A sharp surplus bound for weighted Bernoulli sums with at most seven terms
For positive real weights summing to one, at most seven terms, common Bernoulli bias 0 < p <= 1/3, and every weight below p, a candidate exact computer-assisted argument gives non-strict mean-exceedance probability at least p + 3/256. The constant is a sharp uniform infimum for each fixed support from 4 through 7. An explicit restricted Manickam-Miklos-Singhi consequence follows.
Authors
- Robert Barritz
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-10-29
- DOI
- https://doi.org/10.5281/zenodo.23779014
- Citations
- 1
- Primary Topic
- Advanced Combinatorial Mathematics
- Type
- preprint