{"id":"bc2d680a-b9d0-41f2-a606-0ff57515cdc6","arxiv_id":"2608.12013","paper_version":2,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":8.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"A new family of layer barriers for colour-biased tight Hamilton cycles gives a counterexample to the recent conjecture of Behague, Clemen, Hyde and Morrison on minimum vertex degree thresholds.","lead":"This paper constructs a family of red-blue coloured uniform hypergraphs in which every tight Hamilton cycle is perfectly colour-balanced. It shows that an interior member of this family exceeds the conjectured minimum vertex-degree threshold, disproving a recent conjecture.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant internal objection identified; the only caveat is that the refutation depends on the exact wording of Conjecture 6.1 of [3].","rationale":"The reader's weakest assumption is the same external dependency I reach: the exact statement of Conjecture 6.1 from the cited paper. My independent check of the construction, the colour-balance proof, the degree estimates, and the numerical inequality found no mathematical error. The proofs are short and the crucial arithmetic is reproducible. Therefore no correction to the ACCEPT verdict is warranted. If the original conjecture is exactly as quoted, the counterexample is decisive; if not, the paper's headline claim would need to be re-evaluated, but that is not an internal flaw detectable from the manuscript. Agreement: agree, because the reader identified the same point as the weakest assumption.","tokens_in":6311,"tokens_out":26096,"duration_ms":241255,"concrete_test":"Retrieve arXiv:2607.29628 and compare Conjecture 6.1 verbatim with the statement of Conjecture 1.1 in Section 1, checking in particular whether the degree bound is binom(n,k-1) or binom(n-1,k-1) and whether d_k is given by equation (1). If the original statement differs, recompute the k=17, a=8 counterexample against the original threshold; if it matches, the refutation stands exactly as written.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The mathematics of the paper checks out. Lemma 2.1 is correct: edge types of consecutive windows change by at most one, so a tight cycle cannot cross the forbidden types a-1 and a+2. Proposition 2.2 correctly forces all window B-counts to lie in {a,a+1}, and the incidence count k|B|/n=a+1/2 then forces exactly n/2 edges of each colour. Proposition 2.3's binary-word construction is valid and gives a genuine tight Hamilton cycle. The degree formulae in Proposition 3.1 are the correct inclusion-exclusion counts, and the falling-factorial asymptotics reproduce the two binomial probabilities defining d_{k,a}. For k=17,a=8 the arithmetic is correct: d_{17,8}=1-(C(16,7)+C(16,10))/2^16=5761/8192, and d_{17}=1-(8/17)(33/34)^15; the margin exceeds 0.003, so alpha=0.003 works once n is large. I therefore find no internal gap in the self-contained argument. The one load-bearing external assumption is that Conjecture 1.1 exactly reproduces Conjecture 6.1 of Behague et al. [3]; any difference in the degree normalization, the formula for d_k, or additional hypotheses would require the counterexample computation to be re-run. This is a reference-dependency caveat, not a correctness problem in the paper's own derivation.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper introduces a one-parameter family of red–blue coloured k-uniform hypergraphs H_{k,a}(n), called layer barriers, obtained by deleting edges whose intersection with a small part B has size a−1 or a+2 and colouring the surviving layers a and a+1 red and blue. The authors prove that every tight Hamilton cycle in H_{k,a}(n) is perfectly colour-balanced (Proposition 2.2), that the construction nevertheless contains a tight Hamilton cycle (Proposition 2.3), and that its asymptotic relative minimum vertex degree equals d_{k,a} (Proposition 3.1). At k=17 and a=8, this degree exceeds the conjectured threshold d_17 of Behague, Clemen, Hyde and Morrison, giving a counterexample to Conjecture 6.1 of that paper (Corollary 1.3). The authors also show that interior choices of a make the best barrier have asymptotic relative minimum degree 1−O(k^{-1/2}), in contrast to d_k→1−e^{-1/2}/2 (Corollary 1.4).","tokens_in":6602,"tokens_out":6896,"duration_ms":67799,"significance":"The contribution is a clean, explicitly checkable construction with a concrete numerical counterexample to a published conjecture. The proofs are short and self-contained: Lemma 2.1 gives a sharp window-confinement argument, Proposition 2.2 uses a double-counting incidence identity to force exact colour balance, and Proposition 3.1 computes the degree with standard binomial asymptotics. The counterexample is falsifiable and the constants are explicit. The main caveat is that the refutation depends on the exact wording of Conjecture 6.1 in the cited preprint [3]; the internal mathematics is sound. The paper also gives a new large-uniformity phenomenon, showing that interior layer barriers achieve asymptotic relative minimum degree 1−O(k^{-1/2}), which is substantially denser than the boundary construction.","major_comments":[{"comment":"The counterexample in Corollary 1.3 refutes Conjecture 1.1 as stated, but the paper's central claim depends on this quotation being an exact reproduction of Conjecture 6.1 in [3]. The authors should verify the original statement and either reproduce it verbatim in the introduction or add a remark confirming that the degree normalization (minimum vertex degree as a count of (k−1)-sets), the formula (1) for d_k, and the quantifier order (α before δ and n0) match [3]. If the original conjecture carries any additional hypothesis, the computation in §3.1 would need to be checked against that version. This is a verification requirement rather than an error in the paper's internal derivation.","section":"Section 1, Conjecture 1.1"}],"minor_comments":[{"comment":"The phrase 'colour all remaining edges arbitrarily' is slightly imprecise because the proof of Proposition 2.2 shows that no tight Hamilton cycle uses any edge of these types; a one-sentence remark to that effect would help the reader.","section":"Section 2, equation (3)"},{"comment":"The inequality in (9) is asserted to follow by clearing denominators; displaying the exact integer inequality or stating the verified rational value would make the margin check fully transparent.","section":"Section 3.1, after (9)"},{"comment":"Reference [3] is an arXiv preprint; the authors should ensure they cite the latest version, since the exact statement of Conjecture 6.1 could change between versions.","section":"References"}],"recommendation":"minor_revision","confidential_remarks":"The paper is mathematically sound and the counterexample is convincing provided the quotation of [3, Conjecture 6.1] is accurate. Please ask the authors to confirm the verbatim statement of the conjecture and to add a brief remark on the degree normalization. Once that is done, the paper is suitable for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Solid, short paper. It constructs a one-parameter family of red-blue k-uniform hypergraphs that are layer barriers: every tight Hamilton cycle is perfectly colour-balanced, yet the construction contains at least one tight Hamilton cycle. The boundary case a=0 reproduces the Behague-Clemen-Hyde-Morrison obstruction; the new interior layers are genuinely denser, and at k=17, a=8 the relative minimum vertex degree is ~0.7032473, above the conjectured d_17 ~0.6992772. That refutes the quoted Conjecture 6.1.\n\nThe mathematics is clean and checkable. Lemma 2.1 uses the one-vertex slide between consecutive windows to show a cycle cannot cross the forbidden types. Proposition 2.2 double-counts vertex-edge incidences to force all window types to be a or a+1, then forces exactly n/2 red edges. Proposition 2.3 gives an explicit binary word and verifies the construction actually is Hamiltonian. The degree formulas in Proposition 3.1 are standard inclusion-exclusion, the falling-factorial asymptotics are fine, and the k=17 arithmetic is correct: d_{17,8}=5761/8192, d_17=1-(8/17)(33/34)^15, and the gap exceeds 0.003. The large-k bound 1 - max_a d_{k,a} = O(k^{-1/2}) follows from Stirling, and the choice of a depending on parity is valid.\n\nThe one soft spot is external: the disproof is only as strong as the quoted statement of Conjecture 6.1 in [3]. If the original conjecture has extra hypotheses or a different normalization for d_k, the specific numbers would need rechecking. That is a reference-dependency caveat, not a mathematical gap in this paper. I also note the paper does not treat k=3, but the conjecture is only for k>=4, so this is not a problem.\n\nThis is a worthwhile counterexample, clearly written, with no self-citations and no parameter fitting. It deserves a serious referee; I would support acceptance after the reference check. I'd cite it for the layer-barrier family and the O(k^{-1/2}) behaviour.","headline":"A clean, checkable counterexample to a conjectured vertex-degree threshold; the only caveat is that the refutation hinges on the exact wording of the quoted conjecture.","tokens_in":7135,"tokens_out":2672,"would_cite":true,"duration_ms":24334,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["05C65","05C38","05C15"],"pacs":[],"model":"deepseek-v4-flash","headline":"A one-parameter family of red-blue $k$-graphs forces every tight Hamilton cycle to be perfectly colour-balanced, and at $k=17$ the minimum vertex degree exceeds the conjectured threshold $d_{17}$.","keywords":["uniform hypergraphs","tight Hamilton cycles","colour discrepancy","minimum vertex degree","layer barriers","red-blue edge-colouring","colour-balanced","counterexample"],"falsifier":"Compare the statement of Conjecture 6.1 in the cited reference [3] with the version quoted here as Conjecture 1.1; if the original carries any extra hypothesis (a different degree notion, a restriction on $n$, or an additional structural condition), then the $k=17$, $a=8$ construction may not be a counterexample. If the statements match exactly, the computation $d_{17,8}-d_{17} = \\frac{8}{17}(\\frac{33}{34})^{15} - \\frac{2431}{8192} > \\frac{39}{10000}$ settles it.","tokens_in":6119,"feed_emoji":"⚖️","tokens_out":15198,"duration_ms":134901,"temperature":0.7,"pith_summary":"The paper constructs, for every uniformity $k\\ge 3$ and every parameter $a\\in\\{0,\\dots,k-1\\}$, arbitrarily large red-blue $k$-uniform hypergraphs that contain a tight Hamilton cycle yet have every tight Hamilton cycle perfectly colour-balanced: exactly $n/2$ red and $n/2$ blue edges. The construction is a layer barrier: the vertex set is split into two parts, and edges whose intersection with the smaller part has one of two forbidden sizes are deleted. The asymptotic minimum vertex degree of the $a$-th barrier is $d_{k,a}$ as in (2), and the boundary case $a=0$ reproduces the previously known extremal example with degree $d_k$. The paper's main point is that interior layers are strictly denser: at $k=17$, $a=8$ gives $d_{17,8}=5761/8192\\approx0.703247$, exceeding the conjectured $d_{17}\\approx0.699277$, so Conjecture 1.1, quoted from Conjecture 6.1 of the cited reference [3], is false. For large $k$, choosing $a\\approx k/2$ makes the barrier density $1-O(k^{-1/2})$, far above the limit of $d_k$.","feed_headline":"Colour-balanced barriers beat the conjectured threshold at k=17","feed_subtitle":"At k=17 the new barrier exceeds the conjectured density threshold, and every tight Hamilton cycle stays colour-balanced.","key_machinery":"The layer barrier: fix two forbidden intersection sizes, $a-1$ and $a+2$, where the type of a $k$-set is its intersection size with $B$. Consecutive edges of a tight cycle are consecutive windows of length $k$ in a cyclic vertex order, so their types differ by at most 1; the forbidden layers therefore confine all types of any tight cycle to one of the three intervals $\\{0,\\dots,a-2\\}$, $\\{a,a+1\\}$, $\\{a+3,\\dots,k\\}$. Since the average type over all $n$ edges equals $k|B|/n=a+1/2$, the cycle cannot live in the lower or upper interval, so every edge has type $a$ or $a+1$; counting edges of type $a+1$ against the average gives exactly $n/2$ blue edges and $n/2$ red edges. Hamiltonicity is supplied by a $2k$-periodic binary sequence with $2a+1$ ones per period.","core_discovery":"Fix $k\\ge3$, $a\\in\\{0,\\ldots,k-1\\}$, and let $n$ be divisible by $2k$. Partition the vertex set into $A\\cup B$ with $|B|=(2a+1)n/(2k)$, and put a $k$-set $e$ into the hypergraph exactly when its type $|e\\cap B|$ is neither $a-1$ nor $a+2$; colour type-$a$ edges red and type-$a+1$ edges blue. The paper proves that every tight Hamilton cycle in this $k$-graph has all edge types in $\\{a,a+1\\}$, which by an averaging count forces exactly $n/2$ edges of each type, hence colour sum zero, while a $2k$-periodic binary word with $2a+1$ ones per period exhibits a genuine tight Hamilton cycle. The asymptotic relative minimum vertex degree equals $d_{k,a}=1-\\max\\{P(X_{k,a}\\in\\{a-1,a+2\\}),P(X_{k,a}\\in\\{a-2,a+1\\})\\}$ with $X_{k,a}\\sim\\operatorname{Bin}(k-1,(2a+1)/(2k))$. For $k=17$, $a=8$, binomial symmetry gives $d_{17,8}=5761/8192$, and the paper verifies $d_{17,8}-d_{17}>39/10000$, so with $\\alpha=3/1000$ arbitrarily large counterexamples to Conjecture 1.1 exist. As $k\\to\\infty$ the interior choice $a\\approx k/2$ gives $1-d_{k,a}=O(k^{-1/2})$.","pith_inferences":["The two-forbidden-layer mechanism might carry over to $r$-colourings or to other spanning structures such as perfect matchings, where deleting further layers could yield even denser colour-balanced barriers; this is a testable extension, not a claim of the paper.","The large-uniformity gap suggests that for red-blue $k$-graphs the true threshold for colour-biased tight Hamilton cycles is asymptotically much closer to 1 than to the uncoloured Hamiltonicity threshold, a contrast that could be probed in random hypergraphs.","The binomial symmetry at $k=17$, $a=8$ makes $d_{17,8}=1-2^{-16}(\\binom{16}{7}+\\binom{16}{10})$ a closed rational; searching other uniformities for a similar symmetry might produce an even larger gap over $d_k$."],"forward_implications":["Conjecture 1.1, quoted from Conjecture 6.1 of the cited reference [3], is false: for $k=17$ any threshold forcing a colour-biased tight Hamilton cycle must lie at least at $5761/8192\\approx0.703247$, above the conjectured $d_{17}\\approx0.699277$.","For every $k\\ge3$ and every $a$, the layer barrier $H_{k,a}(n)$ shows that relative minimum vertex degree $d_{k,a}$ is compatible with perfect colour balance, so a true sufficient threshold must exceed every $d_{k,a}$ for the corresponding $k$.","For large $k$, any sufficient threshold must be at least $1-O(k^{-1/2})$, approaching 1, whereas the boundary threshold $d_k$ tends only to $1-e^{-1/2}/2$.","The construction contains a tight Hamilton cycle, so the obstruction is genuinely about colour discrepancy rather than about the failure of Hamiltonicity."],"supporting_citations":[{"why":"Supplies the conjectured threshold $d_k$ and Conjecture 6.1 (quoted as Conjecture 1.1) that the paper refutes by the $k=17$, $a=8$ barrier.","marker":"[3]"},{"why":"Proposes the perfect-matching colour-bias conjecture for $k=4$ that the threshold conjecture generalises, anchoring the boundary case $a=0$ of the family.","marker":"[11]"},{"why":"Establishes the uncoloured minimum-vertex-degree threshold for tight Hamilton cycles in 3-graphs, the baseline degree notion against which the colour-biased threshold is measured.","marker":"[13]"}],"fun_headline_variants":["Colour-balanced barrier beats threshold at k=17","Denser barriers force colour-balanced Hamilton cycles","Threshold counterexample via colour-biased barriers","Interior barriers approach full density as k grows","Colour-balanced Hamilton cycles with densest construction"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The counterexample refutes Conjecture 1.1 only if that conjecture is exactly as quoted from the cited reference [3]: same definition of minimum vertex degree $\\delta_1$, same constant $d_{17}$, and no hidden condition excluding the case $n$ divisible by 34.","fun_headline_variants_meta":{"raw":{"variants":["Colour-balanced barrier beats threshold at k=17","Denser barriers force colour-balanced Hamilton cycles","Threshold counterexample via colour-biased barriers","Interior barriers approach full density as k grows","Colour-balanced Hamilton cycles with densest construction"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00021,"raw_usage":{"total_tokens":1518,"prompt_tokens":1162,"completion_tokens":356,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":778,"completion_tokens_details":{"reasoning_tokens":286}},"tokens_in":778,"tokens_out":356,"duration_ms":3846,"temperature":1.0,"reasoning_tokens":286,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T00:19:09.277498+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compare the statement of Conjecture 6.1 in the cited reference [3] with the version quoted here as Conjecture 1.1; if the original carries any extra hypothesis (a different degree notion, a restriction on $n$, or an additional structural condition), then the $k=17$, $a=8$ construction may not be a counterexample. If the statements match exactly, the computation $d_{17,8}-d_{17} = \\frac{8}{17}(\\frac{33}{34})^{15} - \\frac{2431}{8192} > \\frac{39}{10000}$ settles it.","supporting_citations":[{"cited_title":"A minimum-degree threshold for colour-biased Hamilton cycles in hypergraphs","cited_arxiv_id":"2607.29628","evidence_quote":"Supplies the conjectured threshold $d_k$ and Conjecture 6.1 (quoted as Conjecture 1.1) that the paper refutes by the $k=17$, $a=8$ barrier."},{"cited_title":"H` an, R","cited_arxiv_id":null,"evidence_quote":"Proposes the perfect-matching colour-bias conjecture for $k=4$ that the threshold conjecture generalises, anchoring the boundary case $a=0$ of the family."},{"cited_title":"Reiher, V","cited_arxiv_id":null,"evidence_quote":"Establishes the uncoloured minimum-vertex-degree threshold for tight Hamilton cycles in 3-graphs, the baseline degree notion against which the colour-biased threshold is measured."}],"review_version":1}