Orphan primes of the centered Goldbach decomposition
Every prime v ≥ 7 is written v = q1 + q2 + 1, where (q1, q2) is the most central Goldbach partition of v − 1; q1 and q2 are the two children of v. Some primes are the child of no prime. We call them orphans (463, 607, 691, ...). Up to 2.5·10^9, they make up about 7.3% of the primes, and this share is still slowly increasing; 97% of them are congruent to 1 mod 6. The fate of a prime depends mostly on the small factors of q + 1, and also on its prime neighbours at distances that are multiples of 6, which take parents away by competition. Two tests locate the origin of these effects. With the rule v = q1 + q2 + c, the role of q + 1 passes to q + c. And pseudo-primes sieved up to 1000 reproduce, up to 10^14, the share of orphans in each class and the dispersion of the number of parents: these depend only on the rule, the density of the primes and their small factors. A heuristic model that accounts for the minimality of the gap to the half predicts the mean number of parents and the effect of isolation. For pseudo-primes sieved up to 47, it also predicts the dispersion of the number of parents, and the resulting share of orphans to within a few hundredths of a point; the dispersion of the true primes remains unexplained. We also state a precise conjecture on the limit law of the number of parents. Unless a proof is given, the results are numerical or heuristic. The record also contains the working version of the article (article_en_travail.pdf), which adds the multiplicities and levels of the trees, the gaps L in detail and the Sophie Germain paths, and an archive with the scripts and their outputs.
Authors
- NIZAR MONTACER
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-15
- DOI
- https://doi.org/10.5281/zenodo.24245673
- Primary Topic
- Analytic Number Theory Research
- Type
- preprint