How accurate do you find the 'Common Ancestor' feature on Ancestry?

[Another2Cats]

I'm typically wary about relying on other people's trees without checking them, but I've recently gone back and had a look at how accurate the 'Common Ancestor' (which is based on other people's trees) marker is.

I was surprised to find that around 90% were correct - or, at least, accurate enough that I could find the actual link.

Both my parents have their DNA on Ancestry. My dad has 127 DNA matches with the 'Common Ancestor' marker and, of those, 114 (89%) were correct - or, if not entirely correct, at least correct enough that I could find the actual connection.

My mum has 172 matches of which 154 (89%) are correct or close enough.

When I looked at the matches that weren't correct, some were just so totally and obviously wrong that I don't know why they even showed up on the list.

But with the others, it was not so clear cut. In some cases there were a number of people with the same name and roughly the same age who could have been the parent, one of whom was my relative. But I could find no evidence that it was my relative (rather than any of the other people of the same name) who was the parent.

So the link may not be wrong, just that there's no evidence to back it up.

In other cases, it appears that there is a family connection but it may be further back than Ancestry says. For example, Ancestry may suggest a link to somebody in my tree but the actual link turns out to be with an aunt/uncle or a grand aunt/uncle etc - so, a couple of generations earlier.

I think part of the reason for this might be that the 'Common Ancestor' feature only goes back to 6th cousin. So Ancestry will try and find a link no further back than 6th cousin. That would be your 5x great grandparents.

But I've found on a number of occasions that the actual link is one or two generations back from there. For example, Ancestry may suggest that William Smith is the father of somebody on another person's tree but the dates don't really add up.

Doing a bit of digging I then find that William Smith is in fact either the grandfather or great grandfather instead and Ancestry has just cut out a couple of generations.

Or sometimes it is two people with the same name (something like an aunt and a niece or cousins or 2nd cousins etc) and Ancestry linked it wrongly to the younger person.

There are lots of variations on this theme that I've come across, some are straightforward to sort out but some just spiral into so many possibilities that I don't really have the time to sort it out.

.

I just wondered if my own experience of this feature was similar to how others have found it?

Do you find it accurate or not?

Yes pretty much as you've described, generally very accurate with a couple of absolute headscratchers, completely wrong.

It managed to predict a set of common ancestors in the 18th century which I was impressed by.

Multiple matches from Ancestry Thrulines also made me aware that a genealogist who did my mother's family tree back in the early 1990s took a wrong turn in his research and built a large section of the tree around an 'ancestor' who was no such thing.

I have only 101 DNA matches with the Common Ancestor marker. I’ve not looked into this in detail, but your analysis is interesting.

I will look into this further, although I notice I have a 9cM match with someone. We both have trees on Ancestry that clearly have no connection, but Ancestry thinks that my 2G grandfather is his 5G grandfather. You can actually check the putative connection on ThruLines, and the mistake seems to be because there are two John Clarkes with daughters with similar unusual names, and this mistake seems to have been copied into a large number of trees. But I couldn’t see a way to point this out to AI!