Page 1 of 2 12 LastLast
Results 1 to 10 of 15

Thread: New Beta version of shared matches

  1. #1
    Registered Users
    Posts
    11
    Sex

    New Beta version of shared matches

    I haven't seen an announcement of this anywhere yet, but if you have enabled the New & Improved DNA Matches in Ancestry Lab, then when you view the "Compare" screen for a DNA match and then click on "View all shared matches" in the Shared Matches section of that page, it takes you to a list which may be longer than the list you have previously seen for that person. I think this only started today.

    They are all still estimated 4th-6th cousins or closer to you, but I guess that it is now including any of those who are anywhere on the other person's match list, rather than having to be also estimated 4th-6th cousins to them as well.

    So you may see some who seem to come from a different part of your tree, as they may only share a few cM with the other person. (For instance, I am getting some people who are related on my Mum's side coming up as shared matches with people who are related on my Dad's side. Of course there were a few of those before, but more now.)

    May be worth having another look at those who had "no shared matches" last time you looked.

  2. The Following 5 Users Say Thank You to kiterunner For This Useful Post:

     FionnSneachta (03-20-2019),  Judith (03-22-2019),  kujira692 (03-20-2019),  msmarjoribanks (03-22-2019),  pmokeefe (03-20-2019)

  3. #2
    Registered Users
    Posts
    1,020
    Ethnicity
    Northern Ireland
    Nationality
    Northern Irish
    Y-DNA (P)
    R1b-L193(BY2634)

    Northern Ireland Ireland
    Yep very good. I have some that share with my father and mothers side.

  4. #3
    Registered Users
    Posts
    11
    Sex

    They seem to have turned it off again now.

  5. The Following User Says Thank You to kiterunner For This Useful Post:

     pmokeefe (03-20-2019)

  6. #4
    Gold Class Member
    Posts
    463
    Sex
    Location
    England
    Ethnicity
    Northern English
    Nationality
    British
    Y-DNA (P)
    Brother I2-M26 I2a1a
    mtDNA (M)
    H4a1a1a

    On Friday 22nd it was available and seems to work OK but some improvements would be useful and I have left detailed feedback for them, small stuff like virtually identical colours in the groups. I suspect having paternal matches showing in common with maternal is inevitable in small communities (once you include matches of matches). My lines are from different geographic areas so there has been little marrying between the two sides over 100s of years. I found one potential match from both sides, from our trees, but she was not a DNA match.
    However within my paternal lines there is a lot of intertwining, two bothers marrying two sisters etc but no tree collapses from cousin marriage on that geographically close side, (yet! with ~7 generations so far). My maternal tree has just one 2nd cousin marriage. That will always fool a matches of matches approach. The only true solution is a chromosome browser and I wonder if they will ever do that?

    My lack of marriages between lines seems to be mostly because of the mobility in the early stages of the industrial revolution that there is little intermarriage but further back within small villages there must have been more but I have not found much yet.
    Image “Westray wifie” replica of Neolithic figurine Hidden Content
    Out of 64 pre 1800 births 45% Cheshire, 1% Irish (or Scottish), 25% south Derbyshire, 13% Burton on Trent area (where 4 counties within 10 miles), 7% Shropshire, 1% Staffs, 8% Lancs. So far all British Isles despite what some testing companies say.

  7. The Following User Says Thank You to Judith For This Useful Post:

     msmarjoribanks (03-22-2019)

  8. #5
    Registered Users
    Posts
    329
    Sex
    Location
    Wales
    Ethnicity
    Welsh
    Nationality
    Welsh
    Y-DNA (P)
    R-L21 DF13 (father)
    mtDNA (M)
    J1c9

    Wales England
    I saw the different version for a few hours on about the 20th. There were no differences in format, just more content.

    I looked a few of my puzzling entries before that version disappeared. I have a few plausible “high” matches (35 cM over 3 segments for example) with no shared matches. Our ancestors lived in roughly the same places so they seemed worth looking into but there was no obvious connection. With the different version, there were quite a few shared matches on their lists, almost all of them I had thought were potentially coincidental matches to me. It would be interesting to know whether there is to be a new version, and what it will do differently.
    All 32 3xgreat grandparents were Welsh. Two 6xgreat grandparents from England and a few Irish or English surnames before 1800. Paper trail shows several C11th to C14th Anglo-Norman lines and C11th Norse-Irish lines.

  9. The Following 3 Users Say Thank You to Phoebe Watts For This Useful Post:

     FionnSneachta (03-22-2019),  msmarjoribanks (03-22-2019),  Robert1 (03-29-2019)

  10. #6
    Registered Users
    Posts
    802
    Sex
    Ethnicity
    Irish
    Nationality
    Irish
    Y-DNA (P)
    R-FGC6545-Y160102
    mtDNA (M)
    H6a1b2*
    mtDNA (P)
    H27e

    Ireland
    I signed up for the beta matches at lunch the other day. I saw a few interesting shared matches that were providing hints for connections but said I'd go through them properly in the evening after work. Unfortunately by the time I got home, the extra matches were gone. That was a bit sickening. I was ready to send on my theories to one of my matches but I can't now without anything to back it up. All I can remember is the theory that I formed from seeing the shared matches but not the shared matches themselves.
    Ancestry: Ireland (Paper trail ≅ 81.25% Roscommon, 12.5% Galway, 6.25% Mayo)
    Y-DNA (M) ancestor (Y): Kelly b. c1830 in Co. Roscommon (Uí Maine)
    mtDNA (P) ancestor: Fleming b. c1831 in Co. Roscommon
    mtDNA (M) ancestor: McDermott b. c1814 in Co. Roscommon
    Paternal great grandfather (mt): Connella b. c1798 in Co. Roscommon (T2a1a8)

  11. The Following 3 Users Say Thank You to FionnSneachta For This Useful Post:

     Judith (03-28-2019),  msmarjoribanks (03-22-2019),  Phoebe Watts (03-22-2019)

  12. #7
    Gold Class Member
    Posts
    925
    Sex
    Location
    Brisbane
    Nationality
    Australian
    Y-DNA (P)
    T-P322 (T1a2b1)
    mtDNA (M)
    H6a1

    Australia Cornwall England Scotland Germany Poland
    Elsewhere has been the observation that the shared match threshold fell from 20 cM to 3 or even lower.
    It seems this happened when AncestryDNA was introducing some other tweak.
    And this threshold change happened by accident and in error.
    And was "corrected" after about 24 hours.

    I know from looking at matches who have transferred elsewhere, that for me, gold dust lies between 20cM and 10cM at AncestryDNA.
    (Timber has already trimmed around half of these segments. Now they do the disservice of ignoring them as below 20cM for Shared Matches.)

    My experiments with extending clustering limits down to 15cM show the emergence of much more meaningful Shared Match data just below 20cM for me.
    A whole lot of clusters that meant nothing start coming together.
    Other users will have different points at which this occurs. Some are already happy.
    Many however, would be delighted if they could select a lower Shared Match setting for themselves.

  13. The Following 3 Users Say Thank You to Saetro For This Useful Post:

     Judith (03-25-2019),  msmarjoribanks (03-25-2019),  Phoebe Watts (03-26-2019)

  14. #8
    Registered Users
    Posts
    329
    Sex
    Location
    Wales
    Ethnicity
    Welsh
    Nationality
    Welsh
    Y-DNA (P)
    R-L21 DF13 (father)
    mtDNA (M)
    J1c9

    Wales England
    Quote Originally Posted by Saetro View Post
    Elsewhere has been the observation that the shared match threshold fell from 20 cM to 3 or even lower.
    It seems this happened when AncestryDNA was introducing some other tweak.
    And this threshold change happened by accident and in error.
    And was "corrected" after about 24 hours.

    I know from looking at matches who have transferred elsewhere, that for me, gold dust lies between 20cM and 10cM at AncestryDNA.
    (Timber has already trimmed around half of these segments. Now they do the disservice of ignoring them as below 20cM for Shared Matches.)

    My experiments with extending clustering limits down to 15cM show the emergence of much more meaningful Shared Match data just below 20cM for me.
    A whole lot of clusters that meant nothing start coming together.
    Other users will have different points at which this occurs. Some are already happy.
    Many however, would be delighted if they could select a lower Shared Match setting for themselves.
    That is interesting.

    Yes over 10cM would catch many useful matches for me too. But actually, bringing the threshold down to the bottom of “good” (16cM?) would make a huge difference. It doesn’t really make sense not to report these if there is a good likelihood that they are real.

    I think I read somewhere that the threshold was set at 20cM because of the numbers of matches were becoming to much for American testers. So the facility to choose a lower threshold seems sensible.
    Last edited by Phoebe Watts; 03-26-2019 at 10:52 PM.
    All 32 3xgreat grandparents were Welsh. Two 6xgreat grandparents from England and a few Irish or English surnames before 1800. Paper trail shows several C11th to C14th Anglo-Norman lines and C11th Norse-Irish lines.

  15. The Following 4 Users Say Thank You to Phoebe Watts For This Useful Post:

     Judith (04-10-2019),  msmarjoribanks (03-26-2019),  Robert1 (03-29-2019),  Saetro (03-26-2019)

  16. #9
    Registered Users
    Posts
    329
    Sex
    Location
    Wales
    Ethnicity
    Welsh
    Nationality
    Welsh
    Y-DNA (P)
    R-L21 DF13 (father)
    mtDNA (M)
    J1c9

    Wales England
    Quote Originally Posted by kiterunner View Post
    I haven't seen an announcement of this anywhere yet, but if you have enabled the New & Improved DNA Matches in Ancestry Lab, then when you view the "Compare" screen for a DNA match and then click on "View all shared matches" in the Shared Matches section of that page, it takes you to a list which may be longer than the list you have previously seen for that person. I think this only started today.

    They are all still estimated 4th-6th cousins or closer to you, but I guess that it is now including any of those who are anywhere on the other person's match list, rather than having to be also estimated 4th-6th cousins to them as well.

    So you may see some who seem to come from a different part of your tree, as they may only share a few cM with the other person. (For instance, I am getting some people who are related on my Mum's side coming up as shared matches with people who are related on my Dad's side. Of course there were a few of those before, but more now.)

    May be worth having another look at those who had "no shared matches" last time you looked.
    This facility seems to be there again.

    Edit: no, it’s gone again
    Last edited by Phoebe Watts; 04-10-2019 at 06:49 PM. Reason: Update
    All 32 3xgreat grandparents were Welsh. Two 6xgreat grandparents from England and a few Irish or English surnames before 1800. Paper trail shows several C11th to C14th Anglo-Norman lines and C11th Norse-Irish lines.

  17. #10
    Gold Class Member
    Posts
    463
    Sex
    Location
    England
    Ethnicity
    Northern English
    Nationality
    British
    Y-DNA (P)
    Brother I2-M26 I2a1a
    mtDNA (M)
    H4a1a1a

    Today at about 6pm UK time shared matches went below the 20cM level. From other poster’s experience I knew it might not last so did some quick screen grabs and put the matches into my groups.
    Then I left it and ate supper and after which it had gone back to 20cM only about an hour later. SHAME!

    So if you see your shared matches go below 20cM into remote cousins jump fast! I grabbed 5 lines with brick walls in the few minutes I had.
    Image “Westray wifie” replica of Neolithic figurine Hidden Content
    Out of 64 pre 1800 births 45% Cheshire, 1% Irish (or Scottish), 25% south Derbyshire, 13% Burton on Trent area (where 4 counties within 10 miles), 7% Shropshire, 1% Staffs, 8% Lancs. So far all British Isles despite what some testing companies say.

Page 1 of 2 12 LastLast

Similar Threads

  1. Confusion surrounding shared matches
    By Nqp15hhu in forum AncestryDNA
    Replies: 5
    Last Post: 01-03-2019, 10:51 PM
  2. Yfull: Shared SNPs vs Assumed Shared SNPs?
    By Dibran in forum R1a-Z283
    Replies: 1
    Last Post: 05-26-2018, 12:29 AM
  3. Weirdness with shared matches
    By Loderingo in forum AncestryDNA
    Replies: 2
    Last Post: 04-16-2018, 09:36 PM
  4. Utilizing Shared Matches Tool
    By greerpalmer in forum AncestryDNA
    Replies: 1
    Last Post: 08-01-2017, 06:09 PM
  5. New Version of Big-Y
    By R.Rocca in forum FTDNA
    Replies: 137
    Last Post: 04-25-2016, 01:48 PM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •