Goodbye False Positives! AncestryDNA Updates its Matching Algorithm

AncestryDNAYesterday, when I checked my AncestryDNA account I had 205 pages of matches, or 10,250 matches in the database. Today, thanks to Ancestry’s new matching algorithm, I have 57 pages of matches, or 2,850 matches in the database (a reduction of 72.2%!). I also have DNA Circles, groups of people who MIGHT share a common ancestor (some of whom share DNA with me).

A New Matching Algorithm

I wrote recently about the new algorithm (see “Finding Genetic Cousins – Separating Fact from Fiction“), as did Judy Russell (“When less is more“) and Roberta Estes (“DNA Day With Ancestry“). In essence, AncestryDNA is leveraging information about allele frequencies (“pile-ups”) in which small IBS segments of DNA shared by many, many people are not utilized for matching.

For example, my results show that I share a segment of DNA with 2,500 other people in the database. My next most common segment isn’t even close, coming in at 400. For both these segments, I do not share the DNA with the other people not because we’re all recently related.

Thus, AncestryDNA is identifying the frequency of certain alleles in the database and using that information to reduce false positives. I certainly don’t want to see those 2,500 matches (at least not if the peak is our only shared DNA), or that group of 400 matches. Eliminating the 3,300 false positives in just those three peaks will be a good start!

NOTE: AncestryDNA is NOT using family trees to find genetic matches in their database. The new matching algorithm does NOT utilize your family tree for matches!

NEW NOTE: Check to see if your match fell down your match list rather than being “lost.” If the new phasing algorithm or matching algorithm eliminated a segment or two, then your match may have moved down your list.

So how much did your match numbers change?

Beware the Naysayers

Undoubtedly there will be some concern about the loss of distant matches, and some of that will be valid until we get a better handle on AncestryDNA’s algorithm, including reading the white paper (beware anyone that hasn’t read the white paper!).

However, automatically complaining about the loss of a match because you’ve “proven” a common ancestor and you have shared DNA is NOT a valid complaint without considerable additional evidence. First, it is not difficult to find a common ancestor with millions of other people, including thousands of people in AncestryDNA’s database (especially if the database has a higher-than-average number of people with large family trees, which I would wager it does), and including with false positives. Second, the way this algorithm works is to eliminate shared segments of DNA from the matching algorithm. That means that you will indeed share a segment of DNA with these false positives, but because it is a pile-up, it will not form the basis of a true match. Sometimes the “pile-up” segment can be quite large, possibly even as high as 10 cM or more (although more information will be required to confirm this).

Indeed, even a triangulated group (even a group with 4 or more people) can be incorrectly assigned to an ancestor! Very few people have ALL their lines documented enough to eliminate the possibility of other matches in missing lines (and no one has yet sufficiently proven that all of their recent family tree is free of hidden errors like NPEs). Roberta Estes just posted about this very phenomenon (see “One Match, Two Ancestors – Never Assume“).

So, if someone complains about the loss of a distant match because they’d found a common ancestor and shared DNA, ask them for more evidence. Genetic genealogy, after all, is a hard science and they should be able to back up their assertion with evidence.

How many genetic matches did you lose? What do you think of the change?

P.S. – to download your list of old matches, go to SETTINGS and on the right-hand panel click on “Download v1 DNA Matches.”

By the way, this is VERY useful, here’s the ranking for your matches, which actually gives some information about how much DNA you likely share with many of your matches:

confidence chart

DNA Circles:

Also new today are DNA Circles, a network of related individuals who have a direct line to a shared ancestor, and who share DNA with at least one other person in the network. The purpose of the DNA Circle is to work together to research and study the common shared ancestor. Indeed, you’ll be able to view the DNA profile page for non-matches if they are in your DNA Circle!

Placement within a DNA Circle is dependent upon you having a PUBLIC family tree, and I fully support this decision. The strength of your placement within the DNA Circle will also be “ranked” dependent on a variety of factors, including how many descendants of the ancestor have been tested.

A few tidbits about DNA Circles:

  • You must have a PUBLIC tree to be placed in a DNA Circle.
  • You will share DNA with at least one person in the DNA Circle, but you might not share with everyone in the DNA Circle. The people you don’t share DNA with will share DNA with someone other than you in the circle.
  • Your relative must be beyond the first cousin once removed stage to be included as a different party in a DNA Circle. In other words, a DNA Circle will not be created for just you and your cousin.
  • DNA Circles are refreshed every 3-4 hours!! This is because new trees are added, trees are updated, necessitating new matching.
  • The DNA Circles algorithm looks back 7 generations for a match, so the more complete your tree the more likely you are to find a match.
  • DNA Circles does require a subscription to Ancestry.com, since it is reliant on family trees.
  • About 50% of those with public trees will have at least one DNA Circle.

Here are my DNA Circles (I have two DNA Circles):

12345

Here are the members of my John Cooper DNA Circle (click to enlarge):

1234567

These are all “Emerging,” meaning that it’s a decent match but that I should be cautious. Note that I am a genetic match to two of my circles, but don’t match the third person.

Here’s what my DNA Homepage looks like now:

123456789

Notice that I have 25 ancestor hints, and 63 4th cousins or closer. I wonder what the relationship between # of hints and likelihood of # of DNA Circles is like?

Be sure to read the White Paper! It’s under “Learn More” at the top of your DNA Homepage, and this will pop up:

12345678

What are your thoughts on DNA Circles? Have you found a new, interesting relatives? How many DNA Circles do you have?

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.

52 Responses

  1. Diane Harman-Hoog 19 November 2014 / 2:34 pm

    The implementation of DNA circles is of no help to adoptees whatsoever as they must have private trees due to the speculative nature of making these trees. I have a 51 cM match that is now ranked as a low 4th cousin! The account I am helping lost 3 matches with common ancestor of 4th -5th great grandparents that were shared with closer matches. I am sure glad I found these before they disappeared as they helped to define which sibling might be the base of my accounts match.

  2. Blaine Bettinger 19 November 2014 / 3:11 pm

    Diane – I wonder if the 3 matches moved down the list instead of being lost entirely? That might happen if some of the shared DNA was due to a phasing error.

    Also, there is no doubt that being placed in a DNA Circle without a tree connection is a future goal of genetic genealogists. This seems like a natural progression to me, and would connect very well with the research that AncestryDNA has been conducting (such as reconstructing ancestor’s genomes).

    • Diane Harman-Hoog 19 November 2014 / 3:22 pm

      No they are gone. Many others on the DNAadoption group are reporting the same thing.
      Unless we can get our matches on Ancestry to upload to gedmatch or FTDNA,, adoptees so not have much success with the tree display method of finding ancestors as most do not have their own tree to match with. Having the larger number of matches was more useful to us as the number of occurrences of a surname through time was an important clue to whether a new branch of the tree was valid or not. Not all clusters of smaller length were IBS, American colonial matches in particular often tended to have these. There is speculation in the genetic community that these may in fact hold important clues as to inheritance. I am sure there is a lot of garbage in their but this may be the equivalent of throwing out the baby with the bathwater, I will be the first to admit that I did not spend much time on most of the smaller matches, although I once had a very small match that provided an important clue, He too is no longer there. However, the frequency of the occurrence of the surname is a whole other matter.
      Thanks for answering Blaine!

  3. Blaine Bettinger 19 November 2014 / 3:26 pm

    Diane – your comment makes the assumption that the frequency of the surnames was due to actual matching. I think we only lost false positives today, which means they were not real matches to begin with. They were potentially leading us down rabbit holes, including misleading adoptees. Sure, some of them were undoubtedly genealogically related, but they should not have shown up as genetic relatives.

    Anyway, thank you so much for reading and I hope we all can discuss this more to determine how well it did or didn’t work!

    • FC 25 November 2014 / 12:16 pm

      That is far from the truth. We also lost actual true matches. I have one that was a 96% confidence match under V1; predicted 4th cousin; range 4th-6th. We share 28.8 continuous CMs on Chr 18 and was able to triangulate with another match on Gedmatch. We have followed the paper trail to our common ancestor. I also have the DNA of my grandparents which I also triangulated this match with and this line comes from GA which is the only line of that grandparent from that area of the country. Plus the paper trail is solid with death certificates, census records, and plantation records.

      I did also lose 2 verified cousins (found ancestors living right next door to each other including paper trail of other records to confirm), but I am less concerned with them bc they were in the 10 CM range and I can accept losing a valid match that low.

      I cannot accept losing a valid match of ~30CMs!

  4. Michael Ward 19 November 2014 / 4:13 pm

    A good step! But it just emphasizes that, without actual match data, we’re just groping through the mist.
    Went from 167 pages to 28; probably all to the good. Added a bunch of new 4th cousins, some with a higher level of certainty. My two circles are actually one couple, with well-documented paper trails (if you have a Hapgood, we’re related … that kind of thing). I would have expected more, since shaking leaves have led me to at least two major linkage points on the Mennonite side (if you have a Ream, we’re related &etc.), and no circles yet for them.

    • Blaine Bettinger 19 November 2014 / 9:25 pm

      I’ll be interested to hear if you get more circles over the next few days! Thank you for sharing the information.

  5. Michael C. Dreimiller 19 November 2014 / 5:33 pm

    Pages of matches 214 -> 18
    Number of matches 10,600 -> 884
    # of Hints 0 -> 0
    3rd-4th cousin matches 1 -> 1
    4th-6th cousin matches 14 -> 9
    * 3 unchanged
    * 6 that moved up from 5th-8th
    * 11 that dropped to 5th-8th, including my previous #3, #4, and #5
    * Biggest move down – #5 (4th-6th, 96%) dropped to #684 (5th-8th, Moderate)
    * Biggest move up – #2311 (5th-8th, Very Low) to #10 (4th-6th Very High)

    Is there a way to download a list of v2 matches like the file you can download for v1 matches?

    • Blaine Bettinger 19 November 2014 / 9:27 pm

      Thank you Michael! That is an enormous reduction in matches! As far as I know there is currently no way to download a list of v2 matches. Frustrating!

  6. Lydia Bishop 19 November 2014 / 6:37 pm

    Too many people don’t really understand even the basics of inheritance and genetics. Ancestry needs to be even more transparent with the DNA samples we send. People are upset because many matches they thought were real, were false positives. The Ancestry.com board in Facebook has been burning hot with whiney people bemoaning the loss of their matches. Some folks have lost some matches that may have been valid, but most of those lost matches were false positives. Over at GEDCom I found a 5.5 degree cousin in Finland of all places. I got very excited until I did a “one on one”. The guy in Finland was a false positive. Even one of my 4.0 degree matches in GedMatch is a false positive. But I do seem to have a match in Finland that’s a “live one.”

    Too many people are not studying up on the most basic aspects of DNA and genealogy.

    • Blaine Bettinger 19 November 2014 / 9:29 pm

      Lydia – I agree, there is a great deal of misunderstanding and in many cases a lack of understanding. On the other hand, this is a really challenging concept, and I understand why people have so much difficulty with it. But hopefully blog posts, Facebook posts, and the white papers will help more people understand what AncestryDNA is doing!

  7. Melanie 19 November 2014 / 8:17 pm

    I have 8 circles! Of course, I lost pages of matches. I’m wondering, what about those very low confidence distant matches that had a common ancestor with me. Did those disappear?

    • Blaine Bettinger 19 November 2014 / 9:30 pm

      8 circles, lucky you! I keep hoping I’ll increase from just 2. Many of the distant matches did disappear, even if you’d identified a common ancestor. Those individuals were actually false positives, according to AncestryDNA’s new algorithm.

  8. connie 19 November 2014 / 8:44 pm

    Thank you for this very informative blog update- it has helped tremendously!

    • Blaine Bettinger 19 November 2014 / 9:31 pm

      Thank you Connie! Thank you for stopping by!

  9. Michael Ward 19 November 2014 / 9:01 pm

    Ancestry seems to have deleted my matches below “moderate,” and according to your chart, that corresponds to <6 cM. Darned right; good practice. Of course, even some of the the matches in the very low probability cases may have been correct; or there may be a paper trail / tree match that's real, even if the DNA "match" was for some other reason. Yet more reasons for match data. I wonder if writing to the board "You're leaving money on the table" would get their attention?

    • Blaine Bettinger 19 November 2014 / 9:33 pm

      Michael – it will be really interesting to see how this all plays out. This is certainly not the last update!

  10. George Jones 19 November 2014 / 10:35 pm

    Blaine, you said: “However, automatically complaining about the loss of a match because you’ve “proven” a common ancestor and you have shared DNA is NOT a valid complaint without considerable additional evidence.”

    AncestryDNA says on these new changes: “Less is More”

    George Jones says to AncestryDNA’s Tim Sullivan: “Hogwash” … “LESS is WAY TOO MUCH LESS”. AncestryDNA did not do any or sufficient customer facing Beta Tests and these PhD inspired changes show it. With a Beta Test of 100 or 500 Customers, AncestryDNA would have easily found out IT’S NOT A BRILLIANT IDEA TO ELIMINATE A CUSTOMER’s STARRED RELATIVES … BUT THEY DID IT ANYWAY! This is analogous to Apple coming out with a IOS software update and erasing all your FAVORITES in the Phonebook.

    So, YES! Blaine, I am complaining about the loss of my “GOOD” “STARRED MATCHES” Who in the heck is AncestryDNA or Tim Sullivan to say othervise. He deserves to lose his job over this. I recommend customers switch from AncestryDNA to FTDNA.

    Blaine, you can try to say this Pig is a Silk Purse but others know that AncestryDNA is not worth the money or the headaches any more.

  11. Cyndy 19 November 2014 / 10:58 pm

    I have nine circles. Yay. Eight of them are on my mother’s side. The one on my father’s side has 9 members. When I search for that surname though, only two users come up. Is the search function not working? I had 262 pages of matches; now I only have 69. Am I happy? No, I am not. I have an unusual maiden name that I would be shocked if anyone reading this has in their matches–Sherfick. All Sherficks are related. The German name was changed by my ancestor, who is buried less than a mile from me. Now when I search for the surname, I only get two matches, even though 9 are in the circle. This has destroyed months of searching for connections. Also, my Germanna matches are all gone except for one. I know I am related to all the Holtzclaws as well as several other surnames that have been documented as being intermarried (Otterbach, Fischbach, Young, etc). My parents even have the same Holtzclaw 8th Grandparents, but they are now gone. I’m not against change. I just don’t want to lose hours of research.

    • Blaine Bettinger 20 November 2014 / 8:40 am

      Cyndy – I always have problems with the search function; I recommend using the Snavely tool to create and download a list of your current matches. I think the tool is still being updated in light of AncestryDNA’s changes, so you might want to give it a few days.

      I would be sure to download your list of V1 matches from the Settings page; I’ve heard they have notes intact, if you made any notes.

      Anyone engaged in research that involves distant cousins, be it at FTDNA, 23andMe, GEDmatch, or AncestryDNA, should be aware that an extraordinarily high amount of those matches are false positives, and you are potentially spinning your wheels. The false positive issue is not unique to any one company. On the other hand, hopefully you made some good genealogical connections even if they weren’t actually genetic connections.

  12. Jason Lee 20 November 2014 / 12:08 am

    “…automatically complaining about the loss of a match because you’ve ‘proven’ a common ancestor and you have shared DNA is NOT a valid complaint without considerable additional evidence.”

    True, but Ancestry set themselves up for this kind of complaint by (1) creating match lists that were overloaded with false positives and (2) suggesting shared ancestry on the basis of those false positives and (3) withholding the evidence that would allow customers to evaluate the validity of the suggested relationships.

    It’s a bit unfair to assert that the customers are making tenuous assumptions based on inadequate evidence when the company has adamantly refused to allow the customers to choose to share that evidence freely.

    • Blaine Bettinger 20 November 2014 / 8:43 am

      Jason – You’re right, this problem was created – in part – by AncestryDNA intentionally choosing to share more false positives. Unfortunately, people would have been just as angry if AncestryDNA was eliminating many real matches. So there was no perfect way from the beginning, but all the companies are taking steps to improve matching.

      Hopefully everyone was already suspicious of their distant cousin matches at ALL the companies to begin with, but I fear that is not the case. I almost never wade down “into the weeds” of distant cousin matching because I know most of these matches are either not real, or are far beyond the genealogically-relevant timeframe.

  13. David Negus 20 November 2014 / 2:03 am

    It is hard to evaluate how much has been gained or lost until Jeff Snavely has a chance to update his AncestryDNA Helper. Using those tools in combination with Ancestry’s relatively new sharing feature has been very productive.
    AncestryDNA Helper allows you to download a CSV of all the ancestors of all the matches for each person who shares with you. Comparing the tables of two or three people from the same line allows one to select only those matches which look most promising, and to avoid time spent on false positives.
    Ancestry’s database has grown so large that for my shared people who were wholly American colonial, I was being timed out before the table was finished downloading. I agree Ancestry couldn’t continue with everyone having as many matches as before.
    It’s not really quite fair to require people to read the white paper before having an opinion. I doubt the white paper is comprehensible (even eliminating some of its obvious typos), to over 10% of Ancestry testers.
    Looking at the Timber program, it appears to me to be flawed for anyone who is not 100% American colonial. Ancestry testers must be American. People with a mix of recent immigrant ancestors and American colonial are not going to have the histogram of matches the program assumes. I have two American colonial and two immigrant grandparents. The parts of my genome where I have two American colonial grandparents on both sides will receive a loss in weight for the prospective matches in that area.
    In my group of shared testers, I have two people with all American colonial ancestors. They dropped from c 350-400 pages to 79 and 115 pages respectively. I dropped considerably more from 267 to 37. I would be interested to know if other people notice the same.
    The new “moderate” is not same size as the old “moderate”. The new “moderate” is the lower end to the old “very low” in size.
    If in fact the program performs at the precision which Ancestry claims it will, it should be useful, even it excludes some results that are actually matches—just not in the same way that AncestryDNA Helper and sharing were.
    AncestryDNA Helper still allows one to download a table of ancestors for each match. Presumably Jeff will be able to rebuild it to download the all the matches’ ancestors into one table.
    My reaction to circles is ho hum. I have two third great grandparents, and one great-great grandparent. The two third great grandparents are married to each other.

    • Blaine Bettinger 20 November 2014 / 8:50 am

      David – I agree, looking at triads using the Snavely tool could potentially be a good way to analyze the new match lists.

      I disagree with your sentiment about the White Paper. I think it is silly for people to complain about the outcome without at least trying to understand the process. If you look around, almost every negative comment is more a knee-jerk reaction to loss of matches than any real comment about the process itself. Indeed, your comments are among the very first I’ve read (and I’ve read hundreds over the past 24 hours!) that even mention the process!

      If you’re ONLY interested in the DNA, then DNA Circles won’t be of much value. For genealogists interested in making non-DNA connections, they are an incredible new tool, and I’ve already heard several success stories. I currently see only limited value from my DNA Circles, but keep in mind that this is just the Beta release. They will continue to improve, both as more people test and as the algorithm is refined. And hopefully, in the not-so-distant future, we’ll be placed in DNA Circles without trees! That will truly be remarkable.

      Great comments, thank you!

      • Terresa Kane 22 November 2014 / 9:00 pm

        My fathers test that I manage has no circles, yes the tree is public, yes it has branches going to the Mayflower, 1700’s, and a few the end in the early 1800’s. He now has 56 pages of “high” confidence matches with no new hints. And the couple that I clicked on, there are no common links.

  14. Debbie Renard 20 November 2014 / 11:26 am

    Pages of matches 202 -> 27
    Number of matches 10,086 -> 1,313
    1st & 2nd cousin matches stayed the same (as expected).
    3rd-4th cousin 3 -> 4
    4th-6th cousin matches 53 -> 46
    Haven’t had time to identify biggest moves, yet, but noticed previous #13 dropped to #185 (he’s also a match at the other two companies), and prior #175 went up to #18.
    So glad they did this! I can much better focus my hard-to-come-by research time, instead of wasting it on misleading pile-ups.

  15. Barb LaFara 20 November 2014 / 1:10 pm

    I went from over 18,000 matches to 3,500, yeah! This is a more manageable number. I had all but quit looking at my results, it felt like looking for a needle in a haystack. My best match is still my best match, my known cousins (1st to 5ths) are still among my matches. I only have one circle, that’s a surprise because there are definitely other circles that could be formed. I did lose some of my ‘starred’ matches, but they are members I could not find a common ancestor with but they were good correspondents.

  16. Todd H 20 November 2014 / 2:29 pm

    I have 14 DNA Circles, which means that there has been a critical mass of descendants of those ancestors who have tested. Most of these are from my Southern lines. I went from 77 leaf hints to 53 after the update. Some folks in my old “3rd cousin” category have slid down to “4th cousin”, including a known 2nd cousin once removed. And I know I have lost more than “false positives”, including a formerly-predicted 4th cousin who in GEDmatch shares two segments with me, 24 and 13 cM. A match that strong should not have disappeared from my list (and this is just one example I’ve identified so far). However, I’m glad other matches that failed to make the previous threshold do indeed appear in my match list with their new algorithm, including a known 3rd cousin once removed who was included in my relative lists both at 23andMe and FTDNA.

  17. Susan Howard 20 November 2014 / 4:49 pm

    Blaine, please let me add my thanks to your explanation, It helped immensely. If only Ancestry had sent their customers this same info before switching over, the may have not gotten so many complaints. I believe the problem is that so many get their DNA results and have no idea what to do with them, beyond what Ancestry provides.

  18. Missy 20 November 2014 / 10:39 pm

    I have 16 DNA Circles! 216 pages down to 75, 125 4th cousins or closer. The best thing is that one of my circles included a relative I hadn’t come across before. We aren’t a DNA match but I checked out our tree matches and we share a 3rd great grandfather. Turns out I went to high school with her! Without the circles, I likely would not have found our connection.

  19. len locascio 21 November 2014 / 3:58 pm

    The circle thing is not working, there is no way i do not have any circles, impossible, I don’t think you guys have perfected this program

  20. Joy 22 November 2014 / 2:27 am

    This “upgrade” must have been designed by John Gruber who thinks we’re all stupid! I have lost 90% of my matches, and of the remaining connections very few have a tree. I lost 5 pages of starred matches who all were related to the same Scottish noble familythat I have researched for years. How dare they remove all my hard work! I am furious! Ancestry.com has lost my faith in their business and I will not renew my subscription. This was a huge mistake and I hope the company gets what they deserve for this disservice to their customers… They lost all credibility as far as I’m concerned.

  21. Lydia Bishop 22 November 2014 / 2:23 pm

    Blain,
    Can you or some else more knowledgeable than I in the ways of DNA enhanced genealogical research explain this puzzle.
    Over on GEDMatch I have a 5-6 degree match with an e-mail address in FINLAND! I did the “one-on-one” comparison and she appears to be a genuine match, not an IBC sort of pile-up match false positive.

    I have NO known ancestry in Finland. Her first name is of Finnish origin.

    Now is this person a real match? Or some other kind of “false positive.” I’d like an educated opinion before I write her an e-mail.

    Ideas please.

    Thank you!

  22. Lydia Bishop 22 November 2014 / 2:28 pm

    I have three circles who match paper trails and one match pulled into one circle is a second cousin to a known cousin I met on Facebook. I’m sure if this 4.1 cousin sends her data to GEDMatch we’ll match there and know exactly on which chromosome we match. My only gripe with Ancestry is that we have to transfer our data to GEDMatch for more detailed study and find other folks who are NOT on Ancestry, but have autosomal data on GEDMatch.

    Having rare, uncommon names in my not too distant past has made matching the paper trails to DNA pretty easy. GEDMatch is the proof.

  23. Donna 24 November 2014 / 9:19 pm

    Where oh where have my hints gone? Two years ago I received my DNA results. I have since reviewed all of those with hints that did not have private trees. I had about 167 in all. I now have only 69. I had 4 screens before and only 2 now. I haven’t contacted Ancestry yet. I am very frustrated about all the time put in but am glad I kept a record of them. These matched at family tree level and supposedly at DNA level, too. What was wrong with that?

  24. Sam 9 December 2014 / 2:50 pm

    I have a tree with 3 tested people and over 70,000 people entered and have maintained copies of all AncestryDNA “hints” since it was in beta. “Hints” come and go and return and re-disappear, sometimes repeatedly. With this new system, I’ve gone from over 500 matches to less than 300. I do have 30 circles.

    I also had over 100 “hints” from the 1600’s, 7th to 8th generations back. All of these are gone. In some cases their are more than 20 separate and distinct decendency lines from one couple. Now, I expect that as time goes by, a number of these will reappear. I do not believe that ALL of these lines are all incorrect, just because AncestryDNA no longer shows matches 8 generations back. Granted a number of these matches will eventually be shown to be false. I disagree with your statement that ALL of them are false.

    • Sam 10 December 2014 / 2:55 pm

      Correction after reviewing over one hundred of more than 150 new “hints,” several 7th generation and at least one 8th generation “hint” remain from the test that I am currently reviewing.

  25. Rico Leffanta 11 December 2014 / 3:01 pm

    Less IS more! We must filter out all that ancestral garbage on the internet, most especially those claiming colonial and/or noble ancestry. As one example, Abraham Byram’s Will clearly states Thomas was NOT his biological son, but the youngest son of his “forad wife”, yet almost all descendants of Thomas claim Abraham as their ancestor and will be very upset when their DNA does not link to Abraham. “Rome was not built in one day”, so we must work with what one knows to be true as technology develops instead of wasting time and resources concocted for vanity/racial/religious or property reasons. Their is more to be gained from re-checking original sources (e.g., Abraham’s Will) than persuing one thousand links to alleged ancestors!

  26. Robert Hobdell 18 December 2014 / 1:03 am

    It is quite apparent Ancestry DNA’s new algorithm is throwing out matches based on their coincidence with genetic pile up locations without regard to how much DNA is involved with any particular match being thrown out. How else could 122, or 50%, of my 2nd through 4th cousin range matches have vaporized? No, these matches did not move elsewhere, they are gone. You can not tell me with a straight face these matches were IBS. Particularly not the ones I grew up with.

    75% of my 5th through 8th cousin matches are gone. Among these were a cluster of matches to one surname mostly in appearing Alabama. I do not care whether or not these were IBS. With these matches I was able to find 3 trees with good documentation leading me back to New Castle Co., Delaware, where I found the correct will to document who my 5th Grandmother’s parents and Grandparents are. If I had relied solely on the stronger 3rd – 4th cousin matches to this surname, I would have been lead down the trail of people who copied a hoax from Rootsweb. I would have concluded erroneously my 6th Great Granddad was a totally fictional person from Chester County, Pennsylvania. So I’ll take true IBS matches any day over high confidence matches to utter garbage. Ancestry DNA should have left well enough alone.

    Ancestry DNA is the big Nay Sayer. Since they believe less is more they will be delighted to do without my well documented & researched trees and my continued subscription.

    • Sam 18 December 2014 / 1:49 am

      AncestryDNA “hints” come and go for various reasons including variations of names, dates and locations. One of my “hints” from 2012 and disappearing in 2013 just came back with the new changes. As time goes by, I expect more of the old “hints” to come back. As far as my Brinsley Barnes B:1715 line only one of my previous 19 “hints” still show up in v-2. But, 2 out of 9 of my 2nd cousin 1R Brinsley Barnes B:1715 “hints” are showing up in the new v-2. So my cousin and I combined only show 3 instead of 28 “hints” for our common ancestor Brinsley Barnes B:1715

  27. Tammy Ofsanik 20 December 2014 / 11:39 am

    1. I do not understand why all of my DNA circles are only for relatives on my father’s side – there is not one DNA circle for relatives on my mother’s side, yet I know my mom has 1st cousins with public trees on ancestry.com. And I know my mom is my biological mom. Any thoughts on this?
    2. Also, if I get my brother to test his DNA, will it give me some results not obtained from my female DNA?
    Thanks

  28. weight Loss Articles msn 16 February 2015 / 10:29 pm

    Remarkable things here. I’m very satisfied to see your post.
    Thanks a lot and I am taking a looik forward to contact you.
    Will you please drop me a mail?

  29. marc 31 March 2015 / 1:24 pm

    Re: Ancestry.Com.’s Confdence intervals
    Although I am glad Ancestry has gotten rid of some of our “garbage” matches, I still wish they would provide us with more information about our matches. In particular, I would like a better breakdown of these matches. For instance, are my “extremely high” – over 30 cm matches a result of large chunks of DNA being shared or many inconsequential IBS chunks. Although some people may not think it makes much of a difference, it matters a lot to me when making the decision as to what matches may hold the most promising results and should be investigated. Ftdna assigns a cousin spectrum that looks promising until you look at the actual results – many of my matches rate as being relatively close in distance, but only because we share many small bits of dna – I don’t consider these to be all that promising in that I probably won’t be able to find a recent common ancestor when our biggest chunk of shared dna is 7 CM long and all the rest are tiny. I agree with Gedmatch’s criteria of sharing at least 3 separate 7cm matches to increase your chances of finding a most recent common ancestor. It would also be great if ancestry.com could give us the ability to see “in common matches” as ftdna does, this would greatly help with the ability to triangulate and narrow down lines of research.

  30. Bert 28 July 2015 / 12:23 am

    I share the same concern as Mark in the previous post. It would be great if AncestryDNA either provided a button or link that would elaborate on the details of the match being viewed or else explicitly defined all of the qualification criteria for each pair of relationship predictions & confidence values that Ancestry DNA may use

    Do you know more about the criteria used for segments included in the 30cM of an “Extremely High” match? Surely they must disregard segments of less than 5 cM?

Comments are closed.