Wednesday, May 27, 2015

Follow-up to "You've been assigned a random ancestor for Memorial Day..."

I finally checked my husband's Memorial Day email from FamilySearch. Here are the results:

Ancestor Column
(Seriously!) The rest may be relatives on a family line, but definitely are not ancestors, including the person named in the email.

Ancestor's Headstones
One of them was fun to see since I am not too familiar with my husband's colonial ancestry.

View My Relationship
The listed connection to Nathaniel Tilden (1583-1641) included an incorrect generation, but despite the error in "View My Relationship" it does look like he is an ancestor.

8/30 = 27% accuracy

In Conclusion...
It looks like FamilySearch needs to keep working on its computer matching, with the caveat that since the matching is using user-generated trees, it is unlikely to reach 100% accuracy. Hopefully most people saw accuracy closer to what I saw. (85%)

I haven't kept track, but fortunately FamilySearch hints seem to have a much higher accuracy rate.

Sunday, May 24, 2015

You've been assigned a random ancestor for Memorial Day...

Today FamilySearch sent the following email...

It was bemusing since I have an ancestor with that name, but not with those dates or burial location, but I still followed the link, signed in to FamilySearch, and saw the following list...

It is pleasant to read through and think about the named ancestors and look at their graves, but the last woman on the list is not an ancestor of mine.

This is all computer matching, so how accurate was it? I'll ignore the Map Location data and just look at the three other columns. In all, there are 30 links to my tree and information within the tree.

Ancestor Column 
The entry for John Tanner still has the wrong dates but links to the correct ancestor, correcting the data in the email.
Not sure who Martha Morgan is, and it links to a woman named Matilda Nelson.

Ancestor's Headstone
John Tanner: wrong person
Martha Morgan: wrong person

View My Relationship
Martha Morgan: a relationship is listed that is not supported by the actual Family Tree. Martha Morgan disappears and Matilda Nelson is shown as the fifth great-grandmother of John Tanner, but that relationship is not in the tree.

25.5/30  =   85% accuracy

In Conclusion...
Of course I have no idea how these statistics would represent what all the FamilySearch users saw, but my husband also got an email listing someone not an ancestor. Hence the title of the blog post.

In any case, Happy Memorial Day. We're happy to remember all these people, related or not, but most especially we remember our war dead.