Login to participate
  
Register   Lost ID/password?
Louis Kessler’s Behold Blog » Blog Entry           prev Prev   

MyHeritage’s Record Matches and Discovery Settings - Sun, 23 Mar 2025

One of the best reasons to have your family tree on one of the major online sites MyHeritage, Ancestry or FamilySearch, is because of their record matching systems. They will go out and find records for the individuals in your tree from among the billions of records in their collections and make them available to you.

I’m not saying you shouldn’t go and specifically seek out records that you need. Not all records are online, and there are a lot of records that are not indexed or transcribed incorrectly so that no automated system will find for you. But what you get from record matches are all the straightforward records that you’d never be able to find for yourself in your lifetime. These record matches will also include some record collections that you never would have thought to look at. Automated record matching is such an innovation, that I would go so far as to say it has revolutionized genealogy.


Record Matches on MyHeritage

I last used a desktop program to record my genealogy in 1994. The program I was using at the time was Reunion for Windows. I had at that point connected 1361 people to my family. Over the next 20 years, I worked on my program Behold, expecting it to become the program I would use to be the program I would use to record my family.

But in 2017, the new trend of online record collections and record matching systems became an irresistible draw for me. I opened an account at MyHeritage, initially loaded my tree using a GEDCOM file, and started working on my tree there. Over the next six years, I was fortunate to obtain hundreds of Romanian and Russian records about my relatives in the 1800’s and greatly expanded my ancestry. Meanwhile, the MyHeritage record matches were so valuable to help me work down from my ancestors and discover 3rd, 4th and 5th cousins that I would never have found otherwise.

Currently, I have 8 trees at MyHeritage. My main tree is for my and my wife’s families as well as a place-to-place study and now contains 15,353 people. I have 7 other trees for in-laws and friends that have another 3,339 people. Here’s a bit of the history that I kept about the growth of my family trees at MyHeritage and the number of Record Matches that I’ve confirmed and added to my tree.

image

With MyHeritage, I try to keep up to all the record match suggestions it supplies. Every few weeks, I might get several hundred to several thousand new record matches to look through and process by either confirming or rejecting each one.

The source to person ratio is a metric that I remember Randy Seaver blogging about. In 2019 Randy reported a ratio of 1.94 from his RootsMagic program. I wonder how much he has raised his ratio since then. It’s important to try to ensure your tree has supporting records, and the ratio is a good way to check that you’re going in the right direction. I attribute the growth of my ratio primarily due to the growth of new record collections at MyHeritage.


MyHeritage’s Discovery Settings

When you look at your record matches at MyHeritage, there is a small “cog” on the page that you may never have noticed.

SNAGHTML1862d60

If you click on the “cog”, it brings up this window:

image

I don’t know when MyHeritage added this feature. I didn’t notice that the “cog” was there until sometime last year. The Introduction to Record Matches article by MyHeritage describes the “Confidence Scores” in more detail. The Discovery Settings allow you to set the minimum level of confidence of the record matches that you are given. You can also select only structured records, or only text records if you want.

The confidence level has 11 settings. From zero (lowest confidence) with increments of  0.5 up to 5 (highest confidence). The default is 0.5 which includes almost all the record Matches that are found. Increasing the Confidence Level will reduce the number of Record Matches you are presented with.

I tested most of the confidence levels on my own matches. Below are the numbers of record matches that MyHeritage gives me at each confidence setting. MH sometimes refers to a “Collection” as a “Source”.

SNAGHTML105a5cf

MyHeritage currently has 7466 different collections. Only 485 of those give me record matches at confidence setting 0 or higher.  Confidence setting 5 only gives me record matches from 107 collections.

These are the MyHeritage collections that give me the most record matches at confidence level 0:

image

Note that MyHeritage classifies Geni, FamilySearch and Filae matches as structured record matches. These are actually family tree sites, and genealogists know that other people’s family trees (especially unsourced ones) are not to be trusted until verified with records.

MyHeritage also finds matches with its own member’s family trees. It keeps those separate into what it calls “Smart Matches”. I have  23,678 smart matches with 5,124 trees at MyHeritage. Those are great for clues and really help put families together. MyHeritage is correct not to call those “records”. But IMHO I think MyHeritage should change Geni, FamilySearch and Filae matches to also be smart matches.


Confirming and Rejecting Record Matches

When I learned about the Discovery Settings last year, I changed my confidence setting from 0.5 to 0 and learned that I had 12,000 more matches to review. I wondered if anything with zero confidence was worthwhile spending the time on.

So I spent a few weeks reviewing 6054 of these matches. I confirmed only 2558 of them (42%) and rejected the rest (58%).

Technically, a confirmation of a match should mean that this record does indeed pertain to the person that the match says it pertains to. A rejection should mean that it is not the correct person. What I do is also reject records for the correct person if they are duplicates of another record. I only keep one. For example, the same newspaper article might be printed on consecutive days or in different newspapers.

The one person in my tree who has the most record matches is the astronaut Judith Resnik who was killed in the 1986 explosion of the Space Shuttle Challenger. Needless to say, there were many articles written and the initial articles were reprinted in newspapers everywhere. I have 1208 record matches for Judith and only confirmed 30 of them (2%), rejecting the rest (98%).

So I was wondering if the confidence levels that MyHeritage assigns to records correspond somewhat to the percentage that I have rejected. I downloaded some statistics from my matches and came up with this table:

image

What this says is that MyHeritage does quite well in providing relevant structured record matches, requiring rejection less than 20% of the time with a confidence setting of 0.5 or more. The Free Text records however are much less reliable, with confidence level 0 almost always being rejected.


Conclusion

There’s nothing like MyHeritage record matches to provide you with facts and sources you need to expand your family tree. I enjoy spending the the hours necessary to review new matches whenever they appear for me. It definitely is worth your time doing this!

If you haven’t already, check out the MyHeritage Discovery Settings. You may want to increase the confidence level or only work with structured records if you have too many pending matches. Or you may want to try decreasing the confidence level to 0 and see all the suggestions that MyHeritage has to offer. The choice is yours.

No Comments Yet

Leave a Comment

You must login to comment.

Login to participate
  
Register   Lost ID/password?