You don't need a subscription for ThruLines.
-
Only registered members can see all the forums - if you've received an invitation to join (it'll be on your My Summary page) please register NOW!
-
It's easier than ever before to check your entries from the 1881 Census - more details here
Ancestry Matches 7cM and less
- Thread starter Andrew Lloyd
- Start date
That's right - I have several instances in my tree of two sisters marrying two brothers, though only one in my direct lines. The effect is that descendants of one couple are double cousins to the descendants of the other pair - so they will appear half a generation more recent than they actually are.
Katie Bee
LostCousins Member
Luckily one of my annoying ThruLines potential ancestors has gone - hopefully never to return.
Blaine Bettinger posted the following a couple of days ago
DEDICATED POST for Pre-Timber Shared Amounts
Great news! During a call with Ancestry today, they shared that they will soon provide the total PRE-Timber shared DNA amount (in addition to the pre-Timber longest segment and the post-Timber shared DNA amount). This is a great development, as knowing how much Timber affects a match gives insight into the possible age/distance of a match. The more Timber affects a match (meaning the greater the difference between the pre- and post-Timber total shared amounts), the greater the red flag and the more likely the match doesn't represent recent shared ancestry. There is NO firm date for delivery, but likely within the next month.
REMEMBER! TOIL 90 - Timber Only If Less than 90 cM (thus, Timber is not applied if the match shares 90 cM or more with you)
For more about the Timber algorithm: https://blogs.ancestry.com/ancestry/2015/06/08/filtering-dna-matches-at-ancestrydna-with-timber/
DEDICATED POST for Pre-Timber Shared Amounts
Great news! During a call with Ancestry today, they shared that they will soon provide the total PRE-Timber shared DNA amount (in addition to the pre-Timber longest segment and the post-Timber shared DNA amount). This is a great development, as knowing how much Timber affects a match gives insight into the possible age/distance of a match. The more Timber affects a match (meaning the greater the difference between the pre- and post-Timber total shared amounts), the greater the red flag and the more likely the match doesn't represent recent shared ancestry. There is NO firm date for delivery, but likely within the next month.
REMEMBER! TOIL 90 - Timber Only If Less than 90 cM (thus, Timber is not applied if the match shares 90 cM or more with you)
For more about the Timber algorithm: https://blogs.ancestry.com/ancestry/2015/06/08/filtering-dna-matches-at-ancestrydna-with-timber/
I was in the same Zoom conference call but wasn't particularly excited by this news. I suspect that many users will ignore the post-Timber figure so that they're working with a number that is comparable with other websites; based on my past experience of working with matches in pile-up areas I'm more likely to do the opposite and downgrade matches further than Timber has.
So in a sense it's good news all round - but for me, the problem will be persuading other users that segments in pile-up areas are downgraded for a reason. Ancestry is getting more complicated, and that's bad news for the vast majority.
The really good news from the call is that Ancestry are going to make it easier for those who know little or nothing about their family tree to make use of their DNA matches. If nothing else this should improve the message response rate.
So in a sense it's good news all round - but for me, the problem will be persuading other users that segments in pile-up areas are downgraded for a reason. Ancestry is getting more complicated, and that's bad news for the vast majority.
The really good news from the call is that Ancestry are going to make it easier for those who know little or nothing about their family tree to make use of their DNA matches. If nothing else this should improve the message response rate.
Susan48
LostCousins Superstar
Would someone please explain what pile-up areas are, and how to recognise them in DNA match results? Or a link to an explanation would be helpful. Thanks.
Pile-up areas are parts of the genome where lots of people in a population share the same DNA. For example, colonies are generally established with small numbers of people and so it's very likely that some DNA segments will be shared by more of their descendants than you would otherwise expect - it's rather like certain mutations being more common in particular ethnic groups.
From a practical point of view it can make genetic cousins appear a LOT closer than they really are, and that's why Ancestry's Timber algorithm downgrades the length of the segments - it's an attempt to compensate. Typically you'll have a group of cousins who all match each other, but where you can't see how you're connected to any of them.
Winksetter
LostCousins Member
I wish I could get on with research and not have to worry about the tools that we have available but.....I have been intensively doing research based on DNA results of late and really getting fed up. Today I noticed again that the total cM and longest segment cM are mostly, grossly out of kilter. E.g one of mine has a total of 28 and longest segment 48cM. What use is this sort of rubbish. This is not untypical and is totally useless information. Why not just gives us the definitive facts untouched by algorithms and sticky Ancestry fingers.
Next I went back to a group of 8 relatives of mine All in the same family line, all being actual and individually DNA matches to me. I tried shared matches to the first and received a no matches message. i went the next person and received the same message. All 8 gave the same result. Again I know this result is wrong, each one should have at least 7 shared matches. I think these shared matches are falling foul of the latest purge of lower DNA matches although they are all above 8 cM matches to me and must be higher with each other.
None of this gives me any confidence at all.
As you will know from this discussion we are going to be given the unadjusted shared DNA information in the near future. At the moment we're in a transition period, so - as I explained earlier - we need to be patient.
The information you describe as 'totally useless' is potentially useful when you know the difference in the way that the two figures are calculated - it tells you that 48cM segment has been drastically downgraded by Timber.
There are no definitive facts - DNA is all about probability (for example, see the White Paper). Ancestry's aim is to produce the most helpful information possible - the publication of the unmodified match figures could well have a negative impact, depending on how knowledgeable users are, so I can understand why this information has been held back up to now. I suspect it will hinder more people than it helps.
The information you describe as 'totally useless' is potentially useful when you know the difference in the way that the two figures are calculated - it tells you that 48cM segment has been drastically downgraded by Timber.
There are no definitive facts - DNA is all about probability (for example, see the White Paper). Ancestry's aim is to produce the most helpful information possible - the publication of the unmodified match figures could well have a negative impact, depending on how knowledgeable users are, so I can understand why this information has been held back up to now. I suspect it will hinder more people than it helps.
Shared Matches only shows people who share 20cM or more with both parties. - this has always been the case at Ancestry.
To have confidence you either have to understand how something works, or trust the judgement of others who do. None of us can be an expert on every facet of family history, but this forum is a great place to learn from others.
Winksetter
LostCousins Member
As to shared matches, my point is that all 8 family members are far in excess of 20cM matches with each other although less with me. They all should show up, none do.
As to total cM etc, to give confidence the figures have to be sensible and clearly justified. At the moment they are not useful and not sensible. They are not explained by published information. Why on earth not show really useful facts. Maybe length per which chromosome for example. That really is useful.
As to total cM etc, to give confidence the figures have to be sensible and clearly justified. At the moment they are not useful and not sensible. They are not explained by published information. Why on earth not show really useful facts. Maybe length per which chromosome for example. That really is useful.
Peter
I am getting slightly confused - according to the blog page referred in Blaine's note Timber was here in 2015.
I am assuming that when most people refer to pre-Timber or Post-Timber now - they mean the latest algorithm that has been recently introduced.
However I have a person in my matches that shares 48cM with me (this is unchanged for say 4 years and has not changed as a result of the new algorithm) but when the longest segment was added and it was 60cM.
So is the 48cM the post-Timber shared DNA
and 60cM is the pre-Timber longest segment
Bryman
LostCousins Megastar
I think that you will find that pre- and post- refer to the application of the adjustment, irrespective of when the algorithm became available. Both pre- and post- figures can refer to calculations made today and just refer to whether the Timber adjustment has been applied, or not.
Yes - sorry if that wasn't clear from previous postings.
Just to clarify for the benefit of others reading this thread, when people talk about pre-Timber and post-Timber they are not talking chronologically, but about the figures before and after the application of the Timber algorithm.
Pauline
LostCousins Megastar
Only those who also share over 20cM with you would show up as shared matches. This feature is working fine for me and is unaffected by recent changes.
Your experience must be very different to mine. I have managed multiple kits at FamilyTreeDNA and GEDmatch since 2012 and I can't think of a single instance in which knowing the precise details of a match led to a positive outcome. At best it allowed me to identify which matches involved (wholly or mostly) segments in pile-up areas so that I could discount - and usually discard - those matches. At Ancestry the hard work is done by Timber.
PhilGee
LostCousins Member
It is probably more a case of supporting evidence in the same way the baptism, banns and burial records provide supporting evidence for births, marriages and deaths.
I have a match with 634cM/17segments and 28 shared matches that I have split 22-6 between my gt-GF and gt-GM based on a 5-5 split for identified common ancestors and examining the "shared matches" shared matches. From some spreadsheet experiments with my MyHeritage matches, I believe knowing the segment data would help confirm the split and should identify/confirm a path along which the common ancestors are positioned.
I have a match with 634cM/17segments and 28 shared matches that I have split 22-6 between my gt-GF and gt-GM based on a 5-5 split for identified common ancestors and examining the "shared matches" shared matches. From some spreadsheet experiments with my MyHeritage matches, I believe knowing the segment data would help confirm the split and should identify/confirm a path along which the common ancestors are positioned.
The chance of finding useful evidence has to be weighed against the likelihood of making incorrect deductions. Many people don't fully understand the limitations of DNA, and assume that by drilling down they'll be able to add to their knowledge - but usually it works the other way round.
Think about this: if the information is so useful why don't Ancestry provide it, and why am I not advocating that they provide it?
I can't speak on behalf of Ancestry, but in my case it's because many years of experience have taught me that in most cases the best outcome is that you waste a lot of time. The worst outcome is that you come to the wrong conclusions.
PhilGee
LostCousins Member
The same can be said about tree hints - and I'm not convinced, yet, that hints are better. Especially when Ancestry, it is suggested, actively "suppress" the effects of pile-ups etc where they believe the results are/could be misleading through "Timber".
I'm not sure this is the whole story, it might be to protect their information on pile-up regions etc. I'm also not sure that tree hints/Thrulines come in the category of "so useful" unless used with care, on the basis of a somewhat extreme example:
I have 10 tree "hints" for my 2gt-GF, John Jones. Of these, one tree is a close match for data, perhaps too close (Parents, death index and birth county - but a difference in the place of birth: Llangollen from the 1851 (a miss-transcription as on Ancestry - it should be transcribed as Llansellio/Llansellia as on the 1861 census) whereas mine is Rhydwilym - from the 1891 census. Rhydwilym is in both Pembrokeshire and the Carmarthenshire Parish of Llandissilio.
Of the remaining 9:
- all have the place of birth as Llangyfelach (or the anglicised Llangafelach) or Morriston, that are also the place of residence, in the various corresponding census years;
- two have parents as David/Jane, one (my second cousin!) has John/Mary and six have no parents;
- eight have the year of death as 1914, one has no data - though a quick "burials" check would identify that the death is another person of the same age also on the 1911 census. That leaves four entries between 1918 and 1924 inclusive - I await the 1921 census!
In addition, John is a "Washer" or "Pickler" on census returns, his marriage cert and my gt-GM's birth cert and, having found his parents, a "Tin Lister" in 1851.
To be fair to others, this is my third variation made after I found his parents as a result of being intrigued by finding him living next door to an Elizabeth Davies (nee Jones I found from the children's birth index entries) from Rhydwilym in 1891 and I then found a Narberth 1839 entry for her birth with mother's maiden name of "David". That unlocked the 1841/1851 census entries etc.
I still have one problem, John is 22 on his marriage certificate and a widower! I'm going to need Devine intervention to solve this bit.
Footnote: Whilst writing this, I suddenly thought it was strange that the first three children appear to be named after grandparents but my gt-GM was not. A quick GRO check found a birth 1869Q2 (mmn: Davies and nicely placed between 1867Q3 and my gt-GM 1872Q3) and a death 1869Q3 age 0 for an Elizabeth Jones which would complete the grandparents names before the birth of my gt-GM in 1872! Interesting - now, can I get the finance manager to authorise some expenditure?
You could say it, but it wouldn't be true. Most LostCousins members know how to research their tree properly and are rightly sceptical about the information in Ancestry trees, but I doubt there are more than a very small minority who truly understand the limitations of DNA.
In any case I don't advocate the use of tree hints and don't look at them myself.