genetic testing math probability of relationship

ETA: rather than asking for a refund right away, maybe first ask them for an explanation of the analysis? Their conclusions may be wrong, but there may still be something rather interesting here if we can drill down to the raw data of what they have found.

I suspect that their answer will be something like, “We use Math! to calculate the probabilities, and Math! is never wrong, so you shouldn’t question our results. Who are you going to believe, Math! or your common sense?”.

My maiden name is my grandfather’s stepfather’s name. I understand that this was a fairly common practice. When a woman was widowed young, and remarried, he new husband became the de facto father of any children she had with her first husband, and if they were very young, they assumed his name without any legal proceedings. Before birth registration immediately at birth was mandatory, many people had their baptismal certificate (not a concern for my family), and their school entry. My grandfather was only five when his mother remarried, so I’m not even sure how old he was when his father died. He would have entered school for the first time after his mother remarried, and that is probably the first official record of his name, because I don’t think the certificate you got for your bris had any English on it at the time.

I don’t even know what his biological father’s last name was. He has an older sister with the same biological father, and four siblings with his “adoptive” father, who my father says treated him like his own son, and my father didn’t even find out this wasn’t his biological grandfather until he was 10 or so. My father was his first-born grandchild, and was always called his “kaddishele.” I never met him

My point is, becoming a widow with small children was much more common than it is now, and two world wars made it really common in the 20th century. After about 1930, it probably started to become more common for stepparents to formally adopt stepchildren who did not have another parent, but there were probably lots of informal assumptions of a surname before that. So you can’t assume that a common name means a biological relationship.

To make this claim ever more difficult to support, the database of people who have submitted a Y test might charitably be estimated at one tenth of one percent the total population of males in the US and Europe. If one extrapolates the 2800 to the total population, the claim becomes 4generation male relationship with 2800 x 1000 men

(x.34)

I don’t doubt that the 2800 men detailed in their ‘matches’ report are, in fact, co-male-descendants of some common ancestor. My beef is with the assertion that such a connection might well be within the period of the last 300 or so years. For the wide diversity of locations and surnames they present, the common male ancestor must be far back in the ‘mists of time’, previous to surname commonplace time.

One other possibility that happened to occur to me is that a fairly recent ancestor was a sperm donor to a fertility clinic and ‘fathered’ hundreds of thousands of male children. Nah, that theory doesn’t really work. Hundreds? barely possible. Thousands? Laughable.

No response yet from FTDNA to my claim their claims are really bullshit.

RivkahChayah, yes, that’s another way that surnames might become disconnected from Y chromosomes. But the thing is, even though there are plenty of ways names and chromosomes can become disconnected, the evidence is that, overall, it still doesn’t happen all that much, and Y chromosomes really are correlated. This is more true the older and more distinctive the surname is, of course: “Cohen” and variants has a much better correlation than “Smith”.

I did the DNA 12-marker test at familytreedna a few years ago; all they told me about haplogroup is that it is R-M269, the Neolithic clade ancestral to most of Western Europe.

Someone with the same 12 markers as I have, as well as the same surname, has done the 111-marker test and also shows a specific downstream SNP. Reminded of all this, I just now ordered a $39 test for that SNP. But the MRCA for even that subclade was before the Iron Age — and not just 4 generations ago!

I just think they need over a hundred STP’s or whatever to give you a decent shot at comparing two men’s agnate line. Basing it on test results than millions of men might share seems to be a ‘come-on’ for buying the more expensive, more detailed tests.

To finish this up, this is the actual reply:
Thanks for contacting us. It sounds like you’re looking at the TIP calculator for your 12 marker matches from your Y-DNA test.

The TIP calculator is a rough guide used for estimating how far back your common ancestor statistically could be based on how often mutations occur in the tested markers. [COLOR=“DarkOrange”]Generally it isn’t designed for matches at 12 markers.[/COLOR] It’s designed for matches at higher levels of testing (37, 67, 111 markers) and is a very useful tool at those levels of testing to get an idea of how far back in your family tree you should check for your common ancestor. We recommend you look for a probability of 90% or higher to estimate how far back your common ancestor actually is.

We aren’t trying to suggest that 34% of the matches we’ve found share a common ancestor with you within the last 4 generations. We actually recommend that you consider 12-marker matches to be very distantly related to you. We can only confidently say that they share a common ancestor with you within the last 25 generations.
Well, they claim they recommend that you consider 12-marker matches to be very distantly related. I don’t find that statement, or its equivalent, in any of their promotional materials or the results page. Que Sera.

Even this is very wrong. 25 generations is about 720 years, but I was told on genetics boards that 12-markers are almost useless for even getting to the right thousands-year-old subclade.

The phrase “share a common ancestor within 25 generations” may be correct if all 2[SUP]25[/SUP] = 33 million of your 25-generations-ago ancestors are considered, but it’s an absurd response in the context of a Y-chromosome marker test. Only one of those 33 million ancestors is your Y-chromosome donor.

Perhaps familytreedna.com is a victim of its own success and needs such a large PR department that they’ve had to populate it with idiots.

Perhaps. another correspondent, off board, tells me that my numbers are so close to the average ‘modal’ Western European Haplotype that I was bound to see thousands of ‘matches’. He, himself, only had a few dozen with the same test, and thinks that is more the norm than my plethora of matches.

Thanks for the erudite response. You are a gentleman and a scholar, and there are damned few of us left.