isn’t the first time that Ancestry has removed a large number of matches. In fact, they’ve done it twice previously, first with the introduction of their Timber algorithm resulting in a loss of over 50% of matches, an event nicknamed “Autosomalgeddon.” The screaming could be heard round the world.
In May of 2016, Ancestry revised their algorithm again which resulted in losses.
Then, as now, Ancestry told customers the matches that they lost would be mostly “false matches,” but in many cases, these matches proved to be accurate but from endogamous populations. For example, I lost most of my Acadian matches because Ancestry determined that they were simply a “pile-up region.” Those matches have proven to be valid and triangulate elsewhere.
In any case, there’s no point in crying over spilled milk.
However, the glass is about to be tipped over again so let’s figure out how to make the best of this situation and preserve as much as we can.
Recently, Ancestry announced that they have sold a total of 18 million DNA tests. That’s good news for testers, at least on the surface.
If you’re a regular Ancestry user, you might have noticed that their response recently has been slow and buggy with regular timeouts. I see this notice regularly.
Last month, Ancestry’s lawyers sent cease-and-desist letters to every third party tool (that I know of) that their DNA customers can utilize in order to perform activities such as clusters, downloading a match list and downloading a list of direct line ancestors for your matches into a spreadsheet so that you can search for common surnames that might NOT already be in your tree. That, by the way, combined with triangulation, is the key to breaking down brick walls and pointing the way to previously unidentified ancestors – but I digress.
As you know, I’m a huge fan of both Genetic Affairs and DNAgedcom, both of whom provided tools to enhance and manage the huge list of matches at Ancestry – turning those individual matches into something infinitely more useful.
No one can reasonably evaluate 92,000 matches or make use of most of them.
You can see how many matches fall into each group. Unfortunately, by selecting the custom range, the list of matches is displayed, but the number of matches in that range is not displayed.
The answer is NOT to remove those smaller matches unilaterally, which is the approach that Ancestry is taking, but to utilize better tools to identify valid matches.
These third-party tools signed on to our account, with our permission, on our behalf, and utilized the power of computers to process data that would take us days, if it was possible at all with the huge number of matches that each person has now. This is, after all, the purpose of computers – right?!?
While I was certainly unhappy with the letters threatening the people who provide us with tools to utilize our own results – I was hopeful that it meant that Ancestry was going to provide something similar internally.
Now, retrospectively, I think that Ancestry is trying to find a way to manage their 18 million testers and their matches without adding infrastructure resources. They want to reduce the processing load and when the cease-and-desist letters didn’t have the desired effect on their servers, they looked for other methodologies.
Clearly, providing users with fewer matches means less to manage in a database which equates to freeing up resources.
Ancestry’s commentary is reportedly that this purge will remove “most false matches.” Of course, it will also remove all accurate matches at that level too – and yes, you can in many cases tell the difference.
According to LostCousins and others who were on the call, Ancestry indicates that they will remove most if not all matches less than 8 cM. Today, the matching threshold after Timber and Ancestry’s academic (not parental or family) phasing algorithm is 6 cM. Their current algorithms are intended to remove most false matches.
An 8 cM match can be any of the following relationships, according to Ancestry:
However, as genetic genealogists, we know that with unphased data, 7cM matches are equally as likely to be false matches, identical by chance, as they are to be genuine matches.
There are certainly better ways to assure valid matching other than a mass deletion, such as:
- Clusters (like Genetic Affairs, DNAgedcom and others,) genetic networks that indicate that people in clusters are related to each other. These are like shared ancestors on steroids.
- Phased Data (like FamilyTreeDNA’s Phased Family Matching) that phases your matches with known family members, assigning the match either maternally or paternally.
- Triangulation (like at MyHeritage, FamilyTreeDNA and via third-party tools that received the cease-and-desist letters.)
As Ancestry did in 2016, they apparently will NOT delete matches that you’ve been using, as defined by when:
- You’ve added a match to a colored dot or star match group.
- You’ve entered a note for the match which of course indicates that you’re working with them.
- You’ve sent a message to the match.
I would hope that any matches you’ve placed in your tree would be spared as well, but that criteria is not mentioned on any list I’ve seen. (Update – they are NOT spared and will be deleted.)
I’ve also seen nothing indicating that if you share a match with your parent, which is the definition of parental phasing, that those small matches will be spared either. However, “Shared Matches with Mother” or father is in the group list, so maybe. You could easily add a group for that to be sure everyone is in a group that might be at risk. (Update – shared matches with parents will be deleted.)
It was reported that Ancestry specifically stated that a match showing up in your ThruLines does NOT preserve them in your match list. (Have received confirmation that this is accurate.)
Why Preserve Matches?
You must surely be asking yourself why you need to go to the trouble to preserve these matches – especially if Ancestry seems to think otherwise. Keep in mind that once they are gone, you have no option to work with them, ever.
There are five primary reasons for preserving at least your best matches that are in jeopardy.
- Confirming Ancestors – You can confirm your descent from an ancestor you believe to be yours. Without triangulated segment information, which is not available at Ancestry, the best way to do this is by looking at whether you match the DNA of people who descend from multiple children from that same ancestor. If you only match people descended from one child, the same child as you, it’s certainly possible that you have all mis-identified the same person erroneously in your tree.
- Sharing Information – photos, etc. You never know who is going to have what gem of information. In the past two weeks, I’ve been blessed by a photo of a third-great-grandfather and a letter from his wife to her daughter. On another line, a photo of a watch case and on a Dutch line, a box with a photo of my ancestor’s sibling surfaced. Reach out to see what kind of information your matches might have. Yes, you may have to wade thought a lot of duds, but you just never know where that nugget will be found. They are out there.
- Potential Ancestor Suggestions – Seasoned genealogists may not need potential ancestor suggestions provided by Ancestry, but new researchers certainly do. Old-timers have already done a lot of the digging – but you never know when something useful will turn up. For every brick wall that falls, there are two new opportunities.
- Y and Mitochondrial DNA Candidates – Y and mitochondrial DNA holds information not otherwise available. I wrote a short description of the different kinds of DNA, here. I utilize ThruLines and searches to locate candidates for Y and mtDNA testing for all of my genealogy lines, asking if the person would consider those tests at FamilyTreeDNA. Ancestry doesn’t offer that type of testing. Generally, I offer a DNA testing scholarship. I figure a Y or mitochondrial DNA test is the same price as a reference book (or two) and the resulting information can only be obtained from people descended from those ancestors on a specific path. In other words, that type of DNA is much, MUCH rarer than reference books. As a quick review, Y DNA follows the direct paternal (surname) line in males only, and mitochondrial DNA is inherited by both sexes of children from their mother, but only females pass it on. You inherited your mtDNA directly from your mother’s mother’s mother’s mother on up the direct matrilineal line.
- DNA Transfers – Other vendors, meaning both FamilyTreeDNA and MyHeritage offer unique and much more robust tools that utilize DNA segment information. You can transfer to either company and receive matching for free, paying only to unlock advanced features. GEDmatch, a third party tool doesn’t provide testing, but does provide additional analysis tools as well. Depending on where the majority of your family has tested or “gathered,” you can ask or encourage your Ancestry matches to transfer so that you can confirm that you share triangulated segments. You may be able to provide them with information about their genealogy that they don’t otherwise have access to at Ancestry. I wrote step-by-step transfer instructions, here, for each vendor.
Of course, if you no longer have the matches to work with – these benefits won’t be available to you. This is exactly why it’s critical to identify the most crucial smaller matches and preserve them now. Once Ancestry has removed them, they are gone forever unless they transfer to one of the other vendors.
Ok, so how can we identify and preserve the most important of these matches?
This mass extinction event is supposed to occur about the first of August. When this happened in 2016, we were never given a date and time – just a general date range and one day it just happened.
Here are my recommendations for how to preserve matches that stand the best chance of being relevant, even if they are between 6 and 8 cM.
Please note that I recommend all of these approaches, not just one. Each one will catch people that the others don’t – which will preserve the most likely matches to be useful for you.
First, under DNA Matches, create a “holding group” so you can use that group to preserve matches.
You’ll use a group as a way to prevent the deletion of the match.
Ok, let’s get started.
Auto-Clusters and Third-Party Results
If you ever ran third party cluster tools on your Ancestry data (before the cease-and-desist orders), refer to those clusters now, looking for the size of the matches, focusing on any 8 cM or lower for the longest segment. It’s probably not happenstance that you match all those people and they also match each other.
If you don’t have time to analyze each match now to determine which actual group the match belongs in, create one group at Ancestry that is simply a “Preservation Group” so that you can assign the person to the group in order that they won’t be deleted. Remember, the only matches in jeopardy are the ones from 6 to 8 cM inclusive.
ThruLines is the best tool that Ancestry provides in terms of doing the DNA-plus-tree-matching work for you. ThruLines searches for people whose DNA you match and who also have a common ancestor in your tree. Or at least someone who Ancestry thinks may be a common ancestor. It’s up to you to verify.
On your ThruLines page, click on any Ancestor appearing on that page. The fact that an ancestor appears on ThruLines means that there is at least you and one other person whose DNA matches and you share that common ancestor.
I’m going to click on Lazarus Estes, my great-grandfather, because I have several matches through him.
By clicking on the List option, at the red arrow above, you will see the various matches by their line – meaning which child of Lazarus Estes.
Unfortunately, Ancestry does NOT tell us the individual segment sizes. They tell us the total segment match (after removing anything they think is too matchy) and the total number of segments. You only need to be concerned about segments between 6 and 7.99 cM in size, but currently Ancestry rounds up so segments above 7.5 will show as 8 on your list. You will need to save those as well, or you will lose at least half of your 8 cM matches.
Moving down to the match in the red box, that person matches on 9 cM, so while they are not officially in jeopardy, I’m taking this opportunity to make sure they are assigned to the Lazarus Estes group. Ancestry didn’t say that they won’t delete any matches over 8 cM, so I’m being careful.
To access the area where you can add this person to a group and make a note, click on their profile picture.
You’ll see your photo, plus theirs and the links to add this match to a group, or to add a note.
Ancestry predicted this match to be 5th-8th cousins, but they are my second cousin twice removed.
Shared Matches is not a preservation method, because Ancestry does NOT show any shared matches below 20 cM, unfortunately.
Common Ancestors equates to ThruLines. Click on the Common Ancestors link to view all of your matches with whom a common ancestor can be identified in one list.
These matches will be presented in the largest to smallest match order, not by ancestor like ThruLines. This makes it easier to just keep scrolling and scrolling to the bottom of the match list where your most distant match that can be identified with a confirmed or potential common ancestor is listed.
If you managed to assign all of the matches to groups from your Thrulines, your smallest “common ancestor” matches should all be assigned to groups. Larger matches aren’t in jeopardy.
I have several pages of people who are in jeopardy. Am I ever glad that I checked.
Use the Shared DNA filter as well to select only shared DNA matches of 6-8 cM in order to save these more rapidly.
It’s hard to believe that Ancestry is actually going to take these matches away from me, even though we share DNA, other matches and common ancestors.
Searching for surnames or unique locations among your matches will provide you with additional hints as to possible relationships. Your connection may be to someone who doesn’t connect via a common ancestor, or the spelling might be slightly different. Matching a surname does not mean that’s how your DNA matches, but it’s a hint and can be especially powerful when combined with locations.
You can combine search terms too. In this case, I combined my unusual Dutch surname of Ferwerda and the location of Leeuwarden and found two people. I confirmed one right away shares my line. I’m working on the second.
Both of these matches would have been lost, yet I share both DNA, confirmed ancestors and shared matches with at least one of these people.
It’s time to get busy. You probably have more matches than you think and you don’t have a lot of time between now and the end of July.
I receive a small contribution when you click on some of the links to vendors in my articles. This does NOT increase the price you pay but helps me to keep the lights on and this informational blog free for everyone. Please click on the links in the articles or to the vendors below if you are purchasing products or DNA testing.
Thank you so much.
DNA Purchases and Free Transfers
- FamilyTreeDNA – Y, mitochondrial and autosomal DNA testing
- MyHeritage DNA – ancestry autosomal DNA only, not health
- MyHeritage DNA plus Health
- MyHeritage FREE DNA file upload – transfer your results from other vendors free
- AncestryDNA – autosomal DNA only
- 23andMe Ancestry – autosomal DNA only, no Health
- 23andMe Ancestry Plus Health
Genealogy Products and Services
- MyHeritage FREE Tree Builder – genealogy software for your computer
- MyHeritage Subscription with Free Trial
- Legacy Family Tree Webinars – genealogy and DNA classes, subscription-based, some free
- Legacy Family Tree Software – genealogy software for your computer
- Charting Companion – Charts and Reports to use with your genealogy software or FamilySearch
- Legacy Tree Genealogists – professional genealogy research