severe calculation logic failure
When processing highly endogamous lines, the background relationship engine is failing to prioritize direct, vertical pedigree lines. Instead, it gets trapped in lower mid-19th-century cousin loops, causing it to either assign a flawed sideways shortcut or completely collapse into an "Unknown Relationship" failure further up the tree of the automated relationship calculator on the FamilySearch platform
All biological parent-child links are perfectly attached on these profiles, proving the tree architecture is correct, but the algorithm is broken.
Please investigate these two specific diagnostic test cases:
Case 1: The Sideways Shortcut Loop
The Target Profile: Jonathan Richardson Sr. (P66Y-41M) and Elizabeth Bass (P66Y-JX1).
The Glitch: The system tracks them as a "6th great-grand uncle/aunt."
The True Pedigree: Vertical pedigree mapping and 7-generation Fan Chart tracking prove they are my direct 4th-great-grandparents. The algorithm is bypassing the clean vertical line running through their daughter, Elizabeth Betsy Richardson Moore (P4R7-RVG), and her husband James Moore (P4R7-JBC).
Case 2: Total System Collapse ("Unknown Relationship")
The Target Profile: Henry Bunch (G8DW-QX4), Nancy Bunch (LZDH-8N9), and John Bunch (LLHP-JQB).
The Glitch: The system tracks them as "Relationship Unknown."
The True Pedigree: Unbroken parent-child links prove that Nancy is my direct 6th-great-grandmother, Henry is my 7th-great-grandfather, and John is my 8th-great-grandfather. The underlying cousin loops down in the 1800s (around the Moore/Boon lines) are causing the background calculation script to crash entirely rather than reading the clean vertical links.
Thank you,
Gary71311
[name edited for privacy]
Answers
-
I have been disappointed from the beginning that the default relationship between a parent and child will be no relationship. If you want to show biological, adoptive, foster, guardian, you must manually set it. Thus, for most parent child relationships, there is no relationship type selected since most people don't know they have to take this extra step.
That is why the relationship finder algorithm used in a variety of popular uses, such as finding your relatives at RootsTech, ignores relationship completely. For years, my closest relative at RootsTech has been the descendant of a couple who fostered my great grandmother. Lovely.
So for your 2 cases, the fact that you have established biological relationships is irrelevant.
One other factor that may be involved is when more than one set of parents or marriages is established, and "preferred" is set to one of them, does that play in the algorithm? I don't know the answer to that one. I have step relationships established, which will give the child 2 sets of parents. I do that because I have personal knowledge that the step parent was very active in the step family and is a big part of the family history. But that creates other problems and might affect the algorithm.
0 -
Thank you Gail...appreciate your attempt at this one. Was hoping to get this flagged to march this up the chain, as its really causing problems with my tree
0 -
@sc woz sounds like this needs escalating please
0 -
Your two diagnostic cases are not user error — they match a known, reproducible failure mode in the FamilySearch relationship‑calculation engine when endogamy and mid‑tree cousin loops overwhelm the path‑finding logic. The FamilySearch Community thread you referenced documents the same symptoms, including the exact profiles and miscalculations you’re seeing.
I will send this up to the appropriate experts. I have no way to determine when or if an answer is forthcoming. Hoping it will be soon for you. Thank you for reporting this to FamilySearch and making this a better product.
2 -
We appreciate your feedback. This appears to be resolved. Case 1 looks like it is working as it should with both persons showing direct line relationships. Case 2 also shows all three persons referenced as showing direct line ancestors.
0
