On Ancestry, the shared matches don't seem to be reciprocal. What am I missing?

[AngelaC]

I'm sorry if this has been addressed before - I couldn't find anything similar. I'm also sorry if this is very obvious - it just isn't to me!

My puzzle arises when I'm looking at shared matches. I have a DNA match "A" with whom I share 19cM. If I look at her shared matches, "B" appears in the (very short) list. I share 22 cM with him I think. However, if I look at "B"s shared matches, "A" doesn't appear. Is this something to do with the 20cM limit? If so, I clearly am not understanding it!

If some kind and patient soul could explain what I am seeing, I would really appreaciate it since, to my simple mind, if B is in "'a shared list then A should be in B's ....

 

[Helen7]

Yes, it's because of the 20cM limit. The shared match list will only show those with whom you share 20cM or more. So if you look at A's shared matches, B will appear because B shares over 20cM with you. However, when you look at B's shared matches, A does not appear because A is below the cutoff of 20cM.

Don't forget, you won't see how many cM A shares with B, only the cM you share with each.

 

[Pauline]

However, if I look at "B"s shared matches, "A" doesn't appear. Is this something to do with the 20cM limit?

Yes, it's because you share only 19 cM with A. Ancestry doesn't include people as shared matches if they are in the 'Distant Cousin' list of either of you, and this results in the kind of anomaly you describe.

 

[AngelaC]

Thank you very much, Pauline and Helen. I really appreciate you making sense of things for me. Not that I understand why the rule is there of course, but at least I'm not going completely mad!

 

[peter]

Not that I understand why the rule is there

It's there for the same reason that matches below 8cM aren't shown - because there is significant probability that they will be spurious and thus potentially misleading. The fact that B and C are related to each other doesn't necessarily mean that if A is related to both B and C, the three of them are connected in the same way.

However, whilst the 20cM threshold might be appropriate for the average Ancestry user, who - being American - may have many thousands of close matches (tens of thousands in some cases), it is arguably too high for users in most other countries, who are likely to have only a few hundred close matches (I have 215).

For someone like me, with only 215 close matches, there are just over 46,ooo potential pairs of matches who might be related to each other. For someone who has 5000 close matches there are nearly 25 million, so there will be many more spurious shared matches - so I can understand Ancestry's caution.

 

[PhilGee]

It's there for the same reason that matches below 8cM aren't shown - because there is significant probability that they will be spurious and thus potentially misleading. The fact that B and C are related to each other doesn't necessarily mean that if A is related to both B and C, the three of them are connected in the same way.

(Just a comment, not for further debate) I know you are sure that knowledge of the matching segment(s) is of little help or even misleading, Peter, but I feel this is a case where, for multi-segment matches, it could distinguish between "locations" - in my case, for example, between west Wiltshire and Leicestershire - where trees etc are non-existent/of no help.

 

[peter]

I know you are sure that knowledge of the matching segment(s) is of little help or even misleading, Peter, but I feel this is a case where, for multi-segment matches, it could distinguish between "locations" - in my case, for example, between west Wiltshire and Leicestershire - where trees etc are non-existent/of no help.

There are times when it is useful to know where on the genome matching segments are found, but because very few people truly understand the limitations of DNA there is a great danger that they will draw the wrong conclusions.

The easier it is to make mistakes, the more mistakes will be made - that's why there are so many errors in Ancestry trees. If people start drawing the wrong conclusions from DNA it will undermine trust in the process and ultimately discourage people from testing.

There is nothing to stop Ancestry users transferring their raw DNA to other sites which do identify the matching segments.

Note: it's potentially confusing to refer to 'locations' when you're talking about geographical locations, because when DNA is being discussed the term is usually used in the sense of a location on the genome.

 

[AngelaC]

Thank you for the additional information and comments Peter and Phil. I shall get to grips with this DNA stuff if I live long enough......

Like you, Peter, I don't have that many close contacts at Ancestry - 400 - the vast majority of which are the result of one 3xgt uncle who went to the US in about 1830 with his family and had prolific descendents. If I discount those, I have relatively few.

 

[canadianbeth]

If by "close matches" you mean the number that shows when logged onto the website (4th cousins or closer) then I have 346. Altogether I have 12,703 as of today. I have been gradually going down the list and putting everyone into a group; am down to 17cM at present. The biggest groups of course are no shared matches and unknown. I have found quite a lot of names that when I click on shared matches I get someone that is already in the no shared matches group; I put them into unknown. Only 46 of the 346 have been added to an actual family group; all the rest are either no shared matches or unknown. And I have not marked all of those 46 as confirmed. There are some below the 20cM cutoff that are identified as family as well. I estimate most are from the paternal Joyce side.

 

[peter]

I've identified for certain how I'm related to 41 of my 207 close matches (215 less 8 known cousins), ie 20%, and have a good idea how I'm related to about a dozen more. In other cases there are hints in terms of locations in common, but absolutely no sign of how we our trees connect. Although there is a chance that there's a NPE somewhere it's most likely that these are distant matches that appear closer than they really are.

The vast majority of the unidentified matches (and some of those who have been identified) don't have trees - which is par for the course.

 

[canadianbeth]

Two of the confirmed matches are my niece and my first cousin once removed. Even though my niece has no tree, I know her and as my sister's daughter is pretty much connected to everyone on both sides of the family. And my cousin's daughter is connected on the maternal side; I am assuming her shared matches would not be paternal and I have put them into a probably not paternal group, so therefore no connection to my unknown paternal grandfather.