{"id":"4a80eb83-ed27-49bd-a300-29433542b3d3","arxiv_id":"2505.13849","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The Saxl hypergraph, whose hyperedges are bases of minimum size, is complete exactly for Frobenius groups, alternating and symmetric natural actions, certain projective line actions, Suzuki groups, and five exceptional actions.","lead":"This paper introduces the Saxl hypergraph, whose edges are the minimum-size bases of a permutation group, and classifies when it is complete. It also proves partial results on a common-neighbour conjecture and on when such hypergraphs admit a flag-spanning tour.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1.3 rests on five unverified computational exclusion claims for affine exceptional groups; if any one is actually K(3), the classification's 'only if' direction fails.","rationale":"The reader's CONDITIONAL verdict already identifies the unshown computational checks as a source of risk. My stress-test agrees that this is the weakest point, and I narrow it to one concrete, load-bearing step: the five affine exceptional groups in the proof of Theorem 1.3. The paper has real independent support: Theorem 1.3's proof is mostly a clean reduction via Corollary 3.2 to the standard classification of 2-transitive groups, and the Section 4 results are honestly stated as conditional on Conjecture 4.1. I found no internal contradiction in the non-computational arguments, and the duplicated paragraph in Lemma 4.8 is an editorial flaw rather than a mathematical one. The concern is not that the computations are likely wrong; it is that the central 'only if' direction of the main theorem depends on an unreported finite verification. Because this is easily settled and does not undermine the paper's structure, the correct verdict remains CONDITIONAL, unchanged from the reader's assessment.","tokens_in":26200,"tokens_out":26122,"duration_ms":250239,"concrete_test":"Run a GAP or Magma script for each of the five exceptional affine groups G0 in the list from Theorem 1.3. Construct the natural F2-module V, and up to G0-conjugacy enumerate all pairs (v,w) of distinct nonzero vectors; compute the pointwise stabiliser (G0)_{v,w}. If any non-identity element fixes both v and w, the group is not K(3); if every such stabiliser is trivial, then every triple {0,v,w} is a base, and since one- and two-point stabilisers are nontrivial, b(V:G0)=3, so the group is K(3). Emit, for each group, the maximum order of (G0)_{v,w} and a command log. A negative result for all five groups verifies the omitted computation; a single nontrivial pair would refute Theorem 1.3.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central classification, Theorem 1.3, is proved in Section 3 by reducing to the Dixon–Mortimer list of 2-transitive groups. Most of the case eliminations are genuine derivations, but the affine q=2 case ends with the assertion: 'Otherwise, (d, G0) is one of (4, S6), (4, A6:C2), (4, A7), (6, PΣU3(3)≅G2(2)), and (6, PSU3(3)). Computations confirm none of these correspond to K(3) groups.' No script, certificate, or reproducible computation is supplied. This is load-bearing: K(3) means b(G)=3 and every 3-point set is a base. For an affine group G=V:G0 over F2, this is equivalent to the condition that every pair of distinct nonzero vectors v,w has trivial pointwise stabiliser in G0. If any of the five listed groups satisfies that condition, then Theorem 1.3's 'only if' direction is false. The surrounding reliance on the standard classification of 2-transitive groups is not the soft spot. The conditional results in Section 4 are explicitly stated as dependent on Conjecture 4.1, so they do not affect the central claim. The duplicated paragraph in Lemma 4.8 is an editing error, not a mathematical gap. The unshown finite check is therefore the single most load-bearing concern.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper introduces the Saxl hypergraph H(G) of a permutation group G, whose edges are the bases of minimum size b(G), and studies three themes. Theorem 1.3 classifies the groups whose Saxl hypergraph is complete, extending the Saxl-graph classification for b(G)=2. Section 4 analyses common-neighbour properties: Conjecture 4.3 (a hypergraph strengthening of the Common Neighbour Conjecture) is proved for several large classes of primitive groups assuming the generalised CNC (Theorem 4.5), and gossip numbers are studied (Theorems 4.10 and 4.12). Section 5 treats valency and flag-spanning tours, giving a partial classification for primitive groups with b(G) in {3,4} (Theorems 1.6, 5.2, and 5.4). The paper is clearly organised and makes appropriate use of standard classifications of 2-transitive and primitive groups.","tokens_in":26435,"tokens_out":7402,"duration_ms":63619,"significance":"If correct, Theorem 1.3 provides a complete and elegant classification of permutation groups with complete Saxl hypergraphs, a natural counterpart to the known Saxl-graph results. The conditional results in Section 4 give substantial evidence for a plausible hypergraph analogue of the CNC, and the flag-spanning tour results open a new research direction. The paper is careful in stating which results depend on the unproved generalised CNC. However, several finite computational assertions that are load-bearing for the classifications are made without supplying scripts, certificates, or reproducible details; these should be provided or replaced by proofs.","major_comments":[{"comment":"The exclusion of the five exceptional affine groups (d,G0) = (4,S6), (4,A6:C2), (4,A7), (6,PΣU3(3)≅G2(2)), and (6,PSU3(3)) rests entirely on the assertion 'Computations confirm none of these correspond to K(3) groups' with no code, certificate, or reproducible detail supplied. This check is load-bearing for the 'only if' direction of Theorem 1.3. Since K(3) is equivalent to the condition that every pair of distinct nonzero vectors in V_d(2) has trivial pointwise stabiliser in G0, this is a routine finite verification; please provide a script (e.g., in GAP or Magma) or a short mathematical argument.","section":"§3, proof of Theorem 1.3 (affine q=2 case)"},{"comment":"The assertion 'Computations show PSL3(3).2 gives rise to valency 9750' is a finite check that determines whether this group is included in Table 4, but no computation is supplied. Similar unshown computational checks appear elsewhere, for instance in the proof of Theorem 4.5 where it is stated that one can 'confirm computationally that there are 7 regular orbits on Ω^7' for M24. These computations affect the stated classifications and should be made reproducible or replaced by explicit arguments.","section":"§5, Lemma 5.4 and proof of Theorem 1.6"},{"comment":"The proof of Lemma 4.8 contains a full duplicated paragraph, appearing almost verbatim twice ('It only remains to consider the possibility that span(B′) ≠ span(B) ... An application of Lemma 4.7 completes the proof.'), and the two copies are inconsistent in the displayed inequality (the first reads |Ev| + 1 and the second |Ev−w| + 1). This is a cut-and-paste error that obscures the argument; please remove the duplicate and reconcile the notation so that the intended proof is unambiguous.","section":"§4, Lemma 4.8"}],"minor_comments":[{"comment":"The notation 'PSL2(q2).C2̸≤ PΣL2(q2)' is garbled and does not name a group; it should be written as, for example, 'G = PSL2(q^2).C2 with the C2 not contained in PΣL2(q^2)', or the intended group should be described in words as in the proof.","section":"Theorem 1.3(iii)"},{"comment":"The section heading 'V alency' contains a stray space and should be 'Valency'.","section":"§5 heading"},{"comment":"The phrase 'as occurs for any edge when G is a symmetric group' would be clearer as 'as occurs, for example, when G is a symmetric group in its natural action and the edge is a base containing a pair of points interchanged by a transposition in G'.","section":"§2.3, Remark 2.4"},{"comment":"The sentence 'At most one orbit of such G can have the property that it contains fixed points of more than half the elements of G' would benefit from a short justification or a reference, since it is used to deduce transitivity of K(2) groups.","section":"§3, proof of Lemma 3.1"}],"recommendation":"major_revision","confidential_remarks":"The central concern is reproducibility: Theorem 1.3, Theorem 1.6, and parts of Theorem 4.5 rely on finite computations that are asserted but not supplied. These are likely correct and easily checkable, but the manuscript currently does not meet the standard of verifiability expected for classification results. I recommend asking the authors to provide scripts or certificates as part of the revision. The duplicated paragraph in Lemma 4.8 should also be fixed. Apart from these points, the mathematics appears sound and the paper is a valuable contribution."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The Saxl hypergraph is a natural and genuinely new object, and the completeness classification in Theorem 1.3 is the centerpiece. It cleanly extends the base-2 case and the case analysis is mostly convincing. The paper is also honest about its conditional results: Section 4 explicitly states Conjecture 4.1 and proves theorems relative to it, which is the right way to do business. The valency and flag-spanning tour material in Section 5 is more technical, but it fits the program and contains real information.\n\nThe soft spots are real but narrow. The proof of Theorem 1.3, in the affine q=2 case, terminates with the sentence “Computations confirm none of these correspond to K(3) groups” for five specific groups: (4, S6), (4, A6:C2), (4, A7), (6, PΣU3(3)≅G2(2)), (6, PSU3(3)). This is load-bearing. If any of those groups actually is K(3), the “only if” direction fails. No script, no certificate, no explicit counts are supplied. I believe the computation is probably right, but the text gives a referee no way to verify it, and it is exactly the kind of step that has been wrong in the literature before. I would not desk-reject over this, but I would require the authors to provide the computation, a proof, or at least a precise reference to computer output. The duplicated paragraph in Lemma 4.8 is just an editing error and should be fixed, but it does not suggest a mathematical gap. The GAP experiments mentioned in Section 4 are similarly undocumented, but they are not load-bearing for the main theorem.\n\nThe citation pattern looks fine, and the authors are careful to distinguish their contribution from the prior generalized Saxl graph of Freedman et al. This paper deserves a serious referee. It is not groundbreaking, but it adds a useful tool to the permutation-group base-size program, and the classification in Theorem 1.3 will be cited. The reader's conditional verdict is right. If I were the editor, I would send it out, with an explicit request that the computational claims be supported.","headline":"Solid new invariant and classification, but the main theorem leans on an unshown finite computation that a referee should push on.","tokens_in":26964,"tokens_out":2743,"would_cite":true,"duration_ms":25972,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["20B15","20B05","05C65","05C25"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper classifies the finite permutation groups whose Saxl hypergraph is complete: exactly the Frobenius groups, the natural alternating and symmetric groups, certain projective-line groups, the Suzuki groups in their doubly…","keywords":["Saxl hypergraph","base size","permutation groups","complete hypergraph classification","Common Neighbour Conjecture","flag-spanning tours","primitive groups","2-transitive groups"],"falsifier":"A direct computation in the four sporadic actions of Theorem 1.3(v) would settle the asserted computational confirmation: list all bases of minimal size and verify that every $b(G)$-element subset has trivial pointwise stabiliser. For the common-neighbour part, test Conjecture 4.3 on the affine groups with base size at most seven; a pair of vertices whose minimal-base edges always meet in at least two points would be a counterexample, and under Theorem 4.5 would disprove the generalised Common Neighbour Conjecture for that group.","tokens_in":25983,"feed_emoji":"🕸️","tokens_out":11272,"duration_ms":95416,"temperature":0.7,"pith_summary":"The paper introduces the Saxl hypergraph of a permutation group: the hypergraph whose vertices are the points being permuted and whose edges are exactly the bases of minimal size. Its central result is a classification of the groups for which this hypergraph is complete, meaning every set of $b(G)$ points already forms a base, and the list is short: Frobenius groups, alternating and symmetric groups in their natural actions, certain projective-line groups, Suzuki groups in their doubly transitive action, and four sporadic exceptional actions. The paper also proves that, conditional on the generalised Common Neighbour Conjecture, every primitive group with base size at most seven has the property that any two points lie in two minimal bases meeting in exactly one point, and it gives a partial classification of base-size-three and base-size-four primitive groups whose Saxl hypergraph admits a flag-spanning tour. These results matter because they delimit, in concrete group-theoretic terms, how far the strongest possible base structure can extend beyond the classical base-size-two case.","feed_headline":"Saxl hypergraph completeness classified: five families only","feed_subtitle":"Frobenius, alternating, projective-line, Suzuki, and four sporadic cases are the only groups with a complete Saxl hypergraph.","key_machinery":"The Saxl hypergraph $\\mathcal{H}(G)$ is the hypergraph on $\\Omega$ whose edges are the bases of $G$ of size $b(G)$. The recursion driving the completeness classification is Lemma 3.1: $G$ is $K(n)$, meaning every $n$ points form a base, if and only if each point stabiliser $G_\\alpha$ acting on $\\Omega\\setminus\\{\\alpha\\}$ is $K(n-1)$; this yields Corollary 3.2 that every $K(n)$ group is $(n-1)$-transitive, which funnels the problem into the known list of finite 2-transitive groups. For the flag-spanning tour theorem, the operative mechanism is the criterion that a hypergraph has a flag-spanning tour exactly when it has an even number of vertices and even valency, reducing the proof to parity calculations for the valency of $\\mathcal{H}(G)$.","core_discovery":"The central claim is Theorem 1.3: a group $G \\le \\mathrm{Sym}(\\Omega)$ with $b(G) \\ge 2$ has a complete Saxl hypergraph if and only if $G$ is Frobenius; an alternating or symmetric group in its natural action; a projective-line group among $\\mathrm{PSL}_2(q)$, $\\mathrm{PGL}_2(q)$, or $\\mathrm{PSL}_2(q^2).C_2 \\not\\le \\mathrm{P}\\Sigma\\mathrm{L}_2(q^2)$; a Suzuki group $^2B_2(q)$ in its doubly transitive action; or one of $(\\mathrm{PSL}_2(11),12)$, $(\\mathrm{PGL}_2(11),12)$, $(M_{11},11)$, $(M_{12},12)$. The proof rests on Lemma 3.1, which characterises the property recursively: a group has every $n$ points forming a base exactly when each point stabiliser does the same for $n-1$, so any such group with base size at least three must act $(n-1)$-transitively and the classification reduces to running through the finite 2-transitive groups. Beyond the classification, Theorem 1.5 shows that for primitive groups with base size at least three the common-neighbour number $g_2$ is never exactly one, and Theorem 1.6 gives a partial classification of primitive groups of base size three or four whose Saxl hypergraph admits a flag-spanning tour.","pith_inferences":["A natural test of Theorem 1.3 is to compute, for the four sporadic actions, all bases of minimal size and verify directly that every $b(G)$-set has trivial stabiliser; these computations are asserted but not exhibited, and an independent verification would make the classification self-contained.","The $g_3 = 0$ affine examples in Theorem 4.12 suggest that the Common Neighbour Conjecture cannot be strengthened to triples of vertices, so the hypergraph viewpoint exposes the exact threshold: pairs are special, and any stronger common-neighbour hypothesis would need a different condition.","Because completeness of $\\mathcal{H}(G)$ implies sharp $(n-1)$-transitivity, the classification could be re-derived by enumerating sharply multiply transitive groups and checking the $K(n)$ stabiliser recursion, giving a computational route to verify the theorem without relying on the unstated sporadic computations.","If the generalised Common Neighbour Conjecture is ever proved, Theorem 4.5 would immediately resolve Conjecture 4.3 in all the listed classes; conversely, any counterexample to Conjecture 4.3 within those classes would disprove the generalised conjecture, tying the two conjectures together as a single testable target."],"forward_implications":["Completeness of the Saxl hypergraph is a very rigid property: a group outside the five listed families must contain a $b(G)$-element subset whose pointwise stabiliser is nontrivial, so the classification gives a ready-made certificate for non-completeness.","The recursion of Lemma 3.1 transfers the problem to point stabilisers, so any future classification of $K(n-1)$ groups automatically extends to $K(n)$ groups by taking stabilisers.","For primitive groups of base size at least three, the common-neighbour conjecture, if true, implies not just one but at least two common neighbours for every vertex pair, strengthening the conjectured diameter bound.","Under the generalised Common Neighbour Conjecture, Conjecture 4.3 holds for all primitive groups with base size at most seven, all affine groups, almost simple groups in non-standard actions, many diagonal- and product-type groups, and all groups of degree at most 128.","For primitive groups with base size 3 or 4, a flag-spanning tour exists except in the explicit families listed in Theorem 1.6, including odd-degree groups, certain soluble affine groups, product-type exceptions, $\\mathrm{PSL}_2(q)$ and $\\mathrm{PGL}_2(q)$ with $q \\equiv 3 \\bmod 4$, and $\\mathrm{P}\\Gamma\\mathrm{L}_2(2^e)$ for square-free odd $e$."],"supporting_citations":[{"why":"Supplies the classification of finite 2-transitive groups that the proof of Theorem 1.3 and Theorem 5.2 run through.","marker":"[19]"},{"why":"Defines the Saxl graph for base-size-two groups and states the original Common Neighbour Conjecture that this paper generalises.","marker":"[8]"},{"why":"Introduces the generalised Saxl graph and the walk equivalence used in Lemma 2.1 to transfer connectivity statements between the graph and hypergraph.","marker":"[20]"},{"why":"Gives the base-size bound of at most seven for almost simple groups in non-standard actions, used in Theorem 4.5(iii).","marker":"[12]"},{"why":"Provides the base-size bound for primitive diagonal-type groups used in Theorem 4.5(iv).","marker":"[24]"},{"why":"Gives the criterion for base sizes of wreath products via the distinguishing number of the top group, used in the product-type arguments.","marker":"[2]"},{"why":"Bounds the distinguishing number of soluble and primitive top groups, supplying the regular orbits needed in the product-type proof.","marker":"[35,36]"},{"why":"Supplies the base-size bound for primitive groups of degree $n$ used to handle the degree-at-most-128 case of Theorem 4.5.","marker":"[32]"},{"why":"Establishes the even-vertices-and-even-valency criterion for flag-spanning tours in hypergraphs that underlies Theorem 1.6.","marker":"[1]"},{"why":"Tabulates soluble maximal subgroups and base sizes of almost simple groups used in the valency and tour classification.","marker":"[7]"}],"fun_headline_variants":["Complete Saxl hypergraphs: only five family types exist","Saxl hypergraph complete iff group in one of five families","Only five families: classification of complete Saxl hypergraphs","Five families solve complete Saxl hypergraph classification"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The classification of Theorem 1.3 depends on the standard classification of finite 2-transitive groups and on unshown computer checks for the sporadic actions, while the common-neighbour results in Theorem 4.5 further assume the unproved generalised Common Neighbour Conjecture.","fun_headline_variants_meta":{"raw":{"variants":["Complete Saxl hypergraphs: only five family types exist","Saxl hypergraph complete iff group in one of five families","Only five families: classification of complete Saxl hypergraphs","Five families solve complete Saxl hypergraph classification"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000401,"raw_usage":{"total_tokens":2146,"prompt_tokens":1049,"completion_tokens":1097,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":665,"completion_tokens_details":{"reasoning_tokens":1030}},"tokens_in":665,"tokens_out":1097,"duration_ms":9077,"temperature":1.0,"reasoning_tokens":1030,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T20:08:47.600425+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct computation in the four sporadic actions of Theorem 1.3(v) would settle the asserted computational confirmation: list all bases of minimal size and verify that every $b(G)$-element subset has trivial pointwise stabiliser. For the common-neighbour part, test Conjecture 4.3 on the affine groups with base size at most seven; a pair of vertices whose minimal-base edges always meet in at least two points would be a counterexample, and under Theorem 4.5 would disprove the generalised Common Neighbour Conjecture for that group.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the classification of finite 2-transitive groups that the proof of Theorem 1.3 and Theorem 5.2 run through."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the Saxl graph for base-size-two groups and states the original Common Neighbour Conjecture that this paper generalises."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the generalised Saxl graph and the walk equivalence used in Lemma 2.1 to transfer connectivity statements between the graph and hypergraph."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the base-size bound of at most seven for almost simple groups in non-standard actions, used in Theorem 4.5(iii)."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the base-size bound for primitive diagonal-type groups used in Theorem 4.5(iv)."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the criterion for base sizes of wreath products via the distinguishing number of the top group, used in the product-type arguments."},{"cited_title":"Moscatiello and C","cited_arxiv_id":null,"evidence_quote":"Supplies the base-size bound for primitive groups of degree $n$ used to handle the degree-at-most-128 case of Theorem 4.5."},{"cited_title":"Spanning Euler Tours in Hypergraphs","cited_arxiv_id":"2403.12713","evidence_quote":"Establishes the even-vertices-and-even-valency criterion for flag-spanning tours in hypergraphs that underlies Theorem 1.6."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Tabulates soluble maximal subgroups and base sizes of almost simple groups used in the valency and tour classification."}],"review_version":1}