Scatter the Quadratic Bit: A Dose-Response for the Five-Round Core of a Toy Block Cipher, a Reduction to Unbalanced Pairs, and a Missed Seal
Our previous note showed, by sealed interventions on a 16-bit toy block cipher with the PRESENT S-box, that akey-independent five-round zero structure (the core) is carried by a permutation that gathers the single quadraticoutput bit of every S-box into one S-box. Here we ask how much gathering is needed. Random bit permutationsturned out to be the wrong place to ask: most of them diffuse so poorly that the core survives regardless. Restricted topermutations with full diffusion, the share of ciphers with a core rises with the number g of quadratic bits that meet inone S-box: in the sealed run 65 scattered ciphers (g = 1) had none, and the share was 0.174, 0.526 and 0.950 for g =2, 3 and 4. We prove a reduction: under two conditions met in all 1000 sealed ciphers, the core is decided by whichpairs of input bits of an S-box are unbalanced, and for the PRESENT S-box only three pairs matter. The claim thatscattering removes the core held in the prediction of the division property (100 of 100) and in measurement over 64keys (20 of 20), but it is not proved for the whole family. Eight claims were sealed and the hashes posted before therun; seven came out R. The eighth, made risky on purpose with a forecast of 0.50, came out F, the first miss in thisline after 18 R; eight forecasts are far too few to measure calibration, but a record that contains misses is where sucha measurement can start. The work stays on a toy cipher and published methods. (1) Text: CC-BY-4.0 | Code: GNU Affero General Public License v3.0 or later (AGPL-3.0-or-later) (2) Data and code. The seal document with its SHA-256 list, the frozen results and judgement, the development logs(DEVLOG_01-04) and scripts, the division-trail certificates, and the figure and paper scripts (AGPL-3.0-or-later) arein Bridge_Paper_Scatter_the_Quadratic_Bit_scripts_v1.0.0.zip. Hashes are posted atgithub.com/watabe-masanori/cds-polytope, issue #1. No external data are used. (3) AI use. Designs, code, analyses and text were produced by Kurado (Claude Opus 5.5, Anthropic) working with M. Watabe, who chose the direction and posted the hashes. Comments by ChatGPT and Gemini on a draft, relayed by M. Watabe, prompted Section 8; this is AI commentary, not peer review.
Authors
- Masanori Watabe (ORCID: https://orcid.org/0009-0000-4441-5126)
- Kurado (Claude Opus 5.5)
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-21
- DOI
- https://doi.org/10.5281/zenodo.24413838
- Primary Topic
- Cryptographic Implementations and Security
- Type
- preprint