I'm not going to read a 14-page thread right now, so I will not be commenting on anyone else's opinion.
I am an enthusiast for the use of DNA for genealogy purposes, and I am a reasonably knowledgeable amateur at present. This is entirely a result of the enormously helpful project administrators and members at Family Tree DNA.
I have had male family members from both my parents' families do Y-DNA tests - at present, to 67 markers with a couple of SNPs, in both cases. Coincidentally, the main blockage in my family history research is in one parent's surname line, so Y-DNA was, fortunately, the way to go. The other was just out of interest at the same time.
In the latter case, I have had no meaningful matches at all. This confirms (well, supports) my suspicion that my grx2 grandfather born c1820, who appears to have been the only son of an only son of an only son (that last one born c1735), was in fact the last male in his line in the village in question. He went forth and multiplied mightily, 10 kids with two wives. Through my notes in a census record at Ancestry, I met another male-line descendant of that grx2 grandfather. When I uploaded my relation's Y-DNA markers from FTDNA to Ancestry (when Ancestry still offered that test), I got a 100% match. Lo and behold, it was that cousin, who had tested there. So we have proof of "legitimate" descent from 1820 on, anyway -- always good for a family history researcher to know!
In the other case, a surprising and fun outcome.
My gr-grandfather had a sister, an actress. We had never heard of her. We had never heard of any of his family. After he came to Canada with wife and kids c1910, he never spoke of them. About a decade ago, I embarked on this family history lark, and my first discovery was that he did not exist, before his second marriage at age 30. His one tale had always been that his uncommon surname came from the black sheep younger brother of a peer, on the wrong side of the blanket. It became apparent that the name had been assumed in adulthood. Through the wonders of search engines and my dogged dexterity with them, a set of unimpeachable coincidences in time and space emerged (a person who matched his details precisely who disappeared from the records c1875 when his wife died, with the gr-grfather finally found springing from nowhere in the 1881 census ... living with the mother and daughter of the man who had disapppeared a few years before). Thus I discovered who he "really" was, i.e. his official surname. His sister, the actress, had been given the fake surname as a middle name at birth. He and she, but not their older siblings, adopted it as their surname.
So the Y-DNA testing was intended to find whether there was any truth to his tale about his name, or he really was a "Smith" (his official surname was nearly as common as Smith). There are too few people with the uncommon fake surname, and I have not got up the gumption yet to approach the current peer for his spit, although I've been introduced to him via email. I don't think there's a connection ... although the unmarried younger brother of the peer in question did die at Alma a few weeks before the birth of the actress sister, and the official parents appear to have been estranged around that time ...
I expected I might find a match in the "Smith" surname project, and that the match would likely be descended from a person who emigrated from Cornwall to the US in the mid-1800s in connection with mining -- since most testers at FTDNA are in the US. This would confirm the surname I should be researching, and provide the US member with an answer to the question of their origin.
Not a sausage. Not a single remote match among the hundreds in that project. This seemed highly unlikely, if the real surname in my family was "Smith". But FTDNA did give me one excellent match, one that is very probably within the reach of parish records. With an entirely different surname. But -- it belonged to an elderly man whose grandfather had emigrated from Cornwall to the US in the mid-1800s in connection with mining. From a family long resident about 10 miles from where my gr-grfather was born.
So the possibilities are: that family's surname got switched somewhere, my family's surname got switched somewhere, or our connection predates the adoption of standardized surnames. The switch could have been intentional, as in my gr-grfather's case later on, or could result from a "non-paternal event" where a child took its mother's surname or was given the surname of a husband who was not its father.
and I have to break here for length, with apologies to anyone whose eyes have glazed over ...