{"id":"c7c3ee60-76c4-425f-a948-416e4c2295bd","arxiv_id":"2508.07404","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The Lefschetz homomorphism from endotrivial complexes to orthogonal units of the trivial source ring is surjective for several families of finite groups (2-fusion controlled or dihedral Sylow 2-subgroups for p=2; cyclic Sylow or p-nilpotent for odd p), but not surjective for some groups of p-rank at","lead":"This paper studies a map that turns certain chain complexes of modular representations into units of an algebraic ring, and asks when every such unit comes from a complex. For characteristic 2 it answers yes for groups with dihedral or fusion-controlled Sylow 2-subgroups, and for odd primes it gives both new yes-answers and new no-answers.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 5.10's general-rank counterexample is unproved for n >= p: the proof counts only cyclic p-subgroups inside C_p^n, but once S_n has p-torsion, c(G) exceeds n+1 and the generator bound fails.","rationale":"The reader's stated weakest assumption was the external classification Theorem 2.13, but the reader's rationale also flagged the potential gap in Theorem 5.10 for n >= p. My own reading isolates that gap as the most load-bearing concrete issue in the paper's central claim. The surjectivity theorems for cyclic/dihedral/nilpotent cases appear to rest on the quoted classification and on internal computations that I did not find a specific flaw in. The rank-2 example in Proposition 5.9 is convincing modulo the same external classification. The problem is the passage from rank 2 to arbitrary rank: the group chosen in Theorem 5.10 has extra p-local structure coming from S_n when n >= p, so the proof's lower bound on generators of R_G does not meet the threshold required by Observation 5.3. This is not an ad hominem or a disagreement with consensus; it is a concrete internal gap in one of the paper's advertised main theorems. The paper should either restrict the theorem to n < p, find a different family of examples with elementary abelian Sylow p-subgroups for all ranks, or supply a completed argument for n >= p. The reader's CONDITIONAL verdict remains appropriate: the main positive theorems may well be correct, but the stated universality of the counterexample claim is not established as written.","tokens_in":25253,"tokens_out":9058,"duration_ms":100189,"concrete_test":"Work out the case p=3, n=3 for G=(C_3⋊C_2)^3⋊S_3. Compute (i) the number c(G) of conjugacy classes of cyclic 3-subgroups and (ii) the minimum number of generators of the reduced coherent character tuple group R_G. The element (a,a^{-1},1;(1 2 3)) already shows c(G) >= 5, while the proof's method gives only 4 generators for R_G. If indeed min.gen.(R_G) < c(G), then Theorem 5.10's inference fails and the claim needs a genuinely new construction or a restriction to n < p.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The paper claims in the introduction that for every odd prime p and every integer i >= 2 there are finite groups with elementary abelian Sylow p-subgroups of rank i for which Lambda is not surjective. Proposition 5.9 proves the rank-2 case, and Theorem 5.10 claims the general case by taking G = (C_p ⋊ C_{p-1})^n ⋊ S_n. For n < p this works: S_n is a p'-group, the Sylow p-subgroup is C_p^n, and c(G)=n+1 (including the trivial subgroup), so the n+1 generators exhibited for R_G are enough to apply Observation 5.3. For n >= p, however, S_n has nontrivial p-torsion, so E = C_p^n is not Sylow and the Sylow p-subgroup is a nonabelian extension of E by a Sylow p-subgroup of S_n. There are cyclic p-subgroups not contained in E, for example in the p=3, n=3 case the element (a,a^{-1},1;(1 2 3)) has order 3 and lies outside E. Hence c(G) is strictly larger than n+1. The proof only establishes that R_G has at least n+1 generators, so Observation 5.3 cannot be applied. Moreover, the construction no longer has elementary abelian Sylow p-subgroups, contradicting the introduction's stated theorem. The p-rank may still be n, but the proof of non-surjectivity of Lambda is incomplete for n >= p.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the Lefschetz homomorphism Λ from the group of endotrivial complexes E_k(G) to the orthogonal unit group O(T(kG)) of the trivial source ring. The main results are: for p=2 and k=F_2, Λ is surjective whenever a Sylow 2-subgroup has fusion controlled by its normalizer (Theorem 4.5) and whenever the Sylow 2-subgroup is dihedral (Theorem 4.8 and Corollary 4.10); for odd p, Λ is surjective for p-nilpotent groups (Theorem 5.1) and for groups with cyclic Sylow p-subgroups (Theorem 5.8). The paper also claims counterexamples to surjectivity for p-rank at least 2 (Proposition 5.9 and Theorem 5.10) and gives partial results on ker Λ (Theorems 6.1 and 6.2). The arguments combine the author's prior classification of endotrivial complexes with Boltje-Carman's description of the orthogonal unit group, and in the p=2 case rely on Tornehave/Yalçın surjectivity for p-groups together with explicit bases for the dihedral cases.","tokens_in":25601,"tokens_out":14632,"duration_ms":151966,"significance":"If the stated results hold, this is a substantial contribution to the comparison between splendid Rickard equivalences and p-permutation equivalences: the surjectivity results give large families where every orthogonal unit is the Euler characteristic of an endotrivial complex, while the proposed counterexamples would show that the two families of equivalences genuinely diverge. The p=2 theorems are clean and the cyclic-Sylow odd-p theorem is nontrivial. The paper is honest that the classification theorem [Mil25a] is an external load-bearing input, and the kernel results are clearly presented as partial. However, the general-rank counterexample theorem, which is advertised in the introduction, has a serious gap for n ≥ p, so the scope of the paper's central claim is not currently established.","major_comments":[{"comment":"The proof of Theorem 5.10 is incomplete for n ≥ p. The proof itself notes that E ≅ C_p^n is 'a p-subgroup (not necessarily Sylow)'. When n ≥ p, S_n has p-torsion, so the Sylow p-subgroup of G = (C_p ⋊ C_{p-1})^n ⋊ S_n is strictly larger than E, and there are cyclic p-subgroups not conjugate to subgroups of E; for p=3, n=3, the element ((a,a^{-1},1),(1 2 3)) has order 3 and lies outside E. Hence c(G) > n+1. The proof only establishes that R_G has at least n+1 generators, whereas Observation 5.3 requires at least c(G) generators. Thus non-surjectivity of Λ is not established for n ≥ p. Moreover, for n ≥ p the Sylow p-subgroup is not elementary abelian, contradicting the introduction's claim of groups with elementary abelian Sylow p-subgroups of rank i for every i ≥ 2. The theorem must be repaired, or the claim restricted to n < p with the introduction amended accordingly.","section":"§5.2, Theorem 5.10"},{"comment":"The reduction step asserts that restriction induces an isomorphism T E(G) ≅ T E(N_G(C)) and adds the parenthetical 'this holds for any group containing S'. This is false in general: if H contains a Sylow p-subgroup S, then T E(H) has rank c(H), and G-conjugacy can be coarser than H-conjugacy, so restriction need not be bijective. In the specific case H = N_G(C), where C is the unique subgroup of order p in a cyclic Sylow p-subgroup, Burnside's fusion theorem does give equality of conjugacy classes of cyclic p-subgroups, so the claim is probably true; but it needs to be stated and proved. Please replace the parenthetical with the specific fusion argument.","section":"§5.1, proof of Theorem 5.8"}],"minor_comments":[{"comment":"'Λ is injective' cannot be right. The paper's own Theorem 6.2 shows that for p=2 a nonzero endotrivial complex with even h-marks lies in ker Λ (e.g. h=2f for a nonzero Borel-Smith function), so Λ is not injective even for resistant Sylow 2-subgroups. The sentence presumably should say 'surjective', already proved in Theorem 4.5.","section":"Remark 4.6"},{"comment":"The notation mixes G and G′: the trace should be over N_{G′}(S) (or the common fusion system), not N_G(S). Please rewrite the last sentence of the proof.","section":"Theorem 4.5, final sentence"},{"comment":"'Let p be a prime and let G be a p-group with cyclic Sylow p-subgroup S' should read 'finite group' (and similarly in the following sentence).","section":"Proposition 5.6"},{"comment":"The phrase 'NG(P)/CG(P) is a P-group' should be 'a p-group'; using 'P' for both a subgroup and a property is confusing.","section":"Theorem 5.1, proof"},{"comment":"The headers in Table 3 repeat K1 K2, and the bases for the fusion systems F^I/F^II would be easier to check if the conventions for H1, H2, K1, K2 were repeated in the table captions. Please clarify.","section":"Tables 3 and 4"}],"recommendation":"major_revision","confidential_remarks":"The paper is a serious contribution and most of the p=2 and odd-p-cyclic results appear sound. My main concern is Theorem 5.10, which as written does not prove the advertised general-rank counterexamples. I do not recommend rejection: the flaw is local and an alternative or restricted construction may suffice, but the introduction's sweeping claim must be reconciled with what is actually proved."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a serious paper with real new results, but the general non-surjectivity theorem (5.10) has a gap that the abstract and introduction overstate. The p=2 surjectivity results for fusion-controlled and dihedral Sylow 2-subgroups look well-proved, and the odd-p cyclic Sylow / p-nilpotent surjectivity is credible and carefully built on Miller's earlier classification. If I were refereeing, I'd accept with major revision on the counterexample section.\n\nWhat's genuinely new: Theorems 4.5 and 4.10 give the first surjectivity of the Lefschetz map over F2 for non-p-groups in these classes, using Tornehave–Yalcin plus a transfer argument that appears sound. The odd-p cases (cyclic Sylow, p-nilpotent) are new, and the kernel computation in Section 6 is a useful complement. The heavy reliance on [Mil25a] is legitimate here—this is a continuation, not a circular argument.\n\nThe soft spot is Theorem 5.10. The construction G = (C_p ⋊ C_{p-1})^n ⋊ S_n only has an elementary abelian Sylow p-subgroup when n < p. For n ≥ p, S_n has p-torsion, so C_p^n is not Sylow and there are cyclic p-subgroups outside it. The proof bounds the number of generators of R_G by n+1, but c(G) is then larger, so Observation 5.3 cannot be applied. The paper acknowledges E is 'not necessarily Sylow' but the counting argument only works if it is. The introduction's claim of elementary abelian Sylow p-subgroups of rank i for every i ≥ 2 is therefore unsupported for i ≥ p. The rank-2 example (Proposition 5.9) is fine, so the existence of a non-surjectivity example stands—but the arbitrarily large rank statement is unproved.\n\nMinor: Remark 4.6 says 'injective' where it should say 'surjective'. That's a typo, not a deep issue. The proof of Theorem 5.8 is intricate and I didn't find an obvious flaw.\n\nBottom line: worth a serious referee. The main surjectivity theorems are strong and likely correct; the non-surjectivity section needs a different construction for n ≥ p or a corrected statement. The reader's conditional verdict is right.","headline":"Solid new surjectivity theorems for p=2 and odd p, but the advertised rank-i non-surjectivity examples are only established for rank 2—Theorem 5.10 breaks for n ≥ p.","tokens_in":26135,"tokens_out":2236,"would_cite":true,"duration_ms":20941,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["20J05","19A22","20C05","20C20"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves the Lefschetz homomorphism is surjective for $\\mathbb{F}_2$ when Sylow 2-fusion is normalizer-controlled or dihedral, for odd p when the Sylow p-subgroup is cyclic or the group is p-nilpotent, and gives explicit p-rank $\\ge","keywords":["endotrivial complexes","Lefschetz homomorphism","Euler characteristic","trivial source ring","orthogonal units","Borel-Smith functions","p-permutation modules","surjectivity"],"falsifier":"For $G = N_{S_{2p}}(S)$ with $S \\cong C_p \\times C_p$ and $p$ odd, take the coherent character tuple $\\chi$ that is trivial at every p-subgroup except $S$, where $\\chi_S(x) = -1$ for $x \\in N_G(S) \\setminus N_G(A)$ and $\\chi_S(x) = 1$ otherwise, with $A$ one of the two factors of $S$. The paper's Proposition 5.9 asserts this tuple is not in the image of $\\Lambda$, because the tuple group $R_G$ needs at least three generators while the endotrivial quotient has two. Finding an endotrivial complex $C$ whose h-marks and characters realize $\\chi$, or a direct proof that $R_G$ is generated by two el","tokens_in":25111,"feed_emoji":"🧮","tokens_out":16040,"duration_ms":139368,"temperature":0.7,"pith_summary":"This paper studies the Lefschetz homomorphism $\\Lambda: \\mathcal{E}_k(G) \\to O(T(kG))$ from the group of endotrivial complexes to the orthogonal units of the trivial source ring, and asks when it is surjective. It proves surjectivity over $\\mathbb{F}_2$ for groups whose Sylow 2-subgroup has fusion controlled by its normalizer and for groups with dihedral Sylow 2-subgroups, and over odd p for p-nilpotent groups and groups with cyclic Sylow p-subgroups. It then exhibits, for every odd p and every $n \\geq 2$, groups of p-rank n for which $\\Lambda$ is not surjective, so no general surjectivity theorem can hold for odd p. The paper also computes the kernel of $\\Lambda$ completely for p = 2 and for odd p when the Sylow p-subgroup is cyclic. If these surjectivity results are right, every orthogonal unit in those cases is realized as the Euler characteristic of an actual endotrivial complex, tying algebraic invariants of the trivial source ring to homotopy-theoretic objects.","feed_headline":"Euler characteristic map is surjective in char 2, not always in odd p","feed_subtitle":"Cyclic Sylow and p-nilpotent groups still work; the kernel is now known in char 2 and for cyclic Sylow p-subgroups.","key_machinery":"The load-bearing objects are the h-mark homomorphism $h: \\mathcal{E}_k(G) \\to CF(G,p)$ and the classification theorem that identifies its image with the group $CF^b(G,p)$ of Borel-Smith functions, up to the torsion subgroup $\\mathrm{Hom}(G,k^\\times)$. A Borel-Smith function is a superclass function on p-subgroups satisfying parity conditions and a rank-two additivity relation; it records, for each p-subgroup P, the degree in which the Brauer construction of the complex has nonzero homology. On the target side, the decomposition $O(T(kG)) \\cong (B(S)^G)^\\times \\times (\\prod_{P} \\mathrm{Hom}(N_G(P)/P, k^\\times))'$ separates the problem: for p = 2 surjectivity of $\\Lambda$ is equivalent to surj","core_discovery":"The central claim is that the Lefschetz homomorphism $\\Lambda: \\mathcal{E}_k(G) \\to O(T(kG))$ is governed by the p-local structure of G and is surjective in precisely the cases it lists. Over $k = \\mathbb{F}_2$, Theorem 4.5 proves surjectivity whenever $N_G(S)$ controls fusion in a Sylow 2-subgroup S, and Corollary 4.10 proves it whenever S is dihedral; this covers all groups with abelian or resistant Sylow 2-subgroups, including $A_5$, $A_6$, $A_7$, and $PSL_2(q)$ for odd prime powers q. For odd p, Theorem 5.1 proves surjectivity for p-nilpotent groups and Theorem 5.8 for groups with cyclic Sylow p-subgroups, using the periodicity $2\\Phi(S)$. Proposition 5.9 and Theorem 5.10 then construct,","pith_inferences":["The paper leaves open whether $\\Lambda$ is surjective for every group over $\\mathbb{F}_2$; if it is, then over $\\mathbb{F}_2$ every p-permutation autoequivalence of the trivial source ring would be induced by a splendid Rickard autoequivalence, since orthogonal units are exactly what induce p-permutation equivalences. That implication goes beyond the paper's theorems.","The odd-p rank-counting criterion suggests a cheap way to search for more failures: compute the minimum number of generators of the reduced coherent character tuple group $R_G$ and compare it with the number of conjugacy classes of cyclic p-subgroups minus one. Applying this to families with quaternion or semidihedral Sylow subgroups would test the paper's expectation that most p-rank 2 groups fai","The kernel for odd p would be completely determined if one knew whether $H_C(P)=1$ for all cyclic p-subgroups P forces $H_C \\equiv 1$; a small computer search over groups with non-cyclic Sylow p-subgroups could settle that open question, and a positive answer would complete the kernel description."],"forward_implications":["Over $\\mathbb{F}_2$, every orthogonal unit of the trivial source ring is the Euler characteristic of an endotrivial complex whenever the Sylow 2-fusion is normalizer-controlled or the Sylow 2-subgroup is dihedral; this includes all abelian and resistant Sylow 2-subgroups, and covers $A_5$, $A_6$, $A_7$, and $PSL_2(q)$ for odd q.","For odd p, all orthogonal units are realized for p-nilpotent groups and for groups with cyclic Sylow p-subgroups, so in these families the surjectivity question is closed.","For every odd p and every $n \\geq 2$ there are groups of p-rank n whose Lefschetz homomorphism is not surjective, so a classification of surjectivity for odd p must involve restrictions stronger than rank or Sylow shape alone.","An endotrivial complex over $\\mathbb{F}_2$ with trivial homology has zero Euler characteristic exactly when its h-mark function is even-valued, giving a complete kernel description; for odd p with cyclic Sylow subgroups the analogous criterion is congruence modulo $2\\Phi(S)$."],"supporting_citations":[{"why":"supplies the classification of endotrivial complexes: the h-mark map has image CF^b(G,p) and kernel Hom(G,k^x); all rank, basis, and kernel computations assume it.","marker":"[Mil25a]"},{"why":"gives the decomposition of O(T(kG)) into a Burnside-ring unit group and coherent character tuples, which is the target of the Lefschetz homomorphism.","marker":"[BC23]"},{"why":"defines endotrivial complexes, proves the Euler characteristic is an orthogonal unit, and describes Lambda through h-marks and character tuples.","marker":"[Mil24a]"},{"why":"the p-group surjectivity theorem for the dimension homomorphism; it is the base case for the p=2 results.","marker":"[Tor84]"},{"why":"gives an algebraic induction theorem for unit groups of Burnside rings of 2-groups, used to identify B(S)^x.","marker":"[Yal05]"},{"why":"classifies groups with dihedral Sylow 2-subgroups, reducing the dihedral surjectivity proof to three fusion systems.","marker":"[GW65]"},{"why":"foundational source for Burnside rings, Borel-Smith functions, and the dimension homomorphism.","marker":"[tD87]"},{"why":"groups with cyclic Sylow p-subgroups have periodic cohomology with period 2|Aut_G(S)|, used in the odd-p surjectivity induction and kernel congruences.","marker":"[Swa59]"}],"fun_headline_variants":["Lefschetz map surjective in char 2, not always in odd p","Euler char map: char 2 wins, odd p has exceptions","Sylow structure controls Lefschetz map surjectivity","Char 2 surjectivity, cyclic Sylow works for odd p, but not all"],"cache_read_input_tokens":2816,"weakest_assumption_plain":"All of the paper's rank, surjectivity, and kernel statements assume the previously proved classification of endotrivial complexes: that the h-mark map realizes exactly the Borel-Smith functions, with kernel the one-dimensional characters. If that classification had a counterexample, the whole argument would lose its foundation.","fun_headline_variants_meta":{"raw":{"variants":["Lefschetz map surjective in char 2, not always in odd p","Euler char map: char 2 wins, odd p has exceptions","Sylow structure controls Lefschetz map surjectivity","Char 2 surjectivity, cyclic Sylow works for odd p, but not all"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000303,"raw_usage":{"total_tokens":1640,"prompt_tokens":867,"completion_tokens":773,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":611,"completion_tokens_details":{"reasoning_tokens":689}},"tokens_in":611,"tokens_out":773,"duration_ms":7526,"temperature":1.0,"reasoning_tokens":689,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T22:10:22.348648+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For $G = N_{S_{2p}}(S)$ with $S \\cong C_p \\times C_p$ and $p$ odd, take the coherent character tuple $\\chi$ that is trivial at every p-subgroup except $S$, where $\\chi_S(x) = -1$ for $x \\in N_G(S) \\setminus N_G(A)$ and $\\chi_S(x) = 1$ otherwise, with $A$ one of the two factors of $S$. The paper's Proposition 5.9 asserts this tuple is not in the image of $\\Lambda$, because the tuple group $R_G$ needs at least three generators while the endotrivial quotient has two. Finding an endotrivial complex $C$ whose h-marks and characters realize $\\chi$, or a direct proof that $R_G$ is generated by two el","supporting_citations":[],"review_version":1}