{"id":"19f44805-1314-4abd-a2d4-3667e000fa73","arxiv_id":"2411.16307","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The authors enumerate the non-isomorphic Steiner triple systems of orders 19, 27 and 31 that contain Veblen points: 3, 1736, and 2 with exactly three Veblen points, plus partial counts for order 31 with one Veblen point.","lead":"This paper counts the small hand-shaking puzzle designs, called Steiner triple systems, that contain a highly regular point, for sizes 19, 27 and 31. The exact counts give new data on a classification problem that remains unsolved for larger sizes.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The exact counts for STS(31)s with one Veblen point rest on an unverified 2^35-factor-system enumeration; an independent recomputation of Table 4 is needed.","rationale":"The paper's mathematical framework is coherent: the reductions to Schreier extensions, the coboundary counts, and the orbit computations are internally consistent, and the small cases (Theorems 2.1 and 4.2) are plausible and partly hand-checkable. The most serious issue is not a logical contradiction but the lack of independent confirmation of the large enumeration behind Theorem 4.1 (and, to a lesser degree, Theorem 3.1). The appendix itself notes that storing all 2^35 factor systems would be impractical and that a batch algorithm was required, so a subtle bit-representation or orbit-counting error is possible. If such an error occurred, the exact counts in Table 4 would be wrong. This does not amount to a demonstrated flaw, so the appropriate disposition is conditional acceptance pending independent verification, matching the reader's verdict.","tokens_in":48,"tokens_out":28945,"duration_ms":397715,"concrete_test":"For the quotient Q = PG(3,2), recompute the count 278 with an independent method: enumerate all 2^24 non-equivalent factor systems by a different coset-representative algorithm, apply the action of Aut(Q) using canonical forms, construct the resulting STS(31)s, and use nauty canonical labeling to confirm that exactly 278 isomorphism classes have a single Veblen point. If this count matches, the computational pipeline is validated; if it does not, Table 4 is wrong. A smaller supporting check is to rerun the provided code for the two STS(13) quotients in Theorem 3.1 and independently verify the orbit counts 1504 and 232.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central enumerative claims (Theorems 2.1, 3.1, 4.1, 4.2) all depend on the reduction of isomorphism classes of STS extensions to orbits of Aut(LN) × Aut(LQ) on non-equivalent factor systems, a result imported from the authors' [6]. That is a theoretical dependency, but the more fragile load-bearing step is the computation, especially Theorem 4.1: for each of six STS(15) quotients it enumerates 2^35 factor systems, reduces by coboundaries, and then orbits under Aut(Q), filtering by Remark 1.1. The appendix reports that a naive store of all factor systems would require about 64 PB and that a batch algorithm was used; no independent certificate is supplied. A bug in the encoding of fundamental pairs, in the group action on the 35-bit representation, or in the extra-Veblen filter would change every entry of Table 4, and a similar issue could affect Theorem 3.1. The order-19 and order-27 cases are small enough to be checked, but the order-31 numbers are not backed by any second implementation.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper uses the Schreier-extension theory of Steiner loops, developed by the authors in a previous paper [6], to enumerate Steiner triple systems of orders 19, 27, and 31 that contain Veblen points. The main results are: (i) exactly 3 non-isomorphic STS(19)s have a unique Veblen point, and explicit factor systems are given; (ii) exactly 1736 non-isomorphic STS(27)s have a unique Veblen point, split as 1504 with the non-cyclic STS(13) as quotient and 232 with the cyclic STS(13) as quotient; (iii) exactly 2 non-isomorphic non-projective STS(31)s have exactly three Veblen points; (iv) for six selected quotients Q of order 15, Table 4 lists the numbers of STS(31)s with a unique Veblen point and quotient Q, obtained by enumerating 2^35 factor systems per quotient and reducing by coboundaries and automorphism orbits.","tokens_in":14592,"tokens_out":14421,"duration_ms":118274,"significance":"If correct, these are exact isomorphism-class counts for subclasses that have not been enumerated before, and they demonstrate the reach of the loop-extension method for STS. The paper provides explicit coordinate representations for the three STS(19)s and the two STS(31)s with three Veblen points, and it makes its source code publicly available, which are concrete strengths. The counts are, however, contingent on two external supports: the classification theorem from [6] that identifies isomorphism classes with automorphism orbits, and a large batch computation (2^35 factor systems per quotient) for the order-31 case that is not independently certified in the manuscript.","major_comments":[{"comment":"The theorem statement is broader than the result actually proved: it claims that the number of non-isomorphic STS(31)s with exactly one Veblen point and quotient system Q of order 15 is given by Table 4, but Table 4 lists only six of the 80 non-isomorphic STS(15)s, and the proof explicitly restricts to those six. Please restate the theorem as 'for the quotient systems listed in Table 4' or otherwise qualify it, otherwise the reader may conclude that a complete classification for order 31 has been obtained.","section":"Section 4, Theorem 4.1 and Table 4"},{"comment":"The central reduction that isomorphism classes of extensions are exactly the orbits of Aut(LN) × Aut(LQ) on non-equivalent factor systems when LN is the whole center, as well as the extra-Veblen criterion of Remark 1.1, are imported from [6] without proof or a precise statement of the hypotheses. These are load-bearing in Theorems 2.1, 3.1, 4.1, and 4.2. If the theorem in [6] requires conditions (e.g., that the action is well-defined on cohomology classes, or that every STS with the given center arises from such a factor system), the applications here need to be checked against those conditions. Please include a formal statement of the imported results (or a proof sketch) and verify the hypotheses explicitly for each application.","section":"Section 1.1, Eq. (2) and Remark 1.1"},{"comment":"The order-31 counts in Table 4 rest on an enumeration of 2^35 factor systems for each of the six quotients, and the paper provides no independent certificate or fully reproducible algorithmic specification. The appendix describes the encoding of factor systems as integers, the computation of coboundaries, and a batch approach, but it does not give pseudo-code for the orbit computation or the extra-Veblen filter, nor does it report checks against known counts on smaller orders. Since a bug in the encoding of fundamental pairs, the induced permutation on the 35-bit representation, the orbit computation, or the filter would change every entry of Table 4, the authors should supply a stronger correctness argument: for example, reproductions of the orbit counts for the order-19 and order-27 cases using the same code, or independent verification of Table 4 by a second implementation.","section":"Appendix A.1 and Theorem 4.1"}],"minor_comments":[{"comment":"The word 'Setiner' is a typo and should be 'Steiner'.","section":"p. 4, first paragraph of Section 1"},{"comment":"The statement '4^7 = 16383' is arithmetically incorrect; 4^7 = 16384. The subsequent division by |B^2| = 256 is consistent with the correct value, so this is a typo, but it should be fixed.","section":"Section 4, proof of Theorem 4.2"},{"comment":"The sentence 'these orbits do not necessarily coincide with the isomorphism classes of STS(19)s with three Veblen points' should refer to STS(31)s, not STS(19)s.","section":"Section 4, proof of Theorem 4.2"},{"comment":"The caption reads 'The STS(19) S1 with unique Veblen point 0', but the table is for S2, as the text immediately before it explains.","section":"Table 3 caption"},{"comment":"In the sentence 'Let us now consider a factor systems f represented as a vector ... and two automorphisms α ∈ Aut(LN), β ∈ Aut(LN)', the second automorphism should be β ∈ Aut(LQ), since it acts on the quotient coordinates.","section":"Appendix A.1, paragraph on automorphisms"}],"recommendation":"major_revision","confidential_remarks":"The paper fits the scope of a combinatorial-designs journal and the underlying extension-theoretic framework appears coherent. The main concerns are the overbroad statement of Theorem 4.1, the unproved imported bijection from [6], and the lack of a strong independent correctness check for the large order-31 computation. These are addressable within a revision, so I recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this paper delivers the first exact counts for small Steiner triple systems with Veblen points — three STS(19)s, 1736 STS(27)s, and a partial but substantial table for STS(31)s with one Veblen point, plus two STS(31)s with three Veblen points. The method is the Schreier-extension framework the authors developed in [6], applied systematically. The three STS(19)s are fully written out with factor systems and Pasch switches, which is nice and checkable by hand.\n\nThe real strengths: the counting pipeline is coherent — factor systems, coboundary reduction, then orbits under Aut(Q), with an explicit filter (Remark 1.1) to exclude extensions where the center grows. The order-19 and order-27 parts are small enough to be independently sanity-checked, and the authors ship code and describe hardware. The order-31 three-Veblen-point result reduces to three explicit factor systems on the Fano plane and is fully documented.\n\nSoft spots. The load-bearing bijection between factor-system orbits and isomorphism classes is imported from [6], not proved here. That is acceptable for an application paper, but a referee needs to check it. The bigger issue is Theorem 4.1: the six entries of Table 4 come from enumerating 2^35 factor systems per quotient, with a batch algorithm, and no independent implementation or certificate. A bug in fundamental-pair encoding, the group action, or the extra-Veblen filter would change every row. The order-19 and order-27 numbers are probably safe; I would not trust the order-31 rows until a second implementation or a cross-check exists. There are also smaller signs of carelessness: in Theorem 4.2, 4^7 is written as 16383 instead of 16384, and the text says 'STS(19)s' where it means STS(31)s; the appendix's 64 PB storage estimate for the factor systems is off by several orders of magnitude (the actual naive bit count is around 150 GB, and the formula printed is wrong). None of these touches the counts directly, but they undercut confidence.\n\nBottom line: this is a serious paper for the design-theory/enumeration audience. It deserves peer review — the enumeration approach is legitimate and the small-order results are probably right — but Table 4 needs an independent recomputation before the results should be quoted as established. My honest advice: send it out, but require the authors to address the computation verification and clean up the errors.","headline":"First exact counts for STS(19), (27), and partial (31) with Veblen points; credible for small orders, but the order-31 table rests on an unverified 2^35 enumeration.","tokens_in":15135,"tokens_out":4034,"would_cite":true,"duration_ms":181662,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["05B07","20N05","51E10"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper establishes exact counts of Steiner triple systems with Veblen points at orders 19, 27 and 31.","keywords":["Steiner triple systems","Veblen points","Schreier extensions","Steiner loops","central extensions","STS(19)","STS(27)","STS(31)"],"falsifier":"Search the published complete classification of the 11,084,874,829 STS(19)s for Veblen points: if the number of isomorphism classes with exactly one Veblen point is not three, Theorem 2.1 fails. For the STS(31) counts, re-run the factor-system enumeration with an independently written program; a single discrepancy in Table 4 or in the two three-Veblen-point systems would falsify the corresponding theorem.","tokens_in":14186,"feed_emoji":"🔢","tokens_out":7683,"duration_ms":61330,"temperature":0.7,"pith_summary":"The paper establishes exact isomorphism-class counts for Steiner triple systems of small orders that contain Veblen points — points through which any two triples generate a Pasch configuration. It proves that exactly three of the 11,084,874,829 non-isomorphic STS(19)s have one Veblen point, that exactly 1,736 non-isomorphic STS(27)s have one Veblen point, and that exactly two non-isomorphic non-projective STS(31)s have three Veblen points. It also gives exact counts, for six selected quotient systems, of STS(31)s with exactly one Veblen point. These are complete counts within the corresponding subclasses, not asymptotic estimates. The interest is that counting all Steiner triple systems beyond order 21 is open, while Veblen-point-bearing systems are regular enough that a loop-theoretic construction can enumerate them exactly.","feed_headline":"Only 3 of 11 billion STS(19)s have one Veblen point","feed_subtitle":"Exact counts via Schreier extensions: 1,736 STS(27)s and just 2 non-projective STS(31)s with three Veblen points.","key_machinery":"The machinery is the Schreier extension of Steiner loops. Every Steiner triple system gives a Steiner loop on S ∪ {Ω}, and Veblen points are exactly the nontrivial central elements of that loop. Starting from a central elementary abelian 2-group L_N and a quotient Steiner loop L_Q, every extension is encoded by a symmetric factor system f : L_Q × L_Q → L_N that satisfies a prescribed condition on the triples of Q; the total number of factor systems is |L_N|^b where b is the number of triples of Q. Two extensions are equivalent when they differ by a co-boundary, and when L_N is the full center, isomorphism classes are the orbits of Aut(L_N) × Aut(L_Q) on non-equivalent factor systems. The extra-Veblen criterion of Remark 1.1 decides, via a linear condition on f, whether an element outside N becomes central, and this filter is what separates the STS(31) cases where the quotient itself has Veblen points.","core_discovery":"The central discovery is that the Schreier-extension machinery for Steiner loops yields exact global counts in small orders where a full classification is infeasible. Concretely, Theorem 2.1 identifies the unique three STS(19)s with exactly one Veblen point; Theorem 3.1 gives 1,504 plus 232, hence 1,736, STS(27)s with exactly one Veblen point, split by whether the quotient STS(13) is non-cyclic or cyclic; Theorem 4.2 gives exactly two non-projective STS(31)s with exactly three Veblen points, described by explicit factor systems f1 and f2; and Theorem 4.1 with Table 4 gives exact STS(31) one-Veblen-point counts for six chosen quotient STS(15)s, from 278 for PG(3,2) to 99,952 for STS(15)#61.","pith_inferences":["The same orbit-counting recipe could be applied to order 21, where no complete classification exists, to obtain exact counts for the one-Veblen-point subfamily and thereby a lower bound on the total number of STS(21)s.","The three-Veblen-point count being only two suggests that non-projective systems with many Veblen points may be rare; a testable conjecture is that among non-projective STS(v) the number with at least c Veblen points grows far slower than the total number of systems.","The Appendix describes a batch computation over 2^35 factor systems, so an independent reimplementation on different hardware would directly confirm the six values in Table 4.","Because the two STS(31)s with three Veblen points are presented only through factor systems, an explicit 155-triple listing could be extracted and checked for additional properties, such as which subsystems of order 7 or 9 they contain."],"forward_implications":["Any complete classification of STS(19) must contain exactly the three systems described in Theorem 2.1 as its Veblen-point-bearing members.","Any full enumeration of STS(27) must include exactly 1,736 systems with a single Veblen point: 1,504 with the non-cyclic STS(13) quotient and 232 with the cyclic one.","For order 31, apart from PG(4,2), exactly two non-isomorphic systems have three Veblen points, so the three-Veblen-point family is completely known.","The six counts in Table 4 pin down, exactly, how many STS(31)s over specific quotients have one Veblen point; for example, only 278 arise over PG(3,2) despite 2^24 non-equivalent extension classes."],"supporting_citations":[{"why":"Supplies the Schreier-extension construction, the equivalence and isomorphism criteria for factor systems, and the extra-Veblen criterion used as Remark 1.1.","marker":"[6]"},{"why":"Introduces Schreier loops in the general case that the extension theory is built upon.","marker":"[11]"},{"why":"Provides the complete classification of STS(19)s, giving the total 11,084,874,829 against which the three-Veblen-point count is measured.","marker":"[9]"},{"why":"Source of the 80 STS(15) systems with their labels #1 through #80, as well as the STS(13) tables used as quotient systems.","marker":"[3]"},{"why":"Gives the characterization of projective Steiner triple systems by the property that every point is a Veblen point.","marker":"[4]"},{"why":"Supplies the Pasch-switch terminology and properties used to describe the three STS(19) systems.","marker":"[5]"}],"fun_headline_variants":["Exact count: just 3 STS(19)s with one Veblen point","Only 1,736 STS(27)s have exactly one Veblen point","Two non-projective STS(31)s with three Veblen points","Exact STS counts with Veblen points: 19, 27, 31"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The counts rest on the assumption that every Steiner triple system with the prescribed number of Veblen points arises as a Schreier extension of its full center by its quotient, and that the isomorphism classes of these systems are exactly the automorphism orbits of factor systems described in the paper, with the paper's criterion correctly identifying all additional Veblen points.","fun_headline_variants_meta":{"raw":{"variants":["Exact count: just 3 STS(19)s with one Veblen point","Only 1,736 STS(27)s have exactly one Veblen point","Two non-projective STS(31)s with three Veblen points","Exact STS counts with Veblen points: 19, 27, 31"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000469,"raw_usage":{"total_tokens":2276,"prompt_tokens":826,"completion_tokens":1450,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":442,"completion_tokens_details":{"reasoning_tokens":1362}},"tokens_in":442,"tokens_out":1450,"duration_ms":9345,"temperature":1.0,"reasoning_tokens":1362,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T13:16:04.763564+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Search the published complete classification of the 11,084,874,829 STS(19)s for Veblen points: if the number of isomorphism classes with exactly one Veblen point is not three, Theorem 2.1 fails. For the STS(31) counts, re-run the factor-system enumeration with an independently written program; a single discrepancy in Table 4 or in the two three-Veblen-point systems would falsify the corresponding theorem.","supporting_citations":[{"cited_title":"Extensions of Stei ner triple systems","cited_arxiv_id":null,"evidence_quote":"Supplies the Schreier-extension construction, the equivalence and isomorphism criteria for factor systems, and the extra-Veblen criterion used as Remark 1.1."},{"cited_title":"Schreier loops","cited_arxiv_id":null,"evidence_quote":"Introduces Schreier loops in the general case that the extension theory is built upon."},{"cited_title":"The Steiner triple systems of order 19","cited_arxiv_id":null,"evidence_quote":"Provides the complete classification of STS(19)s, giving the total 11,084,874,829 against which the three-Veblen-point count is measured."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Source of the 80 STS(15) systems with their labels #1 through #80, as well as the STS(13) tables used as quotient systems."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the characterization of projective Steiner triple systems by the property that every point is a Veblen point."},{"cited_title":"Properties of the Steiner triple sy stems of order 19","cited_arxiv_id":null,"evidence_quote":"Supplies the Pasch-switch terminology and properties used to describe the three STS(19) systems."}],"review_version":1}