RootsChat.Com

I have around 40 matches in total mostly on Ancestry who are descendants of the Cornelius and Harriet from the 1861.
Whilst they match me at anything up to 200cM they don't match any of my confirmed maternal relatives. I can now include the recent match on MH at 290cM who descends from Charles/Sidney into that group.  I've traced their kids forwards as best I can but there's still three  I haven't been able to follow completely;
Harriet  b1857
Mary Ann b1858
Henry aka Harry b1869. I think he's the army deserter reported in the Horncastle News in Nov 1888 but then I lose track of him altogether.  The family were involved with travelling shows so can be pretty transient at times, Cornelius snr is a regular in the papers usually being charged with being drunk or failing to quit licenced premises.  Many migrated to Lincoln in the 1880's  and my maternal family are largely in Lincoln from about 1830 and at Wragby earlier than that.  Both families had addresses on Hungate or Hungate Court but slightly different times.



A couple of days after this 280 cM match came up I had another match on MH taken by 'P' who on paper is my full maternal 1c but we only share 343cM, not enough for a full cousin match.  His nephew 'D' and grand nephew 'A' have also tested (matches to me at 319 and 280 respectively), as they are all MH tests I can see how much they share with each other;
P & D share 1537cM
P & A share 855cM
D & A share  3534cM
I would suggest that those amounts are all safely in the 'full' relative bracket so there's nothing dubious in their immediate family.  None of them have any matches within the Wesseldine group
P, D A and myself  only seem to have shared matches through one ancestor; my maternal grandmother and my feeling is my mother is therefore an NPE. My mother (the youngest of 9), was born in 1924 in Lincoln.

I hadn't found the Cockerill/Dobbs marriage so that's a bonus.

Limited value at best in my experience as it's based on where people are living (or claim to be), based on their profile information.  I have matches covering three generations of a family as far apart as Aberdeen, Kent and Cornwall, it's only by building their trees I know they are related and were all born in Sheffield from around 1950-2005. 
That's not to say it is entirely useless but it would take a lot of luck to find a family line with multiple testers who have adequate profile information, accurate trees and haven't moved far for many generations.

I've got the problem of having no living relatives above me now.  My mum's line had died out, she was a single child to a single child but I have had DNA matches with older generations in that family.  My dad and his siblings are all passed now and I've got matches to his mothers families and his dad's mothers side but the direct male line is known on paper but not confirmed by DNA.  They were notorious for affairs so I'd like to sort that one out some how.

Very much like my scenario, nobody from a generation above and my generation range from 57-79 years of age. I'm giving up on a lot of close relatives as they just copy from any old tree and ignore their dna results and the information in my tree and messages to them. I'm sure if I made a 2nd account and a fictitious tree showing a link to Cleopatra they would copy it and message me in a day bragging about the link. 

The downside is that many fall for the marketing thinking they will get an instant tree so the pool of genuine researchers isn't as large as the test numbers indicate.
On top of that the  issue of the 'copy regardless' is ever present. I couldn't make it any clearer on my tree or through 3 acknowledged messages that my great grandmother though married fell pregnant to another man (hence a pile of matches on the Isle of Wight where the father finished up), but my closest matches just copy other trees based on a household in the 1891 and 1901 census so have the wrong ancestors. They may as well skip the dna test for all the notice they take of the results and just go with the majority of tree  just as they do anyway.

I think a further issue is that low response rates to messages just drives the research to working out links to matches further underground to the point where lone research is the default starting point and can remain the prefered method.

I find the MH relationship predictions to be all over the place. I have a maternal 1c, 1c1r and 1c2r (all descendants of one of my uncles), the predicted relationships  are 2c, 1c and 1c respectively.  Yep, the closest of all with the highest dna share is predicted as the most distant of the three.