Shared Match Cluster Hints
I’ve written in this Pro Tools series about the power of Shared Matches. They form manual Clusters of Matches. Like all Clusters, they *tend* to point to a Common Ancestor. Each individual Match has their own ancestry, and they may relate to us in several different ways (particularly with my Colonial Virginia ancestry). With auto-Clustering this is displayed by placing the Match in a Cluster with the strongest ties to other Shared Matches – and using gray-cells to indicate ties to other Clusters. This shows up in a Shared Match list with a mix of Shared Matches tied to one Common Ancestor, along with other Shared Matches who may be related in different ways, and even some Shared Matches who might not be interrelated at all.
So, to make a point: Shared Match Clusters (or concentrations in Shared Match Lists) should be considered as a Hint. The stronger the consensus, the stronger the Hint. The chore that still remains is tracing the genealogy from the Match to a Common Ancestor(s).
I find that consensus is a judgment call. But when I make that call, I usually find other Matches with a genealogy link as expected. But not always…
Segment Triangulation is fairly precise – each of our DNA segments came to us from one particular ancestral path. Shared Matches (aka In Common With, aka Relatives in Common, etc) are not equivalent to Triangulation. When Shared Matches form a Cluster, it’s a strong Hint. And a 20×20 Cluster is much stronger than a 3×3 Cluster. And a 20×20 Cluster where each Match matches almost all of the other Matches is very strong, compared to a 20×20 Cluster where each Match only matches, say, half of the others… I have found large, strong Clusters (beyond close cousins) usually turn out to include one TG (maybe two), but there is no hard rule.
Summary: Shared Matches can grouped into Clusters. Clusters are not the same as Triangulated Groups (TGs), but they can be good pointers and helpful Hints.
[22CW] Segment-ology: Pro Tools Part 15: Shared Match Cluster Hints by Jim Bartlett 20241125
segment-ology 
As always Jim, you make great points. This series has been excellent.
For examining our genetic evidence, I have long pushed the terms ‘clues’ for shared matches and ‘evidence’ for triangulated groups. I liken it to in the paper evidence world, the difference between looking at the ‘index’ vs examining the ‘original’ document such as births or marriages.
In the past we have been able to theorise about our clusters from AncestryDNA and they could give us additional supporting evidence depending as you say, on the size of the clusters and how strong the cross matching is.
However, enhanced shared matching is proving to be a real game changer. Previously we were in the position of being able to be confident about walking back our confirmed segments on our side (through TG’s) and finding clusters with at least one person from a TG was a great way to gather more supporting evidence. Confirming from the other side (our match) was often guesswork, without access to their match lists.
However with ESM even using the small segment matches of their close cousins and building their mini trees we are able to increase our confidence levels that even our matches that are out to quite distant cousins are likely to be coming from the same ancestor. I was very interested to read about your increased confidence levels back to 8C’s yesterday. Unfortunately many of us in Australia aren’t getting the number of matches of our US counterparts but even so there is lots additional clues emerging and tools like WATO coming more into play as we try to fit more of our matches matches (particularly for those who don’t have trees and are their closest matches). Your analogy of the jigsaw is so right, funnily enough my DNA screen saver is just that!
IMO we still need our segment data to really cement our conclusions and fit the pieces together in our chromosome map. However we have a 3 part process at AncestryDNA now:
* Examining shared matches for clues, or hints;
* Gathering additional supporting evidence to build our hypothesis through shared match clusters – as you say these are good pointers with tree triangulation strengthening our argument; and
* Continuing to push our matches to upload to a chromosome site to enable our hypothesis to be confirmed through segment triangulation.
Thanks again Jim, now back to more tree building…..
LikeLike