It would be nice to have some sort of filter for hints. I’m not sure how it could work. Last week Ancestry were really exited to tell me they had found a new DNA shared match hint for me. Little did they know that I had found this match on my own a couple of weeks previously. My new found cousin then added this information to their tree which eventually must have shown up in Ancestry’s searches.
-
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 Public Trees versus Private - a new debate
- Thread starter Bob Spiers
- Start date
Helen7
LostCousins Superstar
Apart from unchecking the 'Display Ancestry member tree hints in your trees' box in Site Preferences, which stops you seeing any hints from other trees, I can't see a way of filtering out the rogue hints.
So far, I haven't found spurious hints to be a problem since attaching a public direct ancestor tree to my DNA about 10 days ago. The only extra hints I have had are from my own public tree to my private tree. That said, it's still early days and my direct ancestors don't have particularly common surnames (with one or two exceptions which I will keep an eye on). However, my husband is half Welsh and has lots of people named Davies and Jones amongst his direct ancestors. The tree attached to his DNA is still private, and after hearing of your experience, is likely to stay that way.
Last edited:
No. I'm talking about where they have added an incorrect person which is now coming up as a hint for me. So it's wrong for them also.
It might take a few weeks. And not all of my Joneses have been copied like that so you may be lucky.
A
At home in NZ
Guest
Are you talking about the new Green Hints, called Potential Ancestors?
I got sick of them very quickly because they are taken from other trees and in may cases when i looked at the tree there weren't any supporting records.
I found out how to turn those off as well as other Hints, go to your Site Preferences and select/deselect whatever you want.
Hope that helps.
jorghes
LostCousins Superstar
And even just correcting mistakes in place names on FTM and then syncing makes another 100 appear; or briefly checking on a branch that you cleared of hints and the next second Ancestry happily gives you another 60 odd to trawl through.... I luckily currently have less than 7000 hints... but most of them are only worth an "ignore".
A
At home in NZ
Guest
I also sync with FTM but i use Ancestry for my data input and sync that. I don't have any issues with place names because I use the actual physical address and ignore any suggestions made by Ancestry.
Helen7
LostCousins Superstar
Yes, I've noticed that the slightest update to the tree causes more hints to appear. I've given up trying to clear them now. I have over 8,000 but don't pay any attention to most of them. And I agree with At home in NZ that the green 'potential ancestors' are a waste of time and I have turned them off too.
jorghes
LostCousins Superstar
I don't use exact addresses, and I like the place to get the little green tick for the Google Maps matching on FTM - but then all I did just then was clean up my five or so versions of "Amsterdam" that were in my tree to just the one... It's still syncing to adjust... so this should be interesting!
Does it ever... I like the "Potential Ancestor" hints - to view, consider, laugh at and ignore. There's a persistent presumption that my 5th great grandfather (the furthest I have managed to get my tree on the direct paternal line) is the son of person X and person Y... but considering I have the death certificate for that particular son (who has the same name) of persons X and Y and he died at the age of 18mths in the same year that my 5th great grandfather died at the age of 74, then them both being the same person is rather slim. I suppose X could be a son (doubtful), grandson (considering I have found the majority, unlikely) or nephew/cousin (possible) but I have yet to find the connection... Most people have them with birthdates far to young to be the parents of the children they have added.
Helen7
LostCousins Superstar
Yes, I was amused by these at first, and noticed several complete mismatches along the lines you describe, but it got quite wearing after a while so I turned them off. What I find worrying is if people happily add these 'potential ancestors' into their tree without bothering to check whether there are supporting sources or even if they make any sense in terms of birthdates, locations etc.
Last edited:
Bryman
LostCousins Megastar
Might this be the source of a request for an improvement at Ancestry? The number of hints would be cut greatly if a check were to be performed before the hint is issued. I raised a similar request at FMP when they started to issue hints. Almost all of my hints were duplicates of facts already incorporated in my tree so I ignored them all.
A
At home in NZ
Guest
Most definitely. They need a decent algorithm as it's obvious the one they are using now is not accurate.
It annoys me when I get a hint that is a totally different person from the one it's hinting about and/or the person died years before the date of the hint.
What are hints? In effect they're search results. The advantage of hints is that we don't have to carry out the searches - Ancestry's AI is doing it for us, constantly. Other applications of the same principle are DNA matches and LostCousins matches.
The disadvantage of hints is that even when we carry out searches ourselves, most of the results we get aren't relevant. So unless the AI that regulates hints is better than the AI that produces search results there's always going to be a problem with hint overload and underwhelm.
There are three situations that spring to mind in which hints are likely to be particularly useful:
What might help is if priority was given to hints from trees of descendants of the individual who is the subject of the hint. We all spend more time and money researching our direct ancestors, and we're more likely to pay for certificates and track down their baptisms.
Something else that might help is if tree owners were able to objectively rate the quality of their information on an ascending scale from:
The disadvantage of hints is that even when we carry out searches ourselves, most of the results we get aren't relevant. So unless the AI that regulates hints is better than the AI that produces search results there's always going to be a problem with hint overload and underwhelm.
There are three situations that spring to mind in which hints are likely to be particularly useful:
- when they relate to a record set which is new or has been updated, or which we hadn't thought to search
- when they provide a possible solution to a 'brick wall', perhaps based on the tree of another user
- when they warn us that an assumption we've made might be wrong
What might help is if priority was given to hints from trees of descendants of the individual who is the subject of the hint. We all spend more time and money researching our direct ancestors, and we're more likely to pay for certificates and track down their baptisms.
Something else that might help is if tree owners were able to objectively rate the quality of their information on an ascending scale from:
- I don't know where this information came from
- I chose this record because it seems to fit
- I chose this record because it's the only one I could find that fits
- I have supporting evidence that strongly suggests this is the right record
- I know this is the right record and can prove it
- I know this is the right record and can prove it, but others might come up with a plausible alternative if they haven't got access to the evidence that I've seen
Pauline
LostCousins Megastar
I have hints from other trees turned off, though I still get hints for stories, picture etc in other trees.
I review all the other hints I receive though as yet I haven't been alerted to any information that I didn't already know about. I regularly receive hints of of events that are already entered in my tree.
Also, if I ever did get a hint with new information, I don't think I would click to add it straight into my tree. Most likely I would need to do a fair bit of checking around before being certain it really did relate to the person in question.
I review all the other hints I receive though as yet I haven't been alerted to any information that I didn't already know about. I regularly receive hints of of events that are already entered in my tree.
Also, if I ever did get a hint with new information, I don't think I would click to add it straight into my tree. Most likely I would need to do a fair bit of checking around before being certain it really did relate to the person in question.
I haven’t been using hints recently as they are overwhelming and generally not as useful as in the past. I think algorithm changes may have made things worse. I was just wondering whether I was missing something and that there might be some bright ideas.
I know I’m missing hints to new and updated record sets, but I like Peter’s list of situations where hints would be useful. It would be good to see activity on particular trees too but tree sharing might be more appropriate there.
Information quality rating would be nice.
I know I’m missing hints to new and updated record sets, but I like Peter’s list of situations where hints would be useful. It would be good to see activity on particular trees too but tree sharing might be more appropriate there.
Information quality rating would be nice.
A
At home in NZ
Guest
Not only do I not fully understand how i can prove a record, for instance many Hints are census records which are proof (of residence) themselves.
How do i prove i have the right baptism record for an ancestor born in the 1700's for example.
Where would we put the rating? This would also mean an enhancement to FTM so that the proof can be synced.
Also, isn't it a bit late for tree owners who have thousands in their trees? What is the likelihood of them going back over very record they have, I certainly would not.
I would put this enhancement in the nice to have category.
There would be no need to rate every entry respectively - in practice it will only be the ones with the highest level of certainty that matter. Prior to 1837 we're only likely to have that level of certainty for our direct ancestors (and perhaps their siblings), and we won't have it for all of them.
PhilGee
LostCousins Member
The "quality" list Peter suggests is interesting, but I wonder how popular it would be. Those who carefully document the information are also the one's most likely to rate their information, whilst those who do not take care probably would not or be ignored on the basis that their ratings are of the same quality as their information.
As I keep my database as a series of spreadsheets, I have a coding system for "vital" records (Q=query, X=ignore as probably incorrect plus others) and can add comments to most records. "Q" results in "needs verification" in the Ancestry "description" box and comments also appear there (such as "date from:", which covers certificate, parish record, Ancestry tree by... , ...). I also add items, such as "name on census", and cover possible name inconsistencies, such as Alice Parsons is probably Louisa Parsons in the birth index (she is also the younger sister of a Louisa) - her eldest sister has Kate Elizabeth probably "Elizabeth Alice Maud Mary"* on the index and "Kate Greenaway"* from the baptism transcription (parents' names fit and she was baptised at the same time as Louisa, I shall, sometime, get more definitive! - famous last words); the baptism is interesting as her maternal aunt is "Greenaway" by marriage.
Phil
* sole occurrence of this name combination
As I keep my database as a series of spreadsheets, I have a coding system for "vital" records (Q=query, X=ignore as probably incorrect plus others) and can add comments to most records. "Q" results in "needs verification" in the Ancestry "description" box and comments also appear there (such as "date from:", which covers certificate, parish record, Ancestry tree by... , ...). I also add items, such as "name on census", and cover possible name inconsistencies, such as Alice Parsons is probably Louisa Parsons in the birth index (she is also the younger sister of a Louisa) - her eldest sister has Kate Elizabeth probably "Elizabeth Alice Maud Mary"* on the index and "Kate Greenaway"* from the baptism transcription (parents' names fit and she was baptised at the same time as Louisa, I shall, sometime, get more definitive! - famous last words); the baptism is interesting as her maternal aunt is "Greenaway" by marriage.
Phil
* sole occurrence of this name combination