General => Ancestral Family Tree DNA Testing => Topic started by: Essnell on Wednesday 11 May 22 01:27 BST (UK)

Title: Leeds method - help
Post by: Essnell on Wednesday 11 May 22 01:27 BST (UK)
Hi everyone,

Feeling a little dull with this and a tad stupid.

   I have started and finished twice, over two very long days, to sort the matches I have by the Leeds process . 
The explanation I found was fine once one got the hang of what was really needing to be done.  So... 

As far as I can see one starts with the first person and checks all the names on their match list ...
then proceeded to work through all the various persons on persons 1's match list    after that I then went to next vacant person on the original list and did similar.  This far I got on the second attempt

Now I started to find that some people were coming up more than once and some were completely new.     I eventually ended up with more colors than four so started allocating those overlaps to        the most matched color group of those specific set of overlaps.  Is this actually useful or is it going to give false ideas.    I actually ended up with 4 colour groups and very few not included anywhere so those marked for removal.
I have realized now that the method actually dose not include this allocating.

As I have very few high cM matches I used those down to 3rd --5th cousins.  I know that is not what usually happens but I wanted to see who in the match list could be a candidate for a pair of grandparents; so I could single out those belonging to my maternal Gt Great grandparents.  This appears to do that.

I thought I might then take that list and try to work out the male and female sides.

Maybe I am hoping for too much here.  Maybe I am totally barking up the wrong tree.  Mmmn

Any thoughts would be appreciated.
Title: Re: Leeds method - help
Post by: Ruskie on Wednesday 11 May 22 04:00 BST (UK)
Itís all double Dutch to me and looks dreadfully complicated, but is there anything here which addresses your issues:

From what little I can comprehend (and I admit that I didnít read the entire article) it looks like you need second cousins for this method to work properly, so if you only have more distant matches maybe that explains why you are having trouble?

Have you uploaded your raw data to FTDNA and My Heritage? I have no idea if they give the same or similar results as the Leeds method, but these two companies have autocluster charts and other tools which ancestry do not, which may help you. There is a fee to unlock these features.
Title: Re: Leeds method - help
Post by: Essnell on Wednesday 11 May 22 06:00 BST (UK)
Hi Ruskie,

Just checked that link and unfortunately I am in a place where this kind of connection just is a nightmare as we have dreadful internet connections. It is country Australia. 

Before I started I did find an explanation of the process with images from  a site but  My DNA just doesn't have any close relatives showing.  I used the closet I have. I did get what looks like four gt. grandparent groups but the one I need isn't at all useful.    Contains a family of 4 plus about 6 odd individuals.   No one other than the family seems to relate to any of the others.  I ran them through the Browser on My Heritage after posting.

I think one can go  to lower cMs but that gives a lot more individuals to deal with and therefore more groups to look at.

My data is from My Heritage but I haven't uploaded it anywhere else.  I actually did it to try to help someone else.  It's been useful to a degree there. I have run the AutoCluster tool twice, a year apart.  The clusters did alter with more matches coming on board.  They don't really help with this.

Looks like another dead end. This Gt Grandmother is so frustrating. 

Anyway after 20 yrs I can only keep looking.

Cheers Essnell  and thank you.

Title: Re: Leeds method - help
Post by: Ruskie on Wednesday 11 May 22 08:54 BST (UK)
I hope when the UK wakes up that youíll get some proper help from the DNA experts.

Title: Re: Leeds method - help
Post by: Biggles50 on Wednesday 11 May 22 22:25 BST (UK)
Please read how I used the Leeds(ish) Method and other visual aids in this recent thread.

Unfortunately the way the Leeds Method is portrayed is that it works, where in reality it does but there are many ifs and buts.

In an ideal scenario you use 2nd - 3rd Cousins so that is about matches in the 90 to 400cM range.

In that range I have 10, of which 4 are in my tree and linked to my Motherís side, so you would assume that the other six could be either. 

Not necessarily the case!

I have a whole mass of Italian and another batch who are Irish, so somewhere there is a NPE and there is a Great Grandmother whose father is unknown.

Then there are the matches who have no trees and user names that are a jumble of letters so the only use is if there are shared matches.

You are on the right track, expand your chart with more colour coded columns.  Then where you have matches that are already in your tree, highlight them as you know which parental side they are on.

Work through the chart trying to assign, using the shared matches and your known linked matches which Grandparent a match aligns with.

A way of doing this is to add four more columns to your chart and label them PGF, PGM, MGF, MGM ie one for each Grandparent.  Then go down each match and assign a colour to the cell under the applicable Grandparent, leave the cell blank if you cannot assign to any ofbthem

When you have done this you can hide the other coloured cells and there is your Leeds Method Chart, completed as much as you can.  It probably will have massive holes ie uncoloured cells but it is what it will be.

Off to bed now so Iíll drop by tomorrow to try and help with any queries

Hope this makes sense to you.

Title: Re: Leeds method - help
Post by: Romilly on Wednesday 11 May 22 22:32 BST (UK)
Iíve never been able to get anywhere with the Leeds methodÖ

I have tried it, but it doesnít seem to work for what Iím looking at.

I have a long-standing Brickwall with my paternal grandfather, who resists all efforts to be found 😵‍💫.

Essnell, - have you tried doing a WATO Chart? It stands for What Are The Odds, and is based on probabilities. Iíve been having a go and can post if it would be of any assistance.

Title: Re: Leeds method - help
Post by: Essnell on Thursday 12 May 22 00:50 BST (UK)
Hi everyone , After some sleep I can think clearly and unstressed...

To Romilly...          I tried the WATO tool too but that was for something different andIi simply don't have  enough information about the particular "Hole". 
I've left it for use later as i think this Leeds thing is possibly more useful for sorting the lines of the matches.  Thanks for suggesting.

Biggles50.  Thank you. Yes that all makes sense.   I have printed you reply so I can access it while working.  Need yet to read the link. 
You have given me the incentive to try  again...   additional ideas are often very helpful. 

Thank you I'll be back later... :)
Title: Re: Leeds method - help
Post by: Rosinish on Thursday 12 May 22 02:12 BST (UK)
Don't forget it's likely certain lines haven't done DNA testing i.e. a lack of matches.

I have numerous links to my DNA but for the life of me, I'm struggling to have matches on my maternal side on all sites I've uploaded to.

I do have the occasional link but nothing worth celebrating.

I've come across links in my DNA who have no other matches  ???

It's annoying but fascinating too.

I find many have either no trees or very little in generations to work out anything as surnames are unknown (probably female connected) & only go as far as their g/parents.

I've come matching DNA trees with names only i.e. no dates or places & have to rely on shared matches to know which side they belong.

It's about time these sites made it compulsory to prove connections with names/dates/places prior to adding useless info.  ::)

Title: Re: Leeds method - help
Post by: Essnell on Thursday 12 May 22 07:43 BST (UK)
Hi Roshinish

I am dead sure some, on the line I am hoping to shed some light on, haven't tested.  Hopefully though some of these many cousin matches might be part of that.

So far in this analysis I have found several people on this particular line and a few  that overlap.   That in itself is interesting.  I am hoping it leads somewhere.  Again the trees to the matches aren't all that helpful.

At this point the little information I have regarding Gt Grandparent through my mother is so limited and all unproven that it's a nightmare even to think of adding it to my tree. 

Your thoughts are just so true and on other research I have done that has been a massive issue.  If I can, I try to verify what I find IF it looks like a possibility to be correct.

I doubt any of the sites could police the accuracy of trees they carry.   One can only wish.

Title: Re: Leeds method - help
Post by: Biggles50 on Thursday 12 May 22 12:20 BST (UK)
Over on Gedmatch the videos there are created by Family History Fanatics and finding them on Youtube gives a very useful resource in learning not only how to use Gedmatch but how to create and modify the Leeds Method chart.

The video in the link explains it far better than I can, but do look at the second video on Modifying the chart as that is imo more useful after you get the basics.
Title: Re: Leeds method - help
Post by: Essnell on Thursday 12 May 22 14:57 BST (UK)

Thank you for the link Biggles50.  I will see if my internet will accommodate the videos.  It's limited a great deal by speed and as I have not got normal internet but depend on a dongle to Wifi it is often slow or none at all.  It's just as well I'm not trying to use it for work.

My Spreadsheet is almost done.  The persons who only have themselves I shall remove as they aren't going to add anything.
There is quite a bit of overlap so I reckon there has to be some cross-link somewhere.  I will need to print this on a much larger paper than A4,  so I can see the whole picture. 

I can definitely see that there isn't much on the line I am after.

Title: Re: Leeds method - help
Post by: Essnell on Friday 13 May 22 13:37 BST (UK)
Hi everyone.
Something odd.  From the Leeds Chart I have been compiling.

I have come across a match  of a match that matches to me within the range I chose to use.  Now this person is not on my list from the testing company yet they are definitely within this family.  Their match list comprises almost exactly to many others Lists and in the same order as well as to mine. 

I  am supposed to match to all of these other persons which would be right according to my paper information. 

Anyone any ideas as to why under these circumstances I don't have this person on my match list.
The match lists are from one compamy, I presume, unless uploaded data from a different company, makes things different.

Title: Re: Leeds method - help
Post by: Biggles50 on Friday 13 May 22 14:25 BST (UK)
DNA matches on Ancestry have tested with Ancestry, they currently do not have any uploads possible from the results of other testing Companies.

Now a person can have tested with Ancestry but chosen in their privacy settings not to be seen by anyone.  Or they may share such a low cM with you it is below the threshold that is displayed as a direct match with you.  I have one match with 6cM which is now linked in my tree, I linked them before Ancestry changed their system and said match would not now be displayed.

Moving on to upload comparison sites like Gedmatch, myfamilytreedna etc raw DNA data can be uploaded to these sites irrespective of which was the testing Company, in Gedmatch using their one to many tool if you scroll along to the right you see the testing Company listed.

You may also find that a given match on Ancestry has a very different cM figure on the comparison sites, apparently this is due to the maths involved in calculating the presented data.

With my own brick wall I have matches listed on my Mystery Tree that all link together yet I gathered their info and our shared matches from three different comparison websites.

Without more info on the sources and cM values its difficult to be more accurate.
Title: Re: Leeds method - help
Post by: Rosinish on Friday 13 May 22 15:16 BST (UK)
Your thoughts are just so true and on other research I have done that has been a massive issue.  If I can, I try to verify what I find IF it looks like a possibility to be correct.

I doubt any of the sites could police the accuracy of trees they carry.   One can only wish.

 ;D It was a 'tongue-in-cheek' comment...of course they can't be policed (such a pity) but in an ideal world it would be beneficial to all involved in 'serious' research, alas  ::)...

Title: Re: Leeds method - help
Post by: Essnell on Saturday 14 May 22 01:47 BST (UK)
Hi Rosinish, 

Yes, I know and I didn't mean to be critical, sorry.  It would be great if they could.  Almost an impossible wish.

Title: Re: Leeds method - help
Post by: Essnell on Saturday 14 May 22 03:50 BST (UK)
Hi Biggles50,
Thank you for the reply. 
I have only tested with MyHeritage so all my match info comes from them not Ancestry. 
I've done some checking.

I didn't add her to my Leeds list, as on my MyH  site match  list she is marked 3rd - distant  -  but  they could be wrong with that seeing  that all the matches she has  are many of the same people from others; Several of whom I have full trees for, which are connected to me.   

My thoughts are that the possibility lies in that these descend from my Great Grandparent's siblings.

My Highest cM Value starts at 202.4cM  so nothing spectacular.  There are exactly 8 before it drops below 90cM. All of them are known.     So : two from one line.    4 from another.  1 each from the remaining two names.   I know where these all fit: one  single unit I know - Its my grandmother's family on my fathers side, the other with one known cousin is my grandmother's family on my mother's line.

 I am trying to find who in this massive full list of 5000 Plus belong to grandmother's family on my mother's line.  I was hoping to do that in order to work from there and try to connect to mu Great Grandmother's line .
 I have the full tree for this cousin - single match and oddly no links to any of that .
 I have two links  that are of matches of matches of this single person:  One I have made contact with but owing to a death there is a delay in getting any dialogue going.  The other is at a stand still - no where to go.

Ok! I think I have confused everyone including myself.