hunting-hadamards
census · order 59

vxvware-kkss-n59

reproduced
Challenge resultNot yet published

Complete vxvware censuses past the range classified by Đoković and Kotsireas [11]: four classes at n=51, three at n=55, none at n=57, and eight quadruples at n=59 that the source counts as two classes and equiv-v1 as one. Existence at 51 and 55 was already known from Spence's infinite family [23] (q = 101 and 109); the class counts at 51 and 55, the nonexistence at 57, and the existence and count at 59 are new. Replays with pinned engines between 2026-11-14 and 2026-11-15 returned the same classes at each order, so the four counts are exact. The empty cell at 53 is not a search result: a (kkss) parameter set exists only when 2v - 1 is a sum of two squares, by the parameter-set condition of [11], and 105 is not.

Original source attribution: vxvware
Submitted / imported by: vxvware · 2026-10-15
Board cell: G-matrices (kkss)
Entry file: census/vxvware-kkss-n59.json · raw JSON

Notes

The source reports a full census with 2 classes under its own equivalence convention and stores 8 quadruples. This board verifies the 1 equiv-v1 representative and records a lower bound; the search has not been replayed here. A single run, calibrated beforehand on the stored v=43 and v=45 censuses. The source counts two classes under its own convention; under equiv-v1 the eight rows fall into one class.

Checks and result

{
  "id": "vxvware-kkss-n59",
  "type": "census",
  "n": 59,
  "stratum": "kkss",
  "equivalence": "equiv-v1",
  "count": 1,
  "bound": "exact",
  "tier": "reproduced",
  "claimed_count": 1,
  "claim_kind": "lower-bound",
  "checks": [
    "pairwise-inequivalence",
    "registered-full-census-replay",
    "representative-existence"
  ],
  "replay": {
    "engine_id": "kkss-census-c-v3",
    "host": "bos-game",
    "threads": 16,
    "date": "2026-11-14",
    "wall_seconds": 9005.0,
    "raw_rows": 4
  },
  "limitations": [
    "Full offline replay of a pinned engine, re-derived from its raw rows here; the same engine can reproduce its own systematic bugs."
  ],
  "contributions": [
    {
      "id": "kkss-new-orders",
      "title": "G-matrix censuses at n=51, 55, 57, and 59",
      "relation": "participant-result",
      "publication_status": "unpublished",
      "contributors": [
        {
          "name": "vxvware",
          "github": "vxvware"
        }
      ],
      "summary": "Complete vxvware censuses past the range classified by Đoković and Kotsireas [@djokovic2018goethals]: four classes at n=51, three at n=55, none at n=57, and eight quadruples at n=59 that the source counts as two classes and equiv-v1 as one. Existence at 51 and 55 was already known from Spence's infinite family [@spence1977skew] (q = 101 and 109); the class counts at 51 and 55, the nonexistence at 57, and the existence and count at 59 are new. Replays with pinned engines between 2026-11-14 and 2026-11-15 returned the same classes at each order, so the four counts are exact. The empty cell at 53 is not a search result: a (kkss) parameter set exists only when 2v - 1 is a sum of two squares, by the parameter-set condition of [@djokovic2018goethals], and 105 is not.",
      "entry_ids": [
        "n51-4eecf9266b906d840eaf",
        "n51-69f675fe8d1ff691c7e7",
        "n51-7a6b68c8e7b2b08bf8cc",
        "n51-f463b622a9ead754a546",
        "n55-5cec90854d11d04eba81",
        "n55-e3c675ce26d23cd5e525",
        "n55-f5bef16e102dfb62585b",
        "n59-99937ade3c4429fa7e85",
        "vxvware-kkss-n51",
        "vxvware-kkss-n55",
        "vxvware-kkss-n57-census",
        "vxvware-kkss-n59"
      ],
      "featured": true,
      "stratum": "kkss",
      "label": "Challenge result",
      "position": 1,
      "recorded_at": "2026-10-15"
    }
  ]
}

Sources

Submission JSON
{
  "schema_version": 1,
  "n": 59,
  "group": {
    "kind": "cyclic",
    "invariant_factors": [
      59
    ]
  },
  "submitted_by": "vxvware",
  "submitted_at": "2026-10-15",
  "discovered_by": [
    "vxvware"
  ],
  "sources": [
    {
      "title": "vxvware/saw-best-matrix: G-matrices (kkss) at n=59 (archived in this repository)",
      "url": "https://github.com/vxvware/goethals-seidel-challenge/blob/f29ee8912110f79163da774997b21f9289d4027f/evidence/participants/saw-best-matrix/gmat/gmat-59.txt",
      "commit": "aaa2c7952b52762b8811e3772ba8760b9eecdbac",
      "source_path": "data/gmat/gmat-59.txt",
      "artifact": {
        "path": "evidence/participants/saw-best-matrix/gmat/gmat-59.txt",
        "sha256": "a1f351b3220afb4cb741fcb9142f9e79c03ba8fd4f2dfcd5464bb01a9a9846b5"
      }
    }
  ],
  "equivalence": "equiv-v1",
  "id": "vxvware-kkss-n59",
  "type": "census",
  "stratum": "kkss",
  "representatives": [
    "n59-99937ade3c4429fa7e85"
  ],
  "claimed_count": 1,
  "claim_kind": "lower-bound",
  "secondary": {
    "stored_rows": 8
  },
  "notes": "The source reports a full census with 2 classes under its own equivalence convention and stores 8 quadruples. This board verifies the 1 equiv-v1 representative and records a lower bound; the search has not been replayed here. A single run, calibrated beforehand on the stored v=43 and v=45 censuses. The source counts two classes under its own convention; under equiv-v1 the eight rows fall into one class."
}