Ann I tried to correct my post: I meant GEDmatch has no filters - I was thinking ahead and typed the wrong company. And I agree that GEDmatch has a high number of IBS in the 7-10cM range. A straight algorithm will pick out ways to match those segment when one or both are IBC/IBS. They make up for it by allowing direct comparisons between any two kits, and Triangulation will cull out most of the IBS. With GEDmatch I have several (bad) pile-on areas about 9cM wide with several hundreds of 7-8cM Match-segments in them - virtually none triangulate. Somehow both FTDNA and 23andMe don't report (filter out) most of these false Matches. Jim - www.segmentology.org > On Nov 9, 2015, at 12:46 PM, Ann Turner via <genealogy-dna@rootsweb.com> wrote: > > That's my interpretation of the word, but I don't understand why Jim > Bartlett said 23andMe didn't have any filters and the 3 (?) "DTC" companies > do have filters. We know about AncestryDNA's Timber, FTDNA's 20 cM filter > (which seems ill-conceived) and the ability to set your own thresholds at > GEDmatch. GEDmatch suffers from the need to accommodate multiple testing > platforms and generates a high percentage of false positives. > > Ann Turner > >> On Mon, Nov 9, 2015 at 9:29 AM, Ann Turner <dnacousins@gmail.com> wrote: >> >> I'm not certain what you mean by filters, but 23andMe does have some >> different thresholds in place for regions known to have "excess" IBD. They >> increase the SNP and/or cM value, so if the "pile-up" is embedded in a >> longer segment, it will pass muster. >> >> Ann Turner >> >> On Mon, Nov 9, 2015 at 7:42 AM, Jim Bartlett via < >> genealogy-dna@rootsweb.com> wrote: >> >>> Andreas, >>> 23andMe doesn't use filters, so we see many Matches there that the 3 >>> DTCs >>> didn't post. In my experience some of them are IBD and Triangulate, >>> but many >>> don't (and go into my IBS bin) >>> >>> >>> >>> Jim Bartlett >