I did similar to Tim (after all, he taught me a lot). So after some time working out the biggest TGs and any other low hanging fruit, I spent 3 or 4 all nighters and went down through my spreadsheet as quickly as I could. I checked each segment against it's neighbors (up and down) in my spreadsheet, to see which ones they matched. Sometimes it was not clear, so I just skipped over those. The point was to make a full pass - beginning of Chr 1 to end of Chr 23 (X) - and group wherever I could. I tried to do several chromosomes at at time. This was in 2013 as I recall, when I had about 1/16 of the Match/segments compared to now. After that first pass, I had TGs that covered about 2/3 of my DNA. I then made another full pass, and fit some more segments into existing TGs and formed a few new ones. And by then I was taking IBC segments (didn't match either side) out of my spreadsheet. Over time, as more and more Match/segments came into my Match lists, the TGs grew a little and started touching each other and filling out large areas. Over these 4 years, the percent "coverage" has grown asymptotically toward 100%. If I were starting today, and had a program that formed the TGs, I could compress 4 years of work into a very short time - and get almost exactly the same result (and the computer wouldn't make the little mistakes I sometimes do). Yes, I had several large TGs with close cousins in them with their large shared segments. If I ignore the large segment and look at the rest of the segments, it almost always naturally breaks down into smaller TGs. So I use the smaller TGs and include the close cousin's segment in each one. The close cousin creates a "knot-hole" and a "funnel" that shrinks the pool of possible Common Ancestors for the TG - see https://segmentology.org/2017/09/10/the-mrca-knothole/ The big TG has a closer MRCA, and the smaller, sub-TGs have more distant MRCAs which are ancestral to the close MRCA. Technically the TGs will continue to subdivide, but there is a limit to what we can "see", which imposed by the shared segment threshold we use. Jim Bartlett [email protected] -----Original Message----- From: Tim Janzen <[email protected]> To: genealogy-dna <[email protected]> Sent: Sun, Oct 1, 2017 2:15 pm Subject: Re: [DNA] Difficulty understand Ancestry's shared matches Dear Andreas, The way I handle this issue is to put everyone in a triangulated group who shares at least a portion of the DNA segment that is the longest segment in the triangulated group. The person in the triangulated group who has the longest segment must not have a known relationship with any of my family members. If this segment is say 30 cMs long there may be people in the triangulated group who don't share a DNA segment with each other, but they could share say 5-10 cMs with the person in the triangulated group who has the longest segment on different portions of that segment. If I ever figure out the genealogical relationship with the person in the triangulated group who has the longest segment, then I can always break the triangulated group down further into two or three triangulated groups. Sincerely, Tim Janzen -----Original Message----- From: GENEALOGY-DNA [mailto:[email protected]] On Behalf Of Andreas West Sent: Sunday, October 1, 2017 10:20 AM To: [email protected] Subject: Re: [DNA] Difficulty understand Ancestry's shared matches Btw, when you say 400 is that with splitting them into the smallest possible groups? Many times we do see TG's where we have smaller TG's at each end which are only connected by some DNA cousins who share the whole segment. That would probably go above 400. Andreas