The Proper Hat-Guessing Number of K₇−e

Matthew Protti

The proper hat-guessing number of K7−e is 12. Either two disjoint S(5,6,12) completion designs or a sharply five-transitive orbit-map construction meets the twin-completion criterion; the general upper bound gives equality.

Sources for this record

What was already known

The general K_n−e upper bound was known. The K5−e and K6−e constructions do not determine K7−e.

AI involvement AI-centered

GPT-5.6 Pro contributed the design and orbit-map constructions and proof within Matthew Protti’s directed project. Protti evaluated and released the result.

GPT-5.6 Pro contributed the retained designs/orbit-map research under Protti’s direction; no individual autonomous origin is shown.

Importance Modest

Determines another open finite case by two distinct constructions.

Limitations

Neither construction establishes the result for all K_n−e. The adversarial reviewer’s human identity is undocumented.

Technical verification Formal (exact computation)

The release and adversarial report document exact regeneration of both 132-block designs, 95,040 group elements, 59,400 ordered lines and all 4,656,960 proper colourings; the reviewer accepted the theorem. The reviewer is unnamed, so only machine checking is annotated.

Formal checking

Exact block/group reconstruction, line tests and Hall-degree census

Checker
Python and independent C++20 constructors, data verifier, adversarial replay
Scope
Both K7−e lower-bound constructions and complete 4,656,960-colouring coverage
Basis
Reported by the linked checker
Evidence source

Checking a formal statement does not establish that it matches the original problem.

Human understanding Few

No identified human gives a separate explanation of the two AI-assisted constructions; the anonymous review is not enough to establish a named person’s understanding. Few is the evidence default.

Correction and publication history