Combinatorics · arXiv:2611.16299 v1 · submitted 12 Nov 2026 · 46 references
Improved Lower Bounds for Mutually Orthogonal Latin Squares (MOLS) at 33 Orders
The paper says
“A Latin square of order n is an n × n array over symbols {0,1,…n−1}, such that the same symbol does not appear twice in any row or any column.”
more of the abstract
Two Latin squares of order n are orthogonal if, when superimposed, they yield each of the n2 ordered pairs of symbols exactly once. A set of two or more Latin squares are mutually orthogonal Latin squares (MOLS) if each pair of distinct squares is orthogonal. Let N(n) be the maximum size of a set of MOLS of order n. For most n which are not prime powers, there is a wide gap between known upper and lower bounds on N(n). This has led to interest in constructing new larger sets of MOLS for explicit values of n, to incrementally improve the lower bounds on N(n). This paper provides MOLS that improve the lower bound on N(n) for 33 values of n in the range from 22 through 494. The constructions were found by GPT-6 Astra, and use established techniques including difference matrices, quasi-difference matrices, orthogonal arrays, and projective planes. A highlight is 4 MOLS of order n = 22; combined with pre-existing results, this establishes N(n) ≥ 4 for all n > 10.
from the abstract
Pith says
Machine review · deepseek-v4-flash · 13 Nov 2026 · what was read was not recorded · not checked by a person
It can be wrong. Challenge this review or reply to it.
The passes disagreed. Reader: holds up. Skeptic: holds up only with conditions. Referee: major revision.
33 new MOLS lower bounds, including N(22)≥4, all from GPT-6 Astra; the math is standard and the certificates are checkable, but the paper's reliance on an AI system it does not describe and the absence of independent verification of the finite checks are real concerns. This line is wrong
4 major objections · 15 minor
How often objections from reviews like this one are right: not yet measured
Not checked: figures, code, what the 46 references say.
The paper says, at §8.1, §2, Theorem 2.2
The review did not quote this passage.
Pith says: Objection 1 of 4 · major
The central claim of Theorem 2.2 (N(n) >= 4 for all n > 10) hinges on the 4 MOLS of order 22. The construction is described via a quasi-difference matrix over Z_21 with one infinite symbol. The authors state that 'Both properties follow by enumerating the displayed rows' (after matrices R and F).
more of objection 1
However, the verification that the 15 finite difference lists are permutations of Z_21 is not shown in the manuscript; it is merely asserted. The repository contains Python code that performs this check, but the manuscript does not reproduce the verification output or provide any independent evidence (e.g., a printed list of the 15 permutations) that the QDM conditions hold. Given that this is the linchpin of Theorem 2.2, the manuscript should either include the verification data or provide a more detailed argument. The same issue applies to all 33 orders: the certificates are summarized but not fully displayed, and the verification code is not audited within the manuscript.
The paper says, at §6.2, §7.1, §7.2, §10.1, §15.6
The review did not quote this passage.
Pith says: Objection 2 of 4 · major
The paper repeatedly asserts that finite checks (e.g., 'all 288 affine partition conditions', 'all 252 character tests', 'all 144 line partitions', 'all 220 small affine partitions') are satisfied, but does not provide the actual verification data or a proof that these conditions hold.
more of objection 2
For example, in §6.2 the authors state 'The data satisfy all 288 affine partition conditions' without showing the data or the verification. In §7.1, they claim 'All 144 groups pass the same check' but only give a single example. In §10.1, they assert 'all 252 character conditions' are satisfied but only sketch a compact check. In §15.6, Table 3 gives intersection histograms but the derivation is not fully explained. While it is reasonable to rely on finite checks for a construction with repository verification, the manuscript should at least provide a representative sample of the verification output or a more detailed argument to convince readers who cannot run the code.
The paper says, at §15, Table 3
The review did not quote this passage.
Pith says: Objection 3 of 4 · major
The geometric constructions in Section 15 rely on the correctness of the line-intersection histograms in Table 3. The authors state that 'The values follow by enumerating the explicit points and lines above' but do not provide the enumeration or a proof that the histograms are correct.
more of objection 3
Given that these histograms are essential for verifying that the PBD filling works (each block size must be a prime power supporting the local filling), the manuscript should either include the enumeration data or provide a more rigorous derivation. The formulas in §15.2 (ray-intersection counts) and §15.3 (hyperbola selections) give some justification, but the actual histograms are not derived from these formulas in the text.
The paper says, at §14, Tables 9-18
The review did not quote this passage.
Pith says: Objection 4 of 4 · major
The affine constructions in Section 14 rely on extensive finite data (offsets, maps, coarse arrays) provided in the appendices. The correctness of these constructions depends on the data satisfying the affine partition conditions. The authors state that 'The finite check is exactly its difference partition condition' but do not provide any verification that the data satisfy these conditions.
more of objection 4
Given the complexity of the data, it would be helpful to include a brief verification argument or at least state that the data have been verified by the repository code.
15 minor objections
The paper says, at §1, Abstract
Pith says: minor objection 1 of 15
The abstract states 'The constructions were found by GPT-6 Astra'. This is interesting but not standard for a mathematics paper. Consider moving this to the acknowledgments or a separate section, as it is not relevant to the mathematical content.
The paper says, at §3, AI Usage
Pith says: minor objection 2 of 15
The AI usage section is detailed but somewhat unusual for a mathematics journal. The paper should clarify whether the AI contributed to the mathematical reasoning or only to the search/implementation. The statement 'The descriptions of the constructions in sections below were written by GPT-6 Astra, with human author review and light editing' raises questions about authorship and attribution. This should be addressed according to the journal's policy.
show 13 more
The paper says, at §4.4, QDM example
Pith says: minor objection 3 of 15
The small example of a QDM with G=Z_3, h=1, w=3 is helpful but the matrix Q is displayed with inconsistent formatting (the matrix appears to have 5 rows but the text says 'five rows'). Also, the example uses infinity symbols but the final Latin square L has entries 0,1,2,3 where 3 represents infinity. This should be clarified.
The paper says, at §5.1, small example
Pith says: minor objection 4 of 15
The small example of two MOLS of order 12 is well-explained but the table of fine differences could be formatted more clearly. The current formatting with braces and commas is a bit hard to read.
The paper says, at §8.1, Order 22
Pith says: minor objection 5 of 15
The matrices R and F are displayed but the formatting is inconsistent: R has 6 rows and F has 5 rows, but the text says 'six base rows and five supplemental rows'. The matrix R appears to have 6 rows, but the first row has only 5 entries (∞,0,0,0,1,4) while the other rows have 6 entries. This should be checked and corrected.
The paper says, at §9, Order 85
Pith says: minor objection 6 of 15
The construction uses cycles A0, A1, A2 with entries modulo 85. The check that the four lists of 21 integers equal R is described but the actual lists are not shown. It would be helpful to include the computed lists or at least state that they have been verified.
The paper says, at §10.2, Order 78
Pith says: minor objection 7 of 15
The verification of the 252 character tests is sketched but the actual checks are not fully explained. The statement 'Alternatively, check all 252 conditions directly' is not helpful. Either provide the verification or state that it is in the repository.
The paper says, at §11, Order 90
Pith says: minor objection 8 of 15
The construction uses a field K of order 81 and a subspace W of order 9. The definition of W as ⟨1,v⟩_{F_3} is clear, but the statement 'The proof requires only that W be closed under addition and negation' is a bit vague. It would be helpful to explain why this is sufficient.
The paper says, at §12, Order 60
Pith says: minor objection 9 of 15
The construction uses operations T0, T1, T2 with constants c0, c1, c2. The definition of c1 and c2 as tuples of length 8 with entries in G is clear, but the addition 'fixes any infinity entry' should be clarified: does it mean that infinity entries are left unchanged by the addition?
The paper says, at §13, Order 225
Pith says: minor objection 10 of 15
The construction uses functions B and C defined by tables. The statement 'Direct inspection of the two short tables shows that id, B, C, id-B, id-C, B-C are permutations of Z_15' is not fully justified. It would be helpful to explain why these differences are permutations, perhaps by noting that the tables are derived from MOLS of order 15.
The paper says, at §15.1, Plane incidence
Pith says: minor objection 11 of 15
The description of the projective plane construction is clear but the notation for direction functions f_m(x) could be simplified. The use of mx+m^3 x^9 for order 456 and mx or mx^3 for order 482 is explained but the justification that these give unique pair coverage is somewhat terse. The reference to Albert's generalized twisted field and André's norm-class switching is helpful but could be expanded.
The paper says, at §15.5, Maximal arcs
Pith says: minor objection 12 of 15
The description of maximal arcs for orders 207 and 175 is very brief. The definition of the arc A uses a quadratic form x^2 + xy + 8y^2 ∈ {0,1,2,3} but it is not explained why this gives a maximal arc. The reference to Denniston's construction is appropriate but the text should at least state that these are known maximal arcs.
The paper says, at Appendix B
Pith says: minor objection 13 of 15
The four squares of order 22 are printed in full. This is helpful for verification but takes up a lot of space. Consider whether this is necessary or if a reference to the repository would suffice.
The paper says, at General
Pith says: minor objection 14 of 15
The paper uses a non-standard notation for encoding vectors (e.g., 'enc' function). This is explained but could be confusing. Consider using standard notation or providing a glossary.
The paper says, at General
Pith says: minor objection 15 of 15
The paper references a repository at https://github.com/Constructive-Codes/MOLS. It would be helpful to include a specific commit hash or version number to ensure reproducibility.
What would settle it
Independently verify the 4 MOLS of order 22 by running the provided Python verification code (or re-implementing the QDM development from Section 8.1) and confirming that the 15 finite difference lists are each a permutation of Z_21. If any difference list fails, the QDM is invalid and Theorem 2.2 collapses. If all pass, the concern is resolved.
from the skeptic's reading
How the review went
4 readings by deepseek-v4-flash · 13 Nov 2026
- Reader holds up
- Skeptic agrees: says it holds up only with conditions The paper's central claim depends on the correctness of 33 finite certificates, but the verification code is not displayed in the manuscript and the repository is not independently audited; the most load-bearing concern is whether the 4 MOLS of order 22 are actually valid.
- Editor wrote the verdict above
- Referee major revision: 4 major objections, 15 minor
Not run on this paper: the simulated authors' defence.
Pith did not check this review's objections against the paper.
The review in full
The review in brief, machine-written
Chris,
This paper reports 33 improved lower bounds for MOLS, the headline being 4 MOLS of order 22, which together with known results gives N(n)≥4 for all n>10. The constructions use standard techniques—difference matrices, quasi-difference matrices, affine lifts, projective plane selections—and the paper provides explicit certificates and a repository with verification code. The results are new by the paper's own comparison to the 2024 compilation, and the general lemmas are coherent and mostly proved. The finite checks are finite, so in principle the claims are checkable.
That said, there are two soft spots worth noting. First, the paper states that all constructions were found by GPT-6 Astra, but it gives almost no detail about how the AI was used, what prompts were tried, or how the outputs were validated beyond the author's own verification code. The prompts in Appendix A are generic and don't explain how the specific constructions were discovered. This makes it hard to assess whether the results are reproducible by others or whether the AI's role was more than a search heuristic. Second, the finite certificates are summarized but not fully displayed in the manuscript; the verification code is in a repository but not independently audited. The most load-bearing case is the 4 MOLS of order 22, where the QDM conditions are asserted to hold by enumerating the displayed rows, but the actual permutations are not printed. The stress-test note flags this correctly. The reader's weakest_assumption also identifies this as the key uncertainty. I think this is a real concern, but it's a concern about verification, not about the mathematics: the construction is finite and checkable, and the repository exists. The paper would be stronger if it included a printed list of the 15 permutations for order 22, or at least a hash of the verification output.
Overall, this is a solid computational construction paper with explicit finite certificates. The results are new and the methods are standard. The reliance on an AI system that is not described in enough detail is a weakness, but not a fatal one. The paper deserves a serious referee who can run the verification code and check the finite claims. I would not cite it in my own work until the certificates are independently verified, but I would bring it to a reading group interested in computational design theory.
Recommendation: send to peer review. The referee should be asked to verify at least the order-22 case using the provided code.
The strongest objection
The paper's Theorem 2.2 (N(n) >= 4 for all n > 10) hinges on the 4 MOLS of order 22. The construction is described in Section 8.1 via a quasi-difference matrix over Z_21 with one infinite symbol. The authors state that 'Both properties follow by enumerating the displayed rows' (Section 8.1, after the matrices R and F). However, the verification that the 15 finite difference lists are permutations of Z_21 is not shown in the manuscript; it is asserted. The repository contains Python code that performs this check, but the manuscript does not reproduce the verification output. The reader's weakest_assumption correctly identifies this as the key uncertainty. While the construction is finite and checkable, the paper provides no independent evidence (e.g., a printed list of the 15 permutations) that the QDM conditions hold. The same issue applies to all 33 orders: the certificates are summarized but not fully displayed. The most load-bearing single point is the order-22 case because it is the linchpin of Theorem 2.2.
The reader's verdict is ACCEPT with MODERATE confidence, and the weakest_assumption is that the finite certificates are correctly transcribed and checked. I agree that this is the central uncertainty. However, I argue that the paper's reliance on unverified certificates is a load-bearing concern that should move the verdict to CONDITIONAL: the paper should be accepted only if the repository verification is independently confirmed. The paper itself states that 'the data satisfy all 288 affine partition conditions' (Section 6.2) and 'all 252 character tests' (Section 10.1), but these are assertions without displayed evidence. The order-22 case is the most critical because it is the basis for Theorem 2.2. The paper provides a linked repository, which is good, but the manuscript does not include the verification output. A CONDITIONAL verdict is appropriate: accept subject to independent verification of the certificates. I do not recommend REJECT because the construction is finite and checkable, and the paper provides the necessary data (matrices R and F) for independent verification. The concrete test is straightforward and would settle the concern definitively.
The referee's report
If correct, the results represent incremental but meaningful progress on a long-standing combinatorial design problem. The improvement for n=22 is particularly notable as it closes the last open case for N(n) >= 4. The paper also provides a strongly regular graph construction from the 4 MOLS of order 22, which was previously open. The use of AI (GPT-6 Astra) to find constructions is novel but not central to the mathematical validity.
Pith's reading of the paper (a machine's summary, not the authors')
A Latin square of size n is an n by n grid in which each row and each column contains every symbol from 0 to n-1 exactly once. Two Latin squares are orthogonal when their cells, read as pairs, cover all n^2 possible symbol pairs exactly once. The paper works on the classical question of how many mutually orthogonal Latin squares (MOLS) exist for a given n, written N(n). For prime powers, the answer is n-1, but for many other values of n the exact answer is unknown.
The paper gives explicit constructions that raise the known lower bound for 33 orders between 22 and 494. Most gains are by 1, several by 2 to 5, and the most consequential is 4 MOLS of order 22. Together with existing published results, this yields N(n) at least 4 for every n larger than 10, closing the last open case for that threshold. The constructions use standard design tools: difference matrices, quasi-difference matrices, lifted orthogonal arrays, and carefully chosen point sets in finite projective planes. All order-specific data appear in appendices, and the paper links to a repository with verification code for each resulting MOLS set.
The constructions were found by an LLM, GPT-6 Astra, and the description text and code were also written by it, with human review. The authors report that an independent language model reconstructed the same results from the descriptions alone, and both versions were verified separately. The value of the paper lies in the explicit certificates and the finite checks, not in new asymptotic theory.
Core claim. The paper's Theorem 2.2: 'The 4 MOLS of order n=22, combined with pre-existing lower bounds [9] and [15, Theorem 5.1], give N(n) >= 4 for all n>10.' The paper also claims Theorem 2.1, that every Constructed value in Table 1 is a valid lower bound on N(n), including N(74) >= 6. If true, explicit MOLS sets exist with these sizes and are checkable from the provided certificates.
The premise it rests on. The theorems depend on the exhaustive finite certificates being correctly transcribed and correctly checked. Section 6.2 says 'The data satisfy all 288 affine partition conditions', Section 10.1 relies on all 252 character tests, and Table 3 gives line-intersection histograms whose correctness is asserted from explicit enumeration. If any certificate or histogram is wrong, the corresponding row of Table 1 fails. This is a reasonable assumption for a finite construction with repository verification, but the checks are not fully displayed inside the manuscript.
Check this review
Pith stored no record of this review's passes and took no fingerprint of it. Reviews made by code from 13 Nov 2026 on carry both.
How often reviews like this one are right
Reviews like this one: written by deepseek-v4-flash, where what was read was not recorded.
How often objections from reviews like this one are right: not yet measured
Reviews by people
This review was written by a machine. No person has reviewed this paper on Pith yet. A scientist listed on Pith can review it for a fee, and their review is published here whatever it concludes.
The paper
Abstract
A Latin square of order n is an n × n array over symbols {0,1,…n−1}, such that the same symbol does not appear twice in any row or any column. Two Latin squares of order n are orthogonal if, when superimposed, they yield each of the n2 ordered pairs of symbols exactly once. A set of two or more Latin squares are mutually orthogonal Latin squares (MOLS) if each pair of distinct squares is orthogonal. Let N(n) be the maximum size of a set of MOLS of order n. For most n which are not prime powers, there is a wide gap between known upper and lower bounds on N(n). This has led to interest in constructing new larger sets of MOLS for explicit values of n, to incrementally improve the lower bounds on N(n). This paper provides MOLS that improve the lower bound on N(n) for 33 values of n in the range from 22 through 494. The constructions were found by GPT-6 Astra, and use established techniques including difference matrices, quasi-difference matrices, orthogonal arrays, and projective planes. A highlight is 4 MOLS of order n = 22; combined with pre-existing results, this establishes N(n) ≥ 4 for all n > 10.
The paper's own sentences, arranged, kept beside the review and separate from it. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. No sentence in this record was written by a model.
abstract compiled · 2026-11-13
The authors' main claim, as the classifier picked it out 0.90
“This paper provides MOLS that improve the lower bound on $N(n)$ for 33 values of $n$ in the range from 22 through 494.”
How the main claim is stated: asserted 1.00
What the abstract itself states, as probabilities: claims first 0.04 · states limitation 0.21 · human participants 0.02 · controlled comparison 0.04 · randomized assignment 0.01 · states quantitative result 0.98 · compares with prior results 0.84 · mentions code or data availability 0.04
The abstract, sentence by sentence
This table labels the abstract only (arXiv listing abstract v1). The number is the probability the classifier gave the label.
| Sentence | Role | The authors' own work | States the main conclusion | Strength |
|---|---|---|---|---|
| A Latin square of order $n$ is an $n\times n$ array over symbols $\{0,1,\ldots n-1\}$, such that the same symbol does not appear twice in any row or any column. | background 1.00 | 0.03 | 0.01 | asserted 1.00 |
| Two Latin squares of order $n$ are orthogonal if, when superimposed, they yield each of the $n^2$ ordered pairs of symbols exactly once. | background 1.00 | 0.03 | 0.01 | asserted 0.99 |
| A set of two or more Latin squares are mutually orthogonal Latin squares (MOLS) if each pair of distinct squares is orthogonal. | background 1.00 | 0.03 | 0.02 | asserted 0.99 |
| Let $N(n)$ be the maximum size of a set of MOLS of order $n$. | background 1.00 | 0.08 | 0.02 | asserted 0.88 |
| For most $n$ which are not prime powers, there is a wide gap between known upper and lower bounds on $N(n)$. | gap 0.75 | 0.05 | 0.02 | asserted 0.99 |
| This has led to interest in constructing new larger sets of MOLS for explicit values of $n$, to incrementally improve the lower bounds on $N(n)$. | background 0.92 | 0.07 | 0.04 | asserted 0.95 |
| This paper provides MOLS that improve the lower bound on $N(n)$ for 33 values of $n$ in the range from 22 through 494. | contribution 0.65 | 0.93 | 0.92 | asserted 1.00 |
| The constructions were found by GPT-6 Astra, and use established techniques including difference matrices, quasi-difference matrices, orthogonal arrays, and projective planes. | method 0.98 | 0.57 | 0.05 | asserted 1.00 |
| A highlight is 4 MOLS of order $n=22$; combined with pre-existing results, this establishes $N(n) \geq 4$ for all $n>10$. | result 0.57 | 0.78 | 0.30 | asserted 1.00 |
What kind of paper the classifier reports this is: theoretical result 0.77 · evidence from simulation or computation 0.65
What was examined, and what was not
Sources read
- arXiv listing abstract v1
Not examined
- full text
- figures and tables
- references and citing sentences
- relations to other papers
Levels run
- P abstract
Not run
- S sections
- T body sentences
- C citations
- R cross-paper relations
How the labels were made
Compiled by classifier-1.13.0 via TypeSafe API, question bank abstract-p1-20260923, 2026-11-13: 1 request, 7,044 input tokens.
Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. This record as JSON
References (46)
- [1] R. J. R. Abel, C. J. Colbourn, and J. H. Dinitz.Mutually orthogonal Latin squares (MOLS). In C. J. Colbourn and J. H. Dinitz, editors, Handbook of Combinatorial Designs, 2nd edition, Chapman & Hall/CRC, Boca Raton, 2007, 160–193.https://doi.org/10.1201/9781420010541 on Pithdoi:10.1201/9781420010541
- [2] R. C. Bose.On the application of the properties of Galois fields to the problem of construction of hyper-Graeco-Latin squares. Sankhy¯ a 3 (1938), 323–338.https://hdl.handle.net/10263/256 on Pith
- [3] R. H. Bruck and H. J. Ryser.The nonexistence of certain finite projective planes. Canadian Journal of Mathematics 1 (1949), 88–93.https://doi.org/10.4153/CJM-1949-009-2 on Pithdoi:10.4153/cjm-1949-009-2
- [4] S. S. Shrikhande.A note on mutually orthogonal Latin squares. Sankhy¯ a, Series A 23 (1961), 115–116. https://zbmath.org/?q=an:0099.35404 on Pith
- [5] R. H. Bruck.Finite nets. II. Uniqueness and imbedding. Pacific Journal of Mathematics 13 (1963), 421–457.https://doi.org/10.2140/pjm.1963.13.421 on Pithdoi:10.2140/pjm.1963.13.421
- [6] C. W. H. Lam, L. Thiel, and S. Swiercz.The non-existence of finite projective planes of order 10. Canadian Journal of Mathematics 41 (1989), 1117–1123.https://doi.org/10.4153/CJM-1989-049 -4 on Pithdoi:10.4153/cjm-1989-049
- [7] W. D. Wallis.Three orthogonal Latin squares. Congressus Numerantium 42 (1984), 69–86.https: //zbmath.org/?q=an:0546.05011 on Pith
- [8] D. T. Todorov.Three mutually orthogonal Latin squares of order 14. Ars Combinatoria 20 (1985), 45–47.https://zbmath.org/?q=an:0596.05009 on Pith
- [9] A. Miller, R. J. R. Abel, I. Valkov, and D. Fraser.Implementing the MOLS table forn up to 500. Symmetry 16 (2024), 1678.https://doi.org/10.3390/sym16121678 on Pithdoi:10.3390/sym16121678
- [10] M. Zhang.The endless exam: Mathematical constructions from today’s models toward superintelligence. arXiv:2609.24555, 2026.https://arxiv.org/abs/2609.24555 on PithPith's reviewarXiv
- [11] Y. Luo, Z. Liu, Y. Yuan, Z. Yang, and A. C.-C. Yao.TCSExam: Evaluating frontier LLMs on the full range of TCS research. arXiv:2611.07962, 2026.https://arxiv.org/abs/2611.07962 on PithPith's reviewarXiv
- [12] C. Bright, A. Keita, and B. Stevens.Orthogonal Latin squares of order 10 with two relations: A SAT investigation. Discrete Mathematics, Algorithms and Applications (2025), Article 2550156. https://doi.org/10.1142/S1793830925501563 on Pithdoi:10.1142/s1793830925501563
- [13] A. Barnoff and C. Bright.Improving SAT solvers on orthogonal Latin square problems. arXiv:2605.02132, 2026.https://arxiv.org/abs/2605.02132 on PithPith's reviewarXiv
- [14] C. Bright, A. Keita, and B. Stevens.Myrvold’s results on orthogonal triples of10× 10Latin squares: A SAT investigation. Electronic Journal of Combinatorics 33 (2026), #P1.30.https://doi.org/10 .37236/13960. 31 Improved Lower Bounds for Mutually Orthogonal Latin Squares (MOLS) ... on Pith
- [15] R. M. Wilson.Concerning the number of mutually orthogonal Latin squares. Discrete Mathematics 9 (1974), 181–198.https://doi.org/10.1016/0012-365X(74)90148-4 on Pithdoi:10.1016/0012-365x(74)90148-4
- [16] A. E. Brouwer.Parameters of strongly regular graphs: 451–500 vertices.https://aeb.win.tue.nl/g raphs/srg/srgtab451-500.html. Accessed November 11, 2026 on Pith
- [17] A. E. Brouwer and H. Van Maldeghem.Strongly Regular Graphs. Encyclopedia of Mathematics and its Applications 182. Cambridge University Press, Cambridge, 2022.https://doi.org/10.1017/97 81009057226 on Pithdoi:10.1017/97
- [18] OpenAI.GPT-6 Astra: A new generation of intelligence. September 3, 2026.https://openai.com/i ndex/gpt-6-astra/ on Pith
- [19] OpenAI.Codex.https://openai.com/codex/. Accessed November 11, 2026 on Pith
- [20] Anthropic.Introducing Claude Opus 5.5. September 22, 2026.https://www.anthropic.com/claude -opus-5-5 on Pith
show 26 more
- [21] R. J. R. Abel and Y. W. Cheng.Some new MOLS of order2np for p a prime power. Australasian Journal of Combinatorics 10 (1994), 175–186.https://ajc.maths.uq.edu.au/pdf/10/ocr-ajc-v 10-p175.pdf on Pith
- [22] K. van Greevenbroek and J. Jedwab.A new structure for difference matrices over abelianp-groups. arXiv:1806.04279, 2018.https://arxiv.org/abs/1806.04279 on PithPith's reviewarXiv
- [23] R. J. R. Abel.Existence of five MOLS of orders 18 and 60. Journal of Combinatorial Designs 23(4) (2015), 135–139.https://doi.org/10.1002/jcd.21384 on Pithdoi:10.1002/jcd.21384
- [24] D. T. Todorov.Four mutually orthogonal Latin squares of order 20. Ars Combinatoria 27 (1989), 63–65 on Pith
- [25] R. J. R. Abel and D. T. Todorov.Four MOLS of orders 20, 30, 38, and 44. Journal of Combinatorial Theory, Series A 64(1) (1993), 144–148.https://doi.org/10.1016/0097-3165(93)90093-N on Pithdoi:10.1016/0097-3165(93)90093-n
- [26] R. J. R. Abel.Some V (12,t )vectors and designs from difference and quasi-difference matrices. Australasian Journal of Combinatorics 40 (2008), 69–85.https://ajc.maths.uq.edu.au/pdf/40/a jc_v40_p069.pdf on Pith
- [27] D. T. Todorov.Four mutually orthogonal Latin squares of order 14. Journal of Combinatorial Designs 20 (2012), 363–367.https://doi.org/10.1002/jcd.21298 on Pithdoi:10.1002/jcd.21298
- [28] R. J. R. Abel and Y. Li.Some constructions fort pairwise orthogonal diagonal Latin squares based on difference matrices. Discrete Mathematics 338 (2015), 593–607.https://doi.org/10.1016/j.disc .2014.11.018 on Pithdoi:10.1016/j.disc
- [29] W. F. Kuhfeld.Difference schemes via computerized searches. Journal of Statistical Planning and Inference 127 (2005), 341–346.https://doi.org/10.1016/j.jspi.2003.09.021 on Pithdoi:10.1016/j.jspi.2003.09.021
- [30] The SageMath Developers.Combinatorial designs database.sage/combinat/designs/database.py, entry MOLS_15_4. https://github.com/sagemath/sage/blob/develop/src/sage/combinat/desi gns/database.py on Pith
- [31] N. J. A. Sloane.A Library of Orthogonal Arrays, array oa.36.13.3.2.https://neilsloane.com/oad ir/oa.36.13.3.2.txt on Pith
- [32] M. Bailey, C. del Valle, and P. J. Dukes.A lower bound on HMOLS with equal sized holes. arXiv:2008.08788, 2020.https://arxiv.org/abs/2008.08788 on PithPith's reviewarXiv
- [33] P. H. J. Lampio.Classification of difference matrices and complex Hadamard matrices. Aalto University doctoral dissertation 177/2015, Section 6.3.1. Electronic classification supplement with P. R. J. Östergård,Matrices over Z7, matrix #2 of order 14.https://wiki.aalto.fi/space... on Pith
- [34] C. Luo, Y. Zhang, and X. Chen.Orthogonal arrays constructed by generalized Kronecker product. Journal of Applied Analysis and Computation 7(2) (2017), 728–744.https://doi.org/10.11948/2 017046 on Pithdoi:10.11948/2
- [35] R. E. A. C. Paley.On Orthogonal Matrices. Journal of Mathematics and Physics 12 (1933), 311–320. https://doi.org/10.1002/sapm1933121311 on Pithdoi:10.1002/sapm1933121311
- [36] A. A. Albert.Generalized twisted fields. Pacific Journal of Mathematics 11 (1961), 1–8. https: //doi.org/10.2140/pjm.1961.11.1 on Pithdoi:10.2140/pjm.1961.11.1
- [37] J. André.Über nicht-Desarguessche Ebenen mit transitiver Translationsgruppe. Mathematische Zeitschrift 60 (1954), 156–186.https://doi.org/10.1007/BF01187370 on Pithdoi:10.1007/bf01187370
- [38] J. M. Dover.Computational Enumeration of André Planes. arXiv:2105.07439, 2021.https://arxiv. org/abs/2105.07439. 32 Improved Lower Bounds for Mutually Orthogonal Latin Squares (MOLS) at 33 Orders on PithPith's reviewarXiv
- [39] A. E. Brouwer.A series of separable designs with application to pairwise orthogonal Latin squares. European Journal of Combinatorics 1 (1980), 39–41.https://doi.org/10.1016/S0195-6698(80)8 0019-9 on Pithdoi:10.1016/s0195-6698(80)8
- [40] M. Greig.Designs from projective planes and PBD bases. Journal of Combinatorial Designs 7 (1999), 341–374.https://doi.org/10.1002/(SICI)1520-6610(1999)7:5<341::AID-JCD5>3.0.CO;2-1 on Pithdoi:10.1002/(sici)1520-6610(1999)7:5
- [41] R. C. Bose, S. S. Shrikhande, and E. T. Parker.Further results on the construction of mutually orthogonal Latin squares and the falsity of Euler’s conjecture. Canadian Journal of Mathematics 12 (1960), 189–203.https://doi.org/10.4153/CJM-1960-016-5 on Pithdoi:10.4153/cjm-1960-016-5
- [42] C. J. Colbourn and J. H. Dinitz.Mutually orthogonal Latin squares: a brief survey of constructions. Journal of Statistical Planning and Inference 95 (2001), 9–48.https://doi.org/10.1016/S0378-3 758(00)00276-7 on Pithdoi:10.1016/s0378-3
- [43] W. D. Wallis and L. Zhu.Orthogonal Latin squares with small subsquares, in L. R. A. Casse (ed.), Combinatorial Mathematics X, Lecture Notes in Mathematics 1036, Springer, 1983, pp. 398–409. https://doi.org/10.1007/BFb0071533 on Pithdoi:10.1007/bfb0071533
- [44] L. Zhu.Six pairwise orthogonal Latin squares of order 69. Journal of the Australian Mathematical Society (Series A) 37 (1984), 1–3.https://doi.org/10.1017/S1446788700021698 on Pithdoi:10.1017/s1446788700021698
- [45] R. H. F. Denniston.Some maximal arcs in finite projective planes. Journal of Combinatorial Theory 6 (1969), 317–319.https://doi.org/10.1016/S0021-9800(69)80095-5 on Pithdoi:10.1016/s0021-9800(69)80095-5
- [46] M. Greig.Some pairwise balanced designs. Electronic Journal of Combinatorics 7 (2000), Research Paper 13.https://doi.org/10.37236/1491. 33 Improved Lower Bounds for Mutually Orthogonal Latin Squares (MOLS) at 33 Orders A Prompts GPT-6 Astra was used in two modes: (a) in ChatGP... on Pithdoi:10.37236/1491
Challenges and discussion
Disagree with the review, or know something it missed? Leave a comment, or challenge this review formally: a challenge becomes a permanent entry in this paper's record, and its decision is published here.
Come back by email
A short list of the week's newly reviewed papers. Nothing is sent until you confirm from your inbox, and every email has a link that stops it. No account.
No comments yet.
Sign in to comment