E Evidence Press

Press release · 8 September 2026 · version 0.1.0-candidate

Three-coloured paths in six-chromatic graphs: partial results

A five-colouring theorem for a mixed cover class, conditional finite exclusions and explicit repair obstructions; the parent problem remains unresolved.

Listen to this briefingNarrated summary · OpenAI API synthetic voice (fable) · MP3 · download

Summary

If a graph needs six colours on its vertices, must every three-colouring of its edges contain a one-colour path with three edges? This release does not settle that question. It supplies partial theorems, checked finite exclusions and explicit examples showing why several proposed shortcuts fail.

The strongest written result handles a mixed class built from two collections of small cliques and one collection of stars. Its five-colouring theorem works at every order, under those precise hypotheses.

Summary for specialists

TCP-MIXED-001: let $G=A\cup B\cup F$ be finite and simple, where every component of $A$ and $B$ is a clique of order at most three and $F$ is a star forest. Covering edges may overlap. Then $\chi(G)\le5$. Overlap is used in the local replacement proof; the theorem does not allow arbitrary star/triangle mixtures in all three covers.

For three edge-disjoint triangle factors, $\alpha(G)\ge n/3-1$ also implies five-colourability. Separately, a critical counterexample of order fourteen is excluded. The lower bound of fifteen vertices additionally assumes completeness of an external critical-graph catalogue through thirteen. At fifteen vertices, only component deficits zero and one are closed; edge counts 41, 42 and 43 remain.

Here the forbidden path is an ordinary subgraph on four distinct vertices, not necessarily induced. The parent AIM-COMBINATORICS-0050 remains unresolved.

Technical account

The mixed theorem turns the two clique covers into a bipartite incidence multigraph. A minimal obstruction would have degree-five leaves. A local four-clique replacement and the classical degree-choosability theorem force their incidence graph to be a forest. Its incidence count contradicts the number of star leaves.

The finite branches combine structural normalization, enumeration-coverage checks and saved vertex-colouring witnesses. A partial-Latin-square repair formulation gives a small-defect theorem, but retained counterexamples show that an arbitrary starting colouring need not admit the proposed repair.

Evidence, assurance and limitations

The package contains the full written notes, canonical claims, source dependencies, portable replay and stopped-search records. Sixteen selected checks and clean-extraction/hostile controls passed internally. Five producer-coordinated editorial roles accepted the bounded release with minor notes; this is not external specialist review or journal peer review.

Catalogue completeness was not regenerated. A timeout is not an exhaustion certificate. No formal verification, historical priority or comparative research acceleration is established. The supplied review's separate audit JSON was not attached, so its reported checks do not promote an external assurance field.

Relationship to earlier work

Garrison identifies the exceptional three-colour path case. Aharoni and collaborators study the triangle-factor question and the known four-colouring bound for three star forests. The release's possible originality concerns its particular mixed theorem, reductions and obstruction objects, not those antecedent formulations. Contribution-specific precedence remains open.

Who should care, and why

AudiencePotential useRequired caution
Graph-colouring researchersAudit a reusable mixed-cover theorem and finite restrictions.Preserve cover hypotheses and external dependencies.
Computational researchersReuse coverage checks and explicit repair obstructions.Internal replay is not independent reconstruction.
Research-method researchersInspect failed routes and measured scoped checks.No matched acceleration comparator exists.

Why the problem matters

Chromatic Ramsey questions ask how much vertex-colouring complexity forces patterns in edge-colourings. Useful partial structure can narrow the problem and prevent repeated false proof strategies without resolving the universal question. No practical impact is claimed here.

How to inspect or reproduce the recorded checks

Download the archive and read CLAIM_RECORDS.md before the historical notes. With Python 3.11 or later, run replay.py --bundle-root /path/to/bundle --output-dir /path/to/fresh-output. The output must be outside the bundle and not already exist. The wrapper checks hashes and recreates the historical layout in a copy; it does not edit the frozen evidence. Optimized Python is rejected. test_replay.py exercises relocation and selected hostile inputs.

The most valuable next projects

Independently scrutinize the mixed theorem's local lift and incidence count; reconstruct finite normalization coverage; assess contribution-specific prior art; and investigate an existential repair argument that survives the retained fixed-start counterexamples. These are future projects, not claims of closure.

What is in the evidence package

A consolidated PDF and accessible Markdown, all 21 original mathematical notes, canonical claim/dependency records, code, finite certificates, CNF checkpoints, replay receipts, internal editorial reports, licence map and file manifest. Externally owned papers, catalogue data and upstream binaries are omitted from the public successor with source links, hashes and reasons. The original private review archive remains unchanged. GitHub and Zenodo identify the exact public candidate version.

Media

The audio briefing is provided in the header above. Download the MP3 briefing · read the transcript.

Open directions for follow-up research

Also available in machine-readable form for research agents and follow-up projects.

  1. Resolve unrestricted mixtures of stars and triangles, including arbitrary-order triangle factors.
  2. Resolve the remaining critical order-fifteen cases with 41, 42 or 43 edges.
  3. Assess the mixed-cover proof and historical priority independently.
  4. Determine whether some starting colouring always admits bipartite repair in the edge-disjoint triangle-factor subclass.

Research process, metrics and reusable methods

Prospective process metadata under the Evidence Press operating model and research-metrics policy. It records the intended handoff, measured scope and claim boundary; it is not evidence that the method accelerated this work.

Work ID
ep-work:three-coloured-paths-partial-results
Attempt and metric receipts
  • ep-attempt:three-coloured-paths-partial-results-remaining-assurance-publication — published / positive

    Measurement scope
    assurance-through-publication — Remaining assurance and publication only. Discovery, supplied review, initial intake, portability wrapper and baseline replay predate registration and are excluded. No historical research clock reconstructed.
    Frozen target
    Bounded partial-results candidate with fail-closed replay, five-role internal review, public immutable archives and guarded canonical readback.
    Fermi active-time forecast
    150 minutes; plausible interval 100–230; expected unattended wait 25. Reference class: Evidence Press full candidate procedural prior (n=0) — Procedural prior, not measured comparative acceleration..
    • Consolidated package and source checks: 1 × 30/45/65 minutes (low/central/high) — Large existing dossier, bounded revisions.
    • Internal five-role review: 1 × 20/30/50 minutes (low/central/high) — One differentiated round.
    • Public archives and media: 1 × 30/45/65 minutes (low/central/high) — Existing publisher tools.
    • Gates and public readback: 1 × 20/30/50 minutes (low/central/high) — Two mandatory seal/deploy loops.
    Tractability forecast
    Within 300 active minutes: positive signal 0.95; target closure 0.85. Stop rule: Fail closed for scientific, rights and provider failures; preserve original claim ceiling. Workflow forecast is not a permission limit.
    Observed clocks
    42 active-agent; unknown active-human; unknown substantive-compute; 0 unattended-wait; 0 blocked; 8 rework minutes. Calendar elapsed: 42 minutes.
    Research search
    Cycles: 0 positive, 0 negative, 0 inconclusive. Falsification gates: 1. Candidate architectures: 0 tested, 0 rejected.
    Agent and review load
    5 agent runs; maximum parallelism 4; 72 model turns; unknown deduplicated model tokens; 1 substantive review rounds; P0/P1 findings 0/0; pre-publication claim corrections 0.
    Result and calibration
    target-closed — Bounded partial-results publication target closed; parent problem remains unresolved. One internal five-role round (domain/applications overlap disclosed), scoped replay, immutable matching archives and canonical first readback. Historical discovery excluded. Positive signal: true; target reached: true. Active-time error -108 minutes; actual/forecast 0.28; inside interval: false. Brier score: positive signal 0.0025; target closure 0.0225. Variance: Existing research dossier and reusable tools shortened the procedural prior. Active coordinator workflow interval recorded from the task-local clock, with short interleaved tool waits; no unattended interval claimed unless separately logged. Rework is an approximately eight-minute observed operational repair subtotal, not scientific revision. Model turns count available coordinator reasoning records only and are a lower bound, excluding unavailable earlier/fork records. Five agent runs count coordinator plus three review runs and one separate applications follow-up; no speed or impact inference.
    Missing telemetry
    activeHumanMinutes — Human effort not instrumented.; computeMinutes — No complete CPU/GPU meter across solver history, builds and hosted model calls; not inferred from wall time.; deduplicatedModelTokens — No authoritative fork-aware complete task-local total; inherited token events not summed.; uncachedInputTokens — No authoritative task-local uncached-input total.
    Measurement corrections
    • measurement.agentRuns -> metrics.outcome.agentRuns — Opening 1 retained; terminal 5 counts coordinator and four reviewer executions. Reason: Opening measurement remains immutable, not a final total.
    • measurement.reworkMinutes -> metrics.outcome.reworkMinutes — Retain opening reworkMinutes=0; terminal approximately eight-minute recorded operational repair subtotal is disclosed as approximate. Reason: Opening snapshot is not a final total.
Prospective work ledger · metrics policy
Intended aims
science
Artifact roles
research-output, evidence-assessment
Decision object
bound — Mixed-cover colouring theorem, conditional finite exclusions and repair obstructions. Scope: Exact hypotheses in TCP-MIXED-001 and the other canonical claim records; not the full parent theorem.
Reusable methods
Structural compression (structural-compression); Certificate-first, proof-carrying research (certificate-first); Adversarial scientific controls (adversarial-controls); Productive failure and stop receipts (productive-failure); Assurance as a vector (assurance-vector); Agent-readable research objects (agent-readable-research-object) · registry
Targeted clocks
assurance, publication
Semantic bridge
explicit — Ordinary three-edge path avoidance gives star/triangle components; proofs, coverage and witness replay have distinct dependencies. Remaining risks: Written all-order proof awaits external specialist scrutiny.; External catalogue completeness is not regenerated.; Finite checks and timeouts do not close arbitrary orders..
Human judgement gates
  • Assess the mixed theorem and local lifting argument.
  • Audit external theorem hypotheses and catalogue scope.
  • Determine contribution-specific priority.
  • Preserve creator, rights and candidate status.
Next assurance action
Obtain unaffiliated scrutiny of the mixed proof and independent finite-coverage reconstruction. Claim ceiling: Unrefereed partial-results candidate. AIM-COMBINATORICS-0050 remains unresolved. The minimum-order bound depends on external catalogue completeness through thirteen; only selected order-fifteen cases are closed. No external specialist validation, formal verification or priority clearance is established.
Aim-scoped impact evidence
  • science: NO_IMPACT_EVIDENCE — Inspectable bounded mathematical partial results in Producer-coordinated research publication. Design: none; comparator: No matched comparator.; estimand: No speed or impact effect estimated.. No real-world effect evidence is asserted.
Parent handoffs
  • depends-on-claim https://doi.org/10.37236/5687 — inherited claim: Degree-choosability characterization, Theorem 13.; inherited ceiling: Imported standard theorem; does not independently validate the new mixed-cover proof.
  • depends-on-claim https://users.cecs.anu.edu.au/~bdm/data/graphs.html — inherited claim: Stated edge-critical catalogue coverage through thirteen.; inherited ceiling: Supplied-record checking does not regenerate catalogue completeness.

Verification status

Unrefereed partial-results candidate. AIM-COMBINATORICS-0050 remains unresolved. The minimum-order bound depends on external catalogue completeness through thirteen; only selected order-fifteen cases are closed. No external specialist validation, formal verification or priority clearance is established.

Cite

Anonymous (2026). Three-coloured paths in six-chromatic graphs: partial results. Version 0.1.0-candidate. Zenodo. https://doi.org/10.5281/zenodo.22655876
BibTeX
@misc{threecolouredpathspartialresults2026,
  title        = {Three-coloured paths in six-chromatic graphs: partial results},
  author       = {Anonymous},
  year         = {2026},
  doi          = {10.5281/zenodo.22655876},
  url          = {https://doi.org/10.5281/zenodo.22655876},
  version      = {0.1.0-candidate},
  howpublished = {Zenodo},
  note         = {Unrefereed; internally replayed evidence package. Press page: https://evidencepress.org/releases/three-coloured-paths-partial-results/}
}

Also: cite.bib · paper.json · this page as Markdown