AncestryDNA uses a composite score called PersonRank to decide if two pedigrees refer to the same person. This includes names, dates, and places, but it ALSO includes family relationships. I had a circle where the husband was listed, but no circle for his wife Lydia Winter, even though the vital statistics were identical in the two pedigrees. When I added the complete family group with Lydia's siblings, she had her own circle. I changed her name to ZZZ and the circle went away. I changed her name to Lidea Winters (a spelling used in some records), and the circle came back. Thus similar spellings were tolerated. It appears to be a cumulative effect: the more elements match, the more confident the algorithm is, but it doesn't depend on exactness of any single element. Ann Turner On Fri, Oct 30, 2015 at 6:55 AM, Jim Bartlett via < genealogy-dna@rootsweb.com> wrote: > Doris > > I've found that Circles depend primarily on the basic info for each > ancestor: name, dates, places. If you and another Match list the same info, > you'll get a Circle. If either of you inserts an asterisk (or an Ahnentafel > number, or a middle name, or rank, etc) in the name field, it probably > won't match = no Circle. I have 11 Matches in a Circle, but no Circle for > her one and only husband because some people insert a rank or false middle > name or guess differently on the place of birth (even tho they all use the > well known/copied bible DOB), and so no Circle for him. > > Jim - www.segmentology.org > > > On Oct 30, 2015, at 8:47 AM, Doris Wheeler via < > genealogy-dna@rootsweb.com> wrote: > > > > Peter, The rules for assigning circles are rather fuzzy. I think you'll > > find most people have few or none. I have just one, despite heavy > colonial > > ancestry, and it is a total mystery to me. View them as "suggestions," > > nothing more. >