Page 1 of 4 123 ... LastLast
Results 1 to 10 of 39

Thread: For anyone out there who doesn't believe Timber is a bad idea

  1. #1
    Gold Class Member
    Posts
    1,125
    Sex
    Location
    Pittsburgh, PA, USA
    Ethnicity
    Ger.-Brit.-Catalan-more
    Nationality
    (U.S.) American
    mtDNA (M)
    H1bg
    Y-DNA (P)
    R-YP619*

    United Kingdom Germany Bayern Catalonia France Ireland Switzerland

    For anyone out there who doesn't believe Timber is a bad idea

    For those of you who may not know what Ancestry's algorithm "Timber" is, it's designed to remove what Ancestry views as "excessive matching". This is matching caused by not by close recent relatedness but by more distant relatedness common in the population. Common enough, presumably, to increase the chance of numerous smaller segments being so located that they give the appearance of a single larger segment.

    The trouble is, even if Timber does the removing that it's supposed to do, it also does a lot that it shouldn't. Timber eliminates a great deal of real -- and recent -- shared DNA.

    One thing that permits me to see Timber at work -- or, so I believe -- is the fact that my daughter and my wife have also tested at Ancestry, so Ancestry conveniently labels which side one of Kathryn's relatives is on. And, the fact is, 100% of her relatives have to be on one side or the other if they really are her relatives. (They could also be on both sides, but not on neither side.)

    So in going through her matches, I can see that most of them are labeled as to side, but not all. And this is not just true for distant matches, but even some matches in the 4th cousin range are not identified as to side. So far nearly all of them seem to be on my side.

    For example, my daughter has a match with whom she shares 34 cM across 3 segments. This is not a trivial amount. But neither of her parents is included among the shared matches! However, the first two shared matches are the daughters of two of my sisters. The third and fourth shared matches are one of my 1st cousins on my father's side, and one of my 1st cousins once removed on this same side.
    In fact, in total my daughter has thirteen shared matches with this relative of hers who somehow isn't on either parent's side, and all but the last four are clearly identified as being on "Father's side".

    I took a look at that "final four", and every one of them has shared matches who are on my side. (And I believe only on my side.) Yet I myself am not listed as a shared match for any of them. Clearly, I should be, but I'm not.

    While there might be reasons other than Timber that I'm not listed as a shared match to these people, I believe that Timber's suppression of some of my matching is likely the main reason that I'm not. You might ask why Timber isn't also suppressing my daughter's matching, but keep in mind that everyone's DNA consists of two strands. My wife and I are not related at all, so far as I know; and we don't even (so far) seem to have any relatives who are shared by both of us. If we do, they aren't close.

    My parents are also not related to each other, but in my case I definitely have a few matches at Ancestry who share matches with me on each side. That is, they share some matches who are definitely on my father's side, and some matches who are definitely on my mother's side.

    In addition, on my mother's side I have some ancestors who were fairly closely related to their spouses -- the closest being a pair of 1st cousins once removed, but with a couple of 2nd cousins as well. Plus, some of my ancestors on this side came from populations in which endogamy was pretty common -- maybe even typical.

    What I'm suggesting is that because of my "opposite parent" (my mother, in this case), I think Timber may have eliminating some matching that was real -- all the matching I had with the particular individuals showing up for my daughter but not for me, and yet clearly on my side.

    There may be another explanation, but these certainly seem like genuine matches for my daughter, and I can't think of a reason why they wouldn't also be genuine matches for me. (They clearly are not on my daughter's mother's side.)

    EDIT:

    There's also an "easy" fix if Ancestry still thinks Timber is important in eliminating "excess matching". Show the matches, but flag them in some way as "possibly unreliable" and then explain why. Customers would then be able to evaluate these matches.

    But if they simply suppress the matches, so you never see them, how would you ever be able to judge for yourself whether they're "real"?
    Last edited by geebee; 07-19-2019 at 04:52 PM.
    Besides British-German-Catalan, ancestry includes smaller amounts of French, Irish, Swiss, Choctaw & another NA tribe, possibly Catawba. Avatar picture is: my father, his father, & his father's father; baby is my eldest brother.

    GB

  2. The Following 11 Users Say Thank You to geebee For This Useful Post:

     boilermeschew827 (07-20-2019),  JMcB (07-24-2019),  JonikW (07-19-2019),  Mike_G (07-20-2019),  msmarjoribanks (07-21-2019),  parasar (07-19-2019),  Phoebe Watts (07-29-2019),  rms2 (07-19-2019),  Robert1 (07-19-2019),  Rufus191 (08-17-2019),  Webb (07-19-2019)

  3. #2
    Gold Class Member
    Posts
    11,638
    Sex
    Location
    Virginia, USA
    Ethnicity
    British and Irish
    Nationality
    USA
    mtDNA (M)
    U5a2c3a
    mtDNA (P)
    K1a1
    Y-DNA (P)
    R1b-FGC36981

    Wales Ireland Scotland France Bretagne England Switzerland
    I've mentioned this before, but I have a y-dna line 5th cousin 1x removed who squeaks in at Ancestry as a 9 cM/1 segment match but who shows up on FTDNA's Family Finder at 31 cMs with a longest segment of 16 cMs. I know he's my 5th cousin 1x removed because he is a 110/111 y-dna str match, a Big Y SNP match, and because of paper trail genealogy.

    He also shares matches with me to a bunch of my y-line relatives, including my dad, my youngest son, and my youngest daughter.

    I thank God he showed up as a match on Ancestry, because that's how I found and recruited him for further testing. But the thought that he might have been missed because of Timber bothers me.
     


    Hidden Content


    Y-DNA: R1b-FGC36981 (L21> DF13> Z39589> CTS2501> Z43690> Y8426> BY160> FGC36974>FGC36982 >FGC36981)

    Additional Data:
    Lactase Persistent:
    rs4988235 AA (13910 TT)
    rs182549 TT (22018 AA)

    Red Hair Carrier:
    Arg160Trp+ (rs1805008 T) aka R160W

    Dad's mtDNA: K1a1

  4. The Following 5 Users Say Thank You to rms2 For This Useful Post:

     boilermeschew827 (07-20-2019),  JMcB (07-24-2019),  JonikW (07-19-2019),  msmarjoribanks (07-21-2019),  Rufus191 (08-17-2019)

  5. #3
    Gold Class Member
    Posts
    1,125
    Sex
    Location
    Pittsburgh, PA, USA
    Ethnicity
    Ger.-Brit.-Catalan-more
    Nationality
    (U.S.) American
    mtDNA (M)
    H1bg
    Y-DNA (P)
    R-YP619*

    United Kingdom Germany Bayern Catalonia France Ireland Switzerland
    Quote Originally Posted by rms2 View Post
    I've mentioned this before, but I have a y-dna line 5th cousin 1x removed who squeaks in at Ancestry as a 9 cM/1 segment match but who shows up on FTDNA's Family Finder at 31 cMs with a longest segment of 16 cMs. I know he's my 5th cousin 1x removed because he is a 110/111 y-dna str match, a Big Y SNP match, and because of paper trail genealogy.

    He also shares matches with me to a bunch of my y-line relatives, including my dad, my youngest son, and my youngest daughter.

    I thank God he showed up as a match on Ancestry, because that's how I found and recruited him for further testing. But the thought that he might have been missed because of Timber bothers me.
    I think it should both you and all of us. As I said, Ancestry doesn't necessarily have to ditch Timber altogether. Just don't use it to eliminate the match (or any part of it), simply flag the match as perhaps requiring caution. By their choose to simply not report such matches, you never have a chance to evaluate it. And I suspect they know that Timber sometimes eliminates genuine matching. They just aren't too worked up about that. (After all, how will their customers know if they aren't even allowed to see the match?)

    I have one match that used to be in a DNA Circle with me, but Ancestry said we didn't share any DNA. Only, (1) I happen to share DNA with at least one of her siblings; and (2) she shares DNA with me and all five of my siblings at GEDmatch. The amount of DNA she shares with some of my siblings is much more than she shares with me, and she shares with the group of us all told on four or five different segments. (At least a couple of which are held in common by more than one of us.)

    I suspect this particular relative would share DNA with me if not for Timber -- meaning that she actually does share DNA with me and Timber hides it -- and as I recall, before Timber she was included in my DNA match list.

    Finally, as I mentioned in my initial post, my daughter matches numerous people who are on my side but don't match me. (Of course, they don't match my daughter's mother, either.) They do, however, match people who are close relatives of mine -- besides my daughter, that is. Several match one my or both of my nieces, and others match some of my 1st or 2nd cousins.

    The only way my daughter could share DNA with any of these folks is either through her mother or through me, so Ancestry is either showing false positives for her or false negatives for me. (Since it's clear that they're on her paternal side.)

    I wouldn't doubt that there are some folks who match my daughter on her mother's side without matching her mother, but I just haven't seen any yet. But you simply can't share DNA with someone unless you inherited that DNA from either your father or your mother (or more rarely, part from each if your parents are related to each other).

    EDIT:

    Let me "2nd" your joy that you Ancestry didn't suppress all of the match. (And I've seen that, too -- matches bigger for my daughter than for me, yet apparently all from my side. That can be caused by things other than just Timber, but Timber seems to sometimes increase the magnitude of the difference.)
    Last edited by geebee; 07-19-2019 at 11:06 PM.
    Besides British-German-Catalan, ancestry includes smaller amounts of French, Irish, Swiss, Choctaw & another NA tribe, possibly Catawba. Avatar picture is: my father, his father, & his father's father; baby is my eldest brother.

    GB

  6. The Following 3 Users Say Thank You to geebee For This Useful Post:

     JMcB (07-24-2019),  msmarjoribanks (07-21-2019),  rms2 (07-24-2019)

  7. #4
    Gold Class Member
    Posts
    1,655
    Sex
    Location
    Chicago, Illinois
    Ethnicity
    mixed European
    Nationality
    USA
    mtDNA (M)
    K2b2
    Y-DNA (P)
    Dad: R1b/L21/DF63

    Yeah, I think flagging the match but showing it is a good idea, and perhaps giving the match amount with and without Timber. I also wish they'd show common matches that are less closely-related. I have a Welsh match and we have no common matches, but I suspect that if they'd show more distant common matches I could figure out which of my most recent Welsh ancestors he connects to. This would be very helpful for him, as my Welsh ancestors who came in the 1840s came from small villages. The match traces his family back only to the next generation (so 1870s or after) and they are already in more urban areas making it very hard to trace given how common Welsh names tend to be. If we could pinpoint which lines we are connected on more specifically, that would be very helpful in pushing his family back to the village (and there with church records and some wills it's easier to get back somewhat farther).

    I recently noticed that DNA painter marks common pile-up areas, which is something I wish all the services with chromosome browsers would do. (Of course, I also wish Ancestry would add a chromosome browser, but oh well.)
    Last edited by msmarjoribanks; 07-21-2019 at 04:48 PM.

  8. The Following 4 Users Say Thank You to msmarjoribanks For This Useful Post:

     FionnSneachta (07-21-2019),  JMcB (07-24-2019),  Phoebe Watts (07-29-2019),  rms2 (07-24-2019)

  9. #5
    Gold Class Member
    Posts
    1,125
    Sex
    Location
    Pittsburgh, PA, USA
    Ethnicity
    Ger.-Brit.-Catalan-more
    Nationality
    (U.S.) American
    mtDNA (M)
    H1bg
    Y-DNA (P)
    R-YP619*

    United Kingdom Germany Bayern Catalonia France Ireland Switzerland
    Quote Originally Posted by msmarjoribanks View Post
    Of course, I also wish Ancestry would add a chromosome browser, but oh well.
    Besides British-German-Catalan, ancestry includes smaller amounts of French, Irish, Swiss, Choctaw & another NA tribe, possibly Catawba. Avatar picture is: my father, his father, & his father's father; baby is my eldest brother.

    GB

  10. The Following 2 Users Say Thank You to geebee For This Useful Post:

     msmarjoribanks (07-22-2019),  rms2 (07-24-2019)

  11. #6
    Gold Class Member
    Posts
    11,638
    Sex
    Location
    Virginia, USA
    Ethnicity
    British and Irish
    Nationality
    USA
    mtDNA (M)
    U5a2c3a
    mtDNA (P)
    K1a1
    Y-DNA (P)
    R1b-FGC36981

    Wales Ireland Scotland France Bretagne England Switzerland
    Sorry to sound overly religious, but I am truly grateful to God for the match that lead to the y-dna test that confirmed who my 5th great grandfather is. If that bothers anyone, oh well. It really is nothing short of a miracle.

    If I get no other matches during the short (or long) time I have left, I'll still be grateful.

    I'm still blown away that it happened.

    Glad I did all this testing. I'd spend ten times as much to find out what I have.
     


    Hidden Content


    Y-DNA: R1b-FGC36981 (L21> DF13> Z39589> CTS2501> Z43690> Y8426> BY160> FGC36974>FGC36982 >FGC36981)

    Additional Data:
    Lactase Persistent:
    rs4988235 AA (13910 TT)
    rs182549 TT (22018 AA)

    Red Hair Carrier:
    Arg160Trp+ (rs1805008 T) aka R160W

    Dad's mtDNA: K1a1

  12. The Following 2 Users Say Thank You to rms2 For This Useful Post:

     JMcB (07-24-2019),  Mike_G (07-28-2019)

  13. #7
    Gold Class Member
    Posts
    1,017
    Sex
    Location
    Pittsburgh, PA
    Ethnicity
    100% European
    Nationality
    American
    mtDNA (M)
    U5a1a2a
    Y-DNA (P)
    DF27>CTS4065>FG23196

    United States of America United Kingdom Germany Ireland Scotland Wales
    Well, all the money I have spent still hasn't helped me deduce who my great great grandpappy was. It has actually created more questions than answers. Don't get me wrong, I don't regret any of the testing I have done and would do it all over again in a second, but my grandfather is 91, and I really started all of this for him because this has been a passion of his since he was a boy and he greatly admired his father, my great grandfather. But time is running out and I haven't gotten any closer. I don't even know if he was who he said he was. I have some very good second and third cousin matches who are descended from my great grandfather's sister. I have used people's trees to make notes on the shared matches who are related to my great grandfather and his sister's mother. This way I can concentrate on the people who I have no idea who they are, but no luck. I have figured out that one shared match is related to my third cousin through her father's family and related to my mother's family so he appears to be a Webb, but is actually not.

  14. #8
    Registered Users
    Posts
    10
    Sex

    I have 12% Ashkenazi Jewish DNA, and on both FTDNA and My Heritage, most of my DNA matches are Jewish people, estimated as much more closely related to me than they actually are, whereas on Ancestry the proportion of Jewish matches fits with the 12% of DNA, so if it is Timber which does that, I am all for it.

  15. #9
    Gold Class Member
    Posts
    1,125
    Sex
    Location
    Pittsburgh, PA, USA
    Ethnicity
    Ger.-Brit.-Catalan-more
    Nationality
    (U.S.) American
    mtDNA (M)
    H1bg
    Y-DNA (P)
    R-YP619*

    United Kingdom Germany Bayern Catalonia France Ireland Switzerland
    Quote Originally Posted by kiterunner View Post
    I have 12% Ashkenazi Jewish DNA, and on both FTDNA and My Heritage, most of my DNA matches are Jewish people, estimated as much more closely related to me than they actually are, whereas on Ancestry the proportion of Jewish matches fits with the 12% of DNA, so if it is Timber which does that, I am all for it.
    If Timber eliminates false matches for some populations at the expense of others, that still isn't right. This is especially true when all Ancestry has to do is use Timber to flag suspicious matches -- as opposed to simply hiding them. The amount Ancestry doesn't report that other companies do can at times be pretty substantial.
    Besides British-German-Catalan, ancestry includes smaller amounts of French, Irish, Swiss, Choctaw & another NA tribe, possibly Catawba. Avatar picture is: my father, his father, & his father's father; baby is my eldest brother.

    GB

  16. The Following 4 Users Say Thank You to geebee For This Useful Post:

     Mike_G (07-28-2019),  msmarjoribanks (07-28-2019),  rms2 (07-28-2019),  Rufus191 (08-17-2019)

  17. #10
    Gold Class Member
    Posts
    1,655
    Sex
    Location
    Chicago, Illinois
    Ethnicity
    mixed European
    Nationality
    USA
    mtDNA (M)
    K2b2
    Y-DNA (P)
    Dad: R1b/L21/DF63

    Quote Originally Posted by Webb View Post
    Well, all the money I have spent still hasn't helped me deduce who my great great grandpappy was. It has actually created more questions than answers. Don't get me wrong, I don't regret any of the testing I have done and would do it all over again in a second, but my grandfather is 91, and I really started all of this for him because this has been a passion of his since he was a boy and he greatly admired his father, my great grandfather. But time is running out and I haven't gotten any closer. I don't even know if he was who he said he was. I have some very good second and third cousin matches who are descended from my great grandfather's sister. I have used people's trees to make notes on the shared matches who are related to my great grandfather and his sister's mother. This way I can concentrate on the people who I have no idea who they are, but no luck. I have figured out that one shared match is related to my third cousin through her father's family and related to my mother's family so he appears to be a Webb, but is actually not.
    My current quest is similar -- I'd like to know who my ggg-grandfather on one line is (although no one else in the family is deeply interested -- my mother and grandmother would have been, but both died assuming he was who they had been told he was, which could be but there's a lot of mystery there).

    All my best matches on this line (descendants of my gg-grandmother) are on Ancestry and most have not done Gedmatch, so I am going to have to be more active in trying to convince people to upload to Gedmatch. If Ancestry would only have a chromosome browser I think I'd be farther along in solving this mystery.

  18. The Following 2 Users Say Thank You to msmarjoribanks For This Useful Post:

     FionnSneachta (07-28-2019),  Webb (07-29-2019)

Page 1 of 4 123 ... LastLast

Similar Threads

  1. AC/Admix on the X Chromosome (doesn't match Hap)
    By Unfounded in forum X Chromosome (X-DNA)
    Replies: 7
    Last Post: 08-13-2019, 04:10 PM
  2. So who has steppe and who doesn’t?
    By Censored in forum Southern
    Replies: 20
    Last Post: 06-28-2018, 02:21 PM
  3. Replies: 7
    Last Post: 08-11-2015, 03:47 PM
  4. Replies: 11
    Last Post: 08-05-2015, 02:17 AM
  5. Exercise Doesn't Help with Weight Loss?
    By Scarlet Ibis in forum Health and Fitness
    Replies: 20
    Last Post: 08-26-2012, 04:12 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
  •