The Boven Olifants cornetcy's surnames tested against the 44,143-person private tree - and a measurement of what a kinship distance is worth

tree query · a copy, index or transcription · consulted 2026-08-20

Type

tree query

Repository

Private family tree, ../the private research folder/source_files/the private family tree

Reference

Queried 20 Aug 2026 on full_name. Breadth-first distance from Frans () over 44,063 linked people.

Consulted

2026-08-20

The citation as it was given

Run at Frans's question: 'do any of the other names of this veldkornetskap show up in the 44k tree?' | *** TWO DATE-VERIFIED NAME MATCHES, which is the part that means something. *** NICOLAAS JOHANNIS GROBBELAAR,, died 19 April 1837 - the 'Nicolaas Joh:s Grobbelaar' who owned Nietgenaamd in 1802 (SRC-531). And MICHIEL HEYNS,, 14 April 1771 to 4 November 1856 - the 'Michiel Heyns' holding Gorgotha on the J 91 roll (SRC-525), who would be in his forties there. Both names match exactly and both sets of dates fit the document. | WEAKER: an Andries van Tonder (, no dates) against the Fontyn applicant of 1832 (SRC-530); a Lourens de Jager, but born 1711 and dead 1765, far too early for the 1832 Roodeheuvel leaseholder. No Anthonie Nortje at all. The Bothas are numerous and none matches Fredrik Jacobus, Johannes Petrus or Johan Adriaan. | *** AND THE MEASUREMENT THAT MATTERS MORE THAN ANY OF THEM: A KINSHIP DISTANCE IN THIS TREE IS ALMOST WORTHLESS AS EVIDENCE. *** All four 'connect' to Frans - Grobbelaar and Heyns at 8 steps, van Tonder and de Jager at 10. Then the distances were measured over the whole tree: | 5 steps 2.2% | 6 steps 6.9% | 7 steps 24.9% | *** 8 STEPS 65.5% *** | 9 steps 81.4% | 10 steps 90.0% | 12 steps 95.8%. | SO 'WITHIN 8 STEPS OF FRANS' DESCRIBES TWO-THIRDS OF THE TREE AND 'WITHIN 10' DESCRIBES NINE-TENTHS. The tree is one giant component and everybody is a cousin of everybody. Reporting these four as 'connected' would have been true and meaningless. | *** THIS BEARS DIRECTLY ON 6, which admits a witness as a person partly on 'a kinship path of <=8 steps to the root couple or to Frans'. THAT CLAUSE ADMITS 65% OF THE TREE AND THEREFORE FILTERS ALMOST NOTHING. *** It is the same shape as the finding already recorded there about clause 1 of the witness rule being nearly inert. The work in that test is done by the OTHER two conditions - the name being unique among 44,143, and the dates putting the person alive and adult at the occasion. The path should be treated as a sanity check, never as evidence, and the number should be published beside it whenever it is cited. | WHAT THE ANSWER TO FRANS'S QUESTION ACTUALLY IS: the surnames of this cornetcy are ordinary Cape surnames that his tree is full of, and presence in it shows nothing on its own. The two individuals above are worth pursuing because their NAMES AND DATES match documents, not because they are in the tree.

Kept verbatim. This project contradicts published genealogies and corrects a family's own oral tradition, so every claim has to be checkable by someone who was not in the room.

Nothing in the model cites this source yet.