How to read →The arrow is not arithmetic. It means: write the number in the exact Hebrew form shown, then calculate the gematria of those words.3641 → „שלוש שש ארבע אחת” → gematria 1918
SUPPLEMENTARY CLOSURE · OUTSIDE 6/6 AND THE NULLFamily bridge — both surnames converge on 2087, then yield Rose Harpaz, the eldest daughter of Luba and Eran
This compact structure is shown before the six Null-tested cores because it is easy to read and captures the family sequence. It uses Shpanberg because that surname remained part of Luba's full name after the wedding. It is not counted in the 6/6 target and is not included in the p-values below.
Groom's surnameהרפז = 292292 → „מאתיים תשעים ושתיים” → gematria 2087
same target2087
Luba's surname, retained as part of her full name after the weddingשפנברג = 635635 → „שש מאות שלושים וחמש” → gematria 2087
What does the convergence mean? 292 and 635 are not added together. Each value is written independently in the Hebrew wording shown; both wordings have the same gematria: 2087.
From the shared node to Luba and Eran's eldest daughter2087 − 970 − 612 = 505
970 = לובה שפנברג הרפז — bride full-name form, where Shpanberg remains part of her name after the wedding612 = ערן הרפז — the groom505 = רוז הרפז — the eldest daughter of Luba and Eran
Plainly: the groom's surname and Luba's surname independently reach the same 2087. Because Shpanberg remained part of Luba's name after the wedding, it continues to participate in the family structure. Subtracting the two parent identity values from that shared node returns exactly רוז הרפז = 505, the eldest daughter of Luba and Eran.
Methodological status: a strong supplementary family closure, displayed separately and not used in the six-core Null result.
Wedding question„מי התחתן בתאריך שבע עשרה בדצמבר אלפיים עשרים ואחד” = 3641
↓ digit wording
3641 → „שלוש שש ארבע אחת” → gematria 1918
=
Ketubah record„שמנה עשר אלף שקלים חדשים” = 1918The full wording „שמנה עשר אלף שקלים חדשים” is shown exactly as it appears in the ketubah.
Anchors used3641 — the full wedding question with the masculine year wording.1918 — the amount phrase recorded in the ketubah.
Plainly: an event-specific question maps in one declared transformation to a value already carried by an external document from the same event.
Route A · question„מי התחתן בתאריך שבע עשרה בדצמבר אלפיים עשרים ואחת” = 4037
4037 → „ארבעת אלפים שלושים ושבע” → gematria 1898
4037 → „ארבע אפס שלוש שבע” → gematria 1422
1898 + 1422 − 983 = 2337Route B · groom + bride + date612 = Eran Harpaz · 983 = Liubov Shpanberg Harpaz
1712 = 17.12 DDMM code · 970 = Luba Harpaz Shpanberg
612 + 983 + 1712 − 970 = 2337Route C · ketubah + bride„שמנה עשר אלף שקלים חדשים” = 1918 — the amount written in the ketubah.
1918 → „אחת תשע אחת שמונה” → gematria 1989
348 = Liubov Harpaz
1989 + 348 = 2337 All three routes meet at23371898+1422−983 = 612+983+1712−970 = 1989+348
Anchors used4037 — full wedding question with the feminine year wording.612 — Eran Harpaz.983 — Liubov Shpanberg Harpaz.970 — Luba Harpaz Shpanberg.348 — Liubov Harpaz.1712 — 17.12 as DDMM.1918 — „שמנה עשר אלף שקלים חדשים”, the amount written in the ketubah.
Plainly: question, identities, date and ketubah do not terminate at unrelated values; three routes land on the same node.
2337 − 1918 + 810 = 1229=
„יום החתונה של ערן ולובה” = 1229Anchors used2337 — the triple-convergence node from closure 02.1918 — „שמנה עשר אלף שקלים חדשים”, the amount written in the ketubah.810 — „אלפיים עשרים ואחד”, the masculine wording of 2021.1229 — „יום החתונה של ערן ולובה”.
Plainly: the central convergence node does not stop as an unlabeled intermediate value; with the ketubah and the wedding-year wording it returns to a phrase explicitly naming the wedding day.
Shared target„אלפיים עשרים ואחד” = 810
Question − year2831 − 2021 = 8102831 = „מי התחתן בתאריך שבע עשרה בדצמבר”.
Question − groom4037 → 1422
1422 − 612 = 8101422 comes from „ארבע אפס שלוש שבע”; 612 = Eran Harpaz.
Date − bride1712 → 1158
1158 − 348 = 8101158 comes from „אחד שבע אחד שתיים”; 348 = Liubov Harpaz.
2831−2021 = 1422−612 = 1158−348 = 810
Anchors used2831 — base wedding question.4037 — full question.2021 — wedding year.612 — groom.1712 — 17.12 DDMM.348 — bride.810 — Hebrew wording of 2021.
Plainly: question–year, question–groom and date–bride are different anchor pairs, yet all three return to the same representation of the wedding year.
Step 1 · identity983 − 612 = 371עדה הרפז = 371
Step 2 · return371 + 371 = 742
1712 − 742 = 970לובה הרפז שפנברג = 970
Anchors used983 — Liubov Shpanberg Harpaz.612 — Eran Harpaz.371 — Ada Harpaz.1712 — wedding-date code.970 — Luba Harpaz Shpanberg.
Plainly: the difference between the bride-name form and the groom equals the groom's mother's name; that same value is then used with the wedding date to return to the bride-name form.
2831 + 983 − 1712 = 2102
↓ write 2102 in Hebrew words
2102 → „אלפיים מאה ושתיים” → gematria 983
=
ליובוב שפנברג הרפז = 983
Anchors used2831 — base “who married?” question.983 — bride identity.1712 — wedding date.
Plainly: the question asks who married; after combining it with the bride identity and the wedding date, the route returns to that same bride identity.
Full event-anchor ledger
Only source anchors are listed here. Derived representations are explained inside the closure where they are used.
612Eran Harpaz — groom
348Liubov Harpaz — bride
970Luba Harpaz Shpanberg
983Liubov Shpanberg Harpaz
371Ada Harpaz — Eran's mother
171217.12 — DDMM date code
2021wedding year
2831„מי התחתן בתאריך שבע עשרה בדצמבר”
3641full question, masculine year wording
4037full question, feminine year wording
1918ketubah amount phrase
1229„יום החתונה של ערן ולובה”
Before the statisticsWhat exactly was tested in the simulation?
The goal was to test whether a network like the wedding network appears easily when the event numbers are replaced by similar alternative values — without changing the rules after seeing the outcome.
1The structure was frozenThe six core closures kept the same roles and operations. A new formula was not invented separately for each alternative world.
2Alternative worlds were generatedIn each world, numerical anchors were replaced by nearby values. Three separate ranges were tested: ±5%, ±10%, and ±25%, while preserving data roles, the question template, calendar rules, and transformations.
3Each world was scored from 0 to 6The real case satisfies 6/6. Twenty million samples were run in each range, and every sample was checked for how many of the six structures survived together.
4The real case was compared with the alternativesIn the 20 million samples of each range, the best result was only 2/6, with no full 6/6 reproduction. The Null space was then examined by exact counting as well.
In other words: the simulation did not ask whether one could find something interesting in other numbers. It asked a narrower question: how often does this same six-closure architecture reappear when the event data are changed under the same rules?
FROZEN EVENT-LEVEL NULL · DEPENDENCY-COMPRESSEDNull test — observed 6/6 versus alternative worlds
Algebraically dependent lines were not counted as independent evidence. The scoring target was frozen to the six core structures shown above.
±5%2/6maximum in 20M · 0 full hits±10%2/6maximum in 20M · 0 full hits±25%2/6maximum in 20M · 0 full hits 60 million draws total across three separate Nulls. They are not pooled into one p-value because they are different distributions.
Exact enumeration · ±5%1 / 9,468,289,223,531,775≈ 1 in 9.47 quadrillionp = 1.056×10⁻¹⁶ · one complete world: the observed case
±10%p = 1.72×10⁻¹⁸≈ 1 : 5.81×10¹⁷
±25%p = 7.26×10⁻²¹≈ 1 : 1.38×10²⁰
Two more conditioned, conservative benchmarks
Fix ketubah 1918 and event phrase 1229 in advance≈ 1 in 399 billion1 : 398,849,539,725Fix all external event data and randomize identities only≈ 1 in 8.34 million1 : 8,341,515 · one full hit in a separate 20M Monte Carlo run How to read this?The 1-in-9.47-quadrillion figure belongs to the fully specified Frozen Event-Level Null. The conditioned benchmarks deliberately fix more external facts first; the network remains rare inside those narrower models too.
Scope: these are benchmarks for declared frozen Null models, not a global p-value corrected for every possible event, wording or search that may have preceded discovery.
Summary
The core finding is a six-family network rather than one isolated equation: question↔ketubah, the 2337 triple convergence, return to the wedding-day phrase, the 810 triple convergence, the Ada/date/bride identity chain, and the “who married?” return to the bride.
6/6 core structures1 / 9.47Q full Frozen Null1 / 8.34M identity-only conditioned benchmark
