In DNA which was a long time ago, I said I was going to do some DNA tests so I could comment on DNA testing for genealogy with some facts.
The first problem is the inordinate amount of time it takes to get things done, mostly caused by the DNA testing company being unorganised and not having enough capacity to cope with their sales and also by postal services who seem to think it is normal for small packets to take a month to deliver.
So after 3 months I got my first results from an autosomal test, and initially it looked like good news with 1800 matches. It did not take long for disappointment to arrive.
The matches list was sorted by size of matches but could be reordered by closeness of relationships. The best results were 20 people who matched as 2nd to 4th cousins, so that seemed the best place to start, they should be easier to find than 6th cousins, and with the size of my tree the links might already be partly in place. The problem noted first was that 60% of the matches had not uploaded any tree for me to compare mine to, some just had two or three people marked private, some had a few 20th century ancestors but often with no places mentioned. I assume these people have done a DNA test so that someone will do the research for them, being to stupid or lazy to do the research themselves. I emailed most of these prospective matches and half have not replied, so not a lot of help there.
The next thing you notice is the relationship calculator is very ambitious, I find it very improbable that a family line that has been exclusively US based for 300 years is going to have a second cousin relationship to my line which has been British for the same period and more.
Uploading my data to gedmatch seemed to produce better results, relationships seemed better estimated comparing gedcom files was a useful start to narrowing results, but once again many people did not have a gedcom to check with and a high number do not respond to questions.
So far I have narrowed my search to two potential matches, in both cases we have numerous family names in common, in the right region at the same time but no obvious common ancestor yet. For the most likely match I have added about 300 people to descendant lines of my direct ancestors, found two marriages with his paternal line that were not connected to him yet.
So are there any conclusions to be drawn yet?
The first conclusion is that you cannot expect to take any DNA test and have your ancestry mapped out for you, you are more likely to match with some other idiot doing the same thing and both of you will get nowhere.
The second point is that to get anything from a DNA test you need to have done a lot of research first, and preferably properly sourced research where you give citations. In my mind if you have done this then DNA will not add anything, except maybe some certainty over some dubious history.
The third conclusion is that testing methods of totally inaccurate and dubious. I studied DNA at university and know that the cheap tests on offer are not good enough for accurate results, evidence for this is I have no matches with known cousins who tested with the same company and the hundreds of matches that cannot be correct. Note that by cheap I mean comparatively cheap, a proper, accurate test would cost about 10 times what the rip off merchants charge, and for what the offer they are really overpriced.
At the moment I cannot recommend or even suggest that anyone buys a DNA test of any sort. If someone really thinks they should test then I would advise them to not test with familytreeDNA, they are incompetent and rude as well as inefficient.
I am still waiting for more test results so I may revisit this and update my opinions, and it could just be that my results are so poor as I am only half human. Surely there is someone who has good results from DNA testing.