PLUNK – An Ancestry DNA game for Christmas (and beyond?)

The “Ancestry DNA Matches” (No Boundaries/DNA/DNA Questions and Answers).thread shows that a fair number of us have an interest in the statistics of our matches, however there are issues over just quoting how many DNA cousins we have, as it is controlled by the number of our cousins who have taken DNA tests. “Distant cousins” is also corrupted by those of us who saved 2,000 or so 6cm and 7cm cousins in a group before Ancestry swung the axe, as we added our 3x gt grandparents surnames in our trees, or our relatives ‘favourite’ locations (or other reasons).

The number of “Common ancestors” is also ‘problematic’ and variable in that they are only picked up from trees where the member’s DNA result has been linked to their tree.

I therefore suggest a statistical ‘game’ that has a target, a score, and the chance to improve over time. I have called it PLUNK (see below for why) – It only uses statistics for close (not “Distant”) cousins.

P = number of DNA close cousins with Private Linked Trees

L = Number of DNA close cousins with Linked Trees

U = Number of DNA close cousins with Unlinked Trees (private and public)

N = Number of DNA close cousins with No trees (includes “Tree Unavailable)

K = Number of close cousins with “Common Ancestors”.

To enter you report each of the above in order, and then calculate and enter your score/percentage based on

=K/SUM(P+L)*100 – that is, the percentage of cousins with linked trees that Ancestry have identified as having common ancestors.

You can increase your score by building your family tree branches so that Ancestry picks up more cousins, but it will go down as you slowly gain more close cousins. It is something of a race. Target is 100% but if you have illegitimate ancestors this could be difficult to attain.

My own figures:

My DNA – P = 26, L = 221, U = 88, N = 114, K = 56 (449 close) and score = 22.6

My wife – P = 23, L = 164, U = 75, N = 87, K = 33 (349 close) and score = 17.6

Worryingly ? less than 50% of our matches have linked trees.

 

Mine come out as:

P=11, L=194, U=79, N=92, K= 48 (376 close matches) making my score 23.4

And just over 54% of my close matches have linked trees.

The figures are similar for my sister (352 close matches) - her score is 25.5, and just over 52% of her close matches have trees.

I'm not quite sure what this proves!

Presumably L is the number with Public Linked Trees?

 

Pauline,

On my DNA cousins private linked trees come up as "private linked trees" so P, public linked trees come up as "linked trees", so L.

Your results are quite successful then, compared to my own. To get to the other side of brick walls you need to identify which of your DNA matches are on the other side of them. By establishing, and increasing, the number that have Common Ancestors, whilst one or two might actually drive straight through a wall, their existence reduces the number of cousins to check on who might be on the other side.

Prior to our DNA tests Tricia's first brick wall was her grandfather, who we eventually discovered had changed his surname when he joined the army at 18. The DNA tests not only proved our theory but also, currently, the highest number of Common ancestors.

The figures show that Ancestry have made about a quarter of the currently possible automatic connections with your DNA cousins using tree information that you and the other cousins have entered. Pretty good. Have you identified where they are all connected yet?

 

My husband and I don't have as many close matches, but our numbers are:

My close matches (277): P = 14, L = 118, U = 60, N =85, K = 20, giving a score of 15.2
His close matches (239): P = 15, L = 119, U = 50, N = 55, K = 22, giving a score of 16.4

Only 47.7% of my close matches have linked trees, whilst 56.1% of my husband's have linked trees.

I notice that about 20% of the linked trees contain <10 people (and sometimes just 3, the DNA match and their parents) which doesn't help the matching process.

 

Thanks for your entries - Re "<10", it gets even more frustrating (not to mention time consuming) to open the tree and find they are all "Private".
I have managed to construct some of the cousins trees from just one pair of parents or grandparents, but all blank is useless.

 

Yes, I agree. Of my husband's 22 matches with "common ancestors", 3 of them have very small trees (<10 people) and in one case just 3 people, all private. In these cases Ancestry has clearly used other people's trees to make the connection, but whether this is accurate is often hard to confirm.

I have also been able to work out how cousins are related from very little information in their trees. It depends how common the names are of course.

 

[QUOTE="]calculate and enter your score/percentage based on

=K/SUM(P+L)*100 –[/QUOTE]

I do not understand how to figure out the math but these are my numbers
P=28
L=155
U=56
N=111
K=39
Total 389

At least two matches had only one person each in their trees.

I am assuming that a lot of those who are not common ancestors are linked to my brick wall, which will never be crossed as long as I do not have a name to go with him.

 

What a jumble of letters and numbers, I have trouble following any of it and am so happy I have not done DNA.
Is Peter's Masterclass of any use in these situations?

 

Do the numbers for each of the sections include those with Common Ancestors already or not?

 

Yes - if you add up John's figures, you will see this is how he did it. The percentage we are working out indicates what proportion of our close matches with trees have been recognised by Ancestry as Common Ancestor matches.

 

Your score from those figures is 21.3

 

Well that's the theory anyway - actually identifying those matches is another matter.

And it does depend on the right people having tested - always assuming that there were relatives sufficiently close to your brick wall who have living descendants.

 

Thanks Pauline - note that the first four figures (PLUN) should = the number of close cousins, the K is the number of these that are identified by Ancestry as having Common ancestors.
And to the 'cousin' in NZ - My wife's family history would have been more difficult without the DNA (see above about her grandfather) and she now has a number of new grateful cousins in the USA because the DNA linked their 'brick wall' great grandfathers with her cousins who had emigrated. To know that their ancestors worked for one of the most powerful Dukes in England has boosted their interest, one family tree going from 300 to 15,000 souls

On results so far - 53.9 % of member's 2431 close cousins had linked trees and Ancestry had identified Common Ancestors for 20.2% of them

John

 

My results:
P = 12, L = 124, U = 63, N =68, K = 22, giving a score of 16.2 (12.5 adjusted for errors)

However, one ThruLines match has gt-grandparents that are not her grandmother's parents (and I have supplied evidence, but left her to decide which is right - there is only one choice!), and five matches have "common ancestors" shown that are not my direct ancestors (confusion over Eliza Doel - one born 1802, the other 1806 - but my ancestor (1802) does not have a baptism record and died 1846 so shows as 1806 in 1841). The adjusted score removes the five errors.

Phil

 

Yes, quite agree, the Common Ancestors (CA) could well be wrong. They all need checking. I have one (correct) daughter whose mother isn't flagged as a CA as she has an incorrect grandparent. The daughter's tree doesn't go back that far.
Further info for the Duke story - the cousin had taken over his father's 'extensive' research which 'had been carried out using the Mormon's information in Salt Lake City' and the top of the line was to a man baptised in a non-existent parish in Cheshire (or anywhere come to that!) The number of his DNA cousins linked to my wife's convinced him of his father's error.

 

Here are two sets of results from my list of DNA - I'm not going to do my grandmother's results, as she has a grand total of 985 "close" cousins and that will take too much time to count it all.

So here are the results from my father's and mother's DNA tests (Their close relation lists are more manageable at 385 and 404 respectively)

Father:
P = 17; L = 159; U = 68; N = 81; K = 35
score: 25.7

Mother:
P = 17; L = 199; U = 88; N = 100; K = 11
Score: 5.09

*Note, my father's results include a large amount of immigration, my mother's does not. My father's (and paternal grandmother's) results have a larger amount of Common Ancestors than my mother's results. [My mother's was an immigrant herself. My father's results include immigration from 1839 onwards.]

 

I can understand why you have opted out of your grandmother's. Just by being from a previous generation she will have closer links to some of your tree branches (one step back allows one step forward on other branches) but that figure is definitely high. Has it helped get through any brick walls?

 

Jorghes - I make your father 19.9% ?? did I miss something.

 

Means I added it up incorrectly - not really all that surprising.

 

It cracked the Ashkenazi Jewish section in some ways, so yes.