AncestryDNA’s new BETA

[peter]

I often think that those linked back on "higher" links are more useful - as it can confirm research you have done already and for a lot of my closer relatives, I already know a lot about them!

I agree - whilst it's good to re-establish links with 2nd and 3rd cousins, it's the matches with 4th, 5th and 6th cousins which I find most useful.

 

[Helen7]

I currently have 39 in my 'common ancestors' list, 8 of these at 4th cousin or closer. (I have 214 '4th cousin' matches in total).
My husband has 26 with common ancestors, 10 listed as 4th cousin or closer. (He has 174 '4th cousin' matches in total).

Of course this is how Ancestry estimates the relationships. In reality some of those listed as 'distant cousins' are actually 3rd or 4th cousins (7 of those in my case). And as others have mentioned, the list does not included cousins who have no linked tree. However, I do find it slightly puzzling that 3 people with 'shared ancestor hints' on the old display have never appeared on my common ancestors list. I also wonder why 3 people previously listed on my common ancestors list have now disappeared from that list (after I'd confirmed the connection).

However, all in all a thumbs up for the new Beta as it is showing me connections I would never have spotted otherwise.

And I agree with jorghes and Peter, matches with more distant cousins often prove more useful.

 

[PhoebeW]

However, I do find it slightly puzzling that 3 people with 'shared ancestor hints' on the old display have never appeared on my common ancestors list. I also wonder why 3 people previously listed on my common ancestors list have now disappeared from that list (after I'd confirmed the connection).

I think Shared ancestors will tend to come and go for now because they depend on ThruLines, which isn’t yet stable.

The algorithm for the old common ancestor hints was quite different though wasn’t it? Some of mine were included because names and localities were similar although there was no “line” to them. I’m almost sure it did include matches with linked private unsearchable trees; those have gone, and I think that’s appropriate.

 

[Helen7]

I’m almost sure it did include matches with linked private unsearchable trees; those have gone, and I think that’s appropriate.

I agree, but the ones I'm talking about have public trees and the common ancestor is easy to see. One of these was wrong (same name and locality but wrong family) but the other two were correct with a clear line that I verified, and I've been in touch with both of these cousins well before the Beta appeared. In that sense the omissions haven't hampered my research, I just wonder why they are not included in the common ancestors list.

 

[PhoebeW]

I agree, but the ones I'm talking about have public trees and the common ancestor is easy to see. One of these was wrong (same name and locality but wrong family) but the other two were correct with a clear line that I verified, and I've been in touch with both of these cousins well before the Beta appeared. In that sense the omissions haven't hampered my research, I just wonder why they are not included in the common ancestors list.

That’s interesting - I’ll check mine too.

 

[PhoebeW]

I found one entry in the old format that should qualify for shared ancestor in Beta. There is an error in her tree though so that can easily be sorted.

 

[jorghes]

I went back to have a look at the old Shared Ancestor hints with a couple of thoughts in my mind about why certain Ancestors may have a "Shaky Leaf" but not a "Common Ancestor".
I had two suppositions - that since ThruLines only goes up to an individual's 5th great grandparents, that any Shaky Leaf hint with a linkage above 5th g-grandparents would not get given a "Common Ancestor" hint (reasonably easy to test, two of my tests had "Shaky Leaves" for 7th and 8th cousins, linked through g-grandparents higher than 5x) and the other of course is that if someone has a private tree which is also unsearchable, they will show for a "Shaky Leaf", but not in ThruLines, as that relies on the private tree still being searchable.

And I found exactly what I was looking for- while adding "SL" comments on the notes I had made for each person with a Shaky Leaf hint, I came across three people who I thought would not have a "Common Ancestor" all of which suggested a link above 5x great grandparents (I wrote the specific relationship in the note, so I retain that). A couple of the results are shared across my father and paternal grandmother's results, but for both in the range of 7th-8th cousins, and one from my paternal line for my father with a 7th/8th cousin link. And when I turned the beta back on and went to have a look, none of those 3 test results came back with a "Common Ancestor" link.

As for the second theory - private and unsearchable trees, it's harder to prove, but I also had two Shaky Leaf hint matches both with private trees who do not appear as "Common Ancestors" where the link is 5x g-grandparents or lower. As I have a few "Common Ancestor" suggestions with private trees and I can see where the link is thanks to ThruLines, I assume that those two results have kept their tree unsearchable.

 

[peter]

I had two suppositions - that since ThruLines only goes up to an individual's 5th great grandparents, that any Shaky Leaf hint with a linkage above 5th g-grandparents would not get given a "Common Ancestor" hint

The problem is that once you get beyond a certain point we're all related, often many times over, so assuming that a matching segment was inherited from an specific ancestor pair is a bit risky.

 

[Helen7]

I found one entry in the old format that should qualify for shared ancestor in Beta. There is an error in her tree though so that can easily be sorted.

I had two suppositions - that since ThruLines only goes up to an individual's 5th great grandparents, that any Shaky Leaf hint with a linkage above 5th g-grandparents would not get given a "Common Ancestor" hint... and the other of course is that if someone has a private tree which is also unsearchable

I wouldn't really expect any linkage to be shown above 5x g-grandparents - nor would I expect a DNA match at that level, though I know it is possible.
However, I've now found that one of the missing shared ancestor hints has turned up in the common ancestors list (the Beta is clearly still rather fluid), but the remaining absentee is not a distant relative - we share 2x great-grandparents. He has a public tree and I can't see any obvious errors in it, so it's a mystery why he is still omitted from the common ancestors list (and also ThruLines of course).

 

[DavidF]

I know I'm a late comer to this thread but what I have noticed with the results that surprises me is how variable the Ancestry 'likely relationship' is compared to the actual relationship.
Here are some examples: I have matched 9 x 2nd cousins - Ancestry has these as 5 x 2nd cousin; 3 x 3rd cousin; 1 x 4th cousin. I have matched 9 x 2nd cousin once removed - Ancestry has these as 1 x 2nd cousin; 2 x 3rd cousin; 5 x 4th cousin; 1 x distant cousin.
Clearly this is not a very exact science.

 

[jorghes]

DavidF - I believe that is because the amount of DNA that you can share with your cousins is highly variable. For example, my brothers and I do not share the same amount of DNA with our grandmother, one of my brothers has a large proportion of DNA that matches.

As this is variable, then Ancestry is probably best to give approximate relationships rather than trying to give exact estimates.

 

[PhoebeW]

I know I'm a late comer to this thread but what I have noticed with the results that surprises me is how variable the Ancestry 'likely relationship' is compared to the actual relationship.
Here are some examples: I have matched 9 x 2nd cousins - Ancestry has these as 5 x 2nd cousin; 3 x 3rd cousin; 1 x 4th cousin. I have matched 9 x 2nd cousin once removed - Ancestry has these as 1 x 2nd cousin; 2 x 3rd cousin; 5 x 4th cousin; 1 x distant cousin.
Clearly this is not a very exact science.

The new analysis of probability of relatedness that came with the DNA matches Beta is a great improvement and shows the range of possible relatedness for each match.

You might find some of your relatives are being suggested as more distant relatives than the paper trail says. That might be because Ancestry adjusts the size of the match. I’m affected by this on two branches. My 3rd cousins show up on Ancestry as about 9cM matches - unlikely for 3rd cousins. But taking them over to GEDmatch or MyHeritage shows that the match is larger.

 

[peter]

I know I'm a late comer to this thread but what I have noticed with the results that surprises me is how variable the Ancestry 'likely relationship' is compared to the actual relationship.
Here are some examples: I have matched 9 x 2nd cousins - Ancestry has these as 5 x 2nd cousin; 3 x 3rd cousin; 1 x 4th cousin. I have matched 9 x 2nd cousin once removed - Ancestry has these as 1 x 2nd cousin; 2 x 3rd cousin; 5 x 4th cousin; 1 x distant cousin.
Clearly this is not a very exact science.

If you study the chart in my Masterclass you'll see that the amount of DNA shared between cousins of a given degree varies enormously - indeed, if it didn't vary we'd never match anyone more distant than a 4th cousin once removed, or half 4th cousin, whereas in practice almost all of our matches are with cousins who are more distant (see the table in the Masterclass).

Ancestry now provide detailed statistics on each match showing the likelihood of different relationships (based on actual matches); but the information in the chart is probably easier to digest.

 

[Katie Bee]

I am trying to understand the DNA matches!
I still cannot use ThruLines, so I will ask Ancestry support why that has not been activated.
I have a private, searchable tree, I have enabled New & Improved DNA Matches Beta.
When I view all DNA matches I have a 3rd - 4th Cousin and from their public tree I can see that we have the same 2xgreat grandparents.
When I click on Compare, Ancestry tells me that I do not share any common ancestors.
Comparing our trees, our 2x great grand father has the same name, birthdate and place of birth and marriage date and place of marriage.
Our 2xgreat grandmother I have Ann, they have Anne, I don't have an exact birthdate, they do.
Is this enough to make Ancestry ignore our common ancestors or am I missing something?
Reading all your comments on ThruLines, I did not think Ancestry would be so particular!

 

[Helen7]

Comparing our trees, our 2x great grand father has the same name, birthdate and place of birth and marriage date and place of marriage. Our 2xgreat grandmother I have Ann, they have Anne, I don't have an exact birthdate, they do.

I would think Ancestry should show at least the 2x g-grandfather as a common ancestor if all the details match as you say, and it seems strange that the very slight discrepancies in your common 2x g-grandmother would cause a problem. Do you have other DNA matches showing common ancestors? If not, your lack of ThruLines might be the problem, as the 'new and improved DNA matches' seems to be interlinked with ThruLines.

I looked again at my 3rd cousin's tree I mentioned in #169 above, and I see his tree does not show any dates for our common 2x g-grandparents (though their names fully match and are quite unusual), so I'm guessing this lack of dates is why Ancestry tells me we don't share any common ancestors.

 

[Katie Bee]

Do you have other DNA matches showing common ancestors? If not, your lack of ThruLines might be the problem

I haven't found any matches that show common ancestors.
I have asked Ancestry why I cannot get to Thrulines, so we will hopefully find out what the problem is.
Thanks for your reply Helen7

 

[Katie Bee]

I haven't found any matches that show common ancestors.
I have asked Ancestry why I cannot get to Thrulines, so we will hopefully find out what the problem is.

I have had a reply from Ancestry.
"I can confirm that the reason you do not have any ThruLines is because the way you have your tree set up that your DNA is linked to. ThruLines looks at your name in the tree and then follows down the generations and compares them to other peoples trees in your DNA matches. In your case you have the name of yourself as "Private person 1" then your parents as "Private person 2 and 3". This would not match up with your DNA name or the name of ancestors in other peoples trees. You would be required to enter the real names of yourself and your parents in your tree for you to be able to avail of ThruLines. I hope this answers your query."

I used FTAnalyser to create a minimalist DNA GEDCOM and when it asked "Do you want living people replaced with 'Private Person' and their details hidden?" I replied YES.
So I have updated my tree and we will see if ThruLines appears

For the record the few days was of the order of about 4.

I will let you know if it is any quicker for me.

 

[PhoebeW]

Has anyone else seen their ThruLines become far more accurate?

By today, I have only one missing ancestor and only two replaced ancestors. The quality of the suggested ancestors is a bit better too.

New shared ancestor hints are trickling in. They seem to be recent emigrants who have documented ancestry in the UK.

 

[Tim]

I used FTAnalyser to create a minimalist DNA GEDCOM and when it asked "Do you want living people replaced with 'Private Person' and their details hidden?" I replied YES.
So I have updated my tree and we will see if ThruLines appears

A great new feature from Alexander, I also used it but I said No, as Ancestry already privatises living names. I should load a new tree, with all the new matches I've made.

 

[Pauline]

Has anyone else seen their ThruLines become far more accurate?

To some extent, but not as much as yours. I’m OK up to 2x great grandparents, then I have 3 replaced 3x gt grandparents. For the 4x and 5x greats I have quite a few replaced ancestors, almost half of the 5x.

But I’m faring a lot better than my brother in law who still has a replacement for his mother, and consequently no maternal ancestors showing at all.