A Segment-ology TIDBIT

Thinking about Small Segments and Distant Matches…

Many have used the Speed and Balding IBD Statistics in Figure 2 of their Paper …  This chart has often been used to scare us away from small segments [by small I mean 7-to-15cM Shared DNA Segments – I do not encourage anyone to use smaller/”tiny” segments].

The vast majority of our Matches at AncestryDNA fall into this 7-to-15cM category, and I get many ThruLines Matches which have valid paper genealogies. They may not all link to the DNA, but I see no reason to discount them based on the small size of the Shared DNA alone. ThruLines is limited to Matches who are related as 6C or closer – not what I would call a “distant” Match. Only the small Shared Segments and the constant reference to the Speed and Balding chart, warning that small segments are usually distant, stand in the way.

This got me to thinking (watch out!)… The AncestryDNA Timber algorithm is well known to “down weight” the cM of many of our Shared DNA segments. Click on the “DNA” line in any Match Profile to see the “Unweighted share DNA” amount – often somewhat larger than the amount shown on the DNA Profile. This is Timber at work, downweighting the DNA that would be shown at, say, GEDmatch. One of the effects of this downweighting is that many of the AncestryDNA customers who would show up as a Match at GEDmatch are never shown as a Match to us at AncestryDNA!  It seems to me that AncestryDNA has already compensated for the statistics reported by Speed and Balding. It is thus unfair to compare our Match lists with the Speed and Balding statistics.

I’m not saying that some of our Matches are not distant – some of them are. What I am saying is to let the chips fall where they may. If we can find a Common Ancestor – at *any* Shared cM amount – why not accept it (if it also passes a genealogy review). The Shared cM Project clearly shows small Shared DNA Segments in the range for cousinships at 3C and more distant. Why should we be frighted away when our Match falls into the small segment category?

My blog post about a Common Ancestor Spreadsheet (here), now has over 8,000 rows of Matches with Common Ancestors with me. I sort them to get “nested” family groups, and draw comfort as I see the closer families and note they are Shared Matches with each other. New ThruLines have been pouring in recently (and the quality is dropping off a little). As expected in my Common Ancestor spreadsheet, a majority are in the small segment range. I am not worried about the cM size as long as the genealogy is valid!  

Bottom line: Let the chips (small Shared cMs) fall where they may; and focus on the genealogy.

Happy New Year!

[22CE] Segment-ology: Let the Chips Fall Where They May TIDBIT by Jim Bartlett 20240101