{"id":"fdeca8a0-1069-4c9c-baee-2258b6ac3ba9","arxiv_id":"2509.02882","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"For rational product Cantor sets in R^d (d≥2) with cyclotomic conditions on digit sets, the Favard length of N^{-1}-neighborhoods decays as N^{-ε}, new for d≥3 and for fibered digit sets.","lead":"This math paper proves that the average projected length (Favard length) of thin neighborhoods of certain self-similar fractal sets in any dimension decays at least like a power law, extending a known planar result to higher dimensions and to a new family of digit sets. The result gives the first non-trivial upper bounds of this kind in dimensions three and higher.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The proof of the main theorem contains an invalid final contradiction: choosing ε0>1/α² makes N^{1-α√ε0}=N^{-c}, which is compatible with m≈log N; the SSV-case exponent α in (4.16) also misses a factor 2.","rationale":"The reader's verdict identifies the dependence on the unpublished preprint [17] and the missing explicit exponent as the main weakness. That is a legitimate concern for the fibered-subset extension, but it does not affect Theorem 1.2 for the 2^d-corner Cantor set, which only needs #Ai≤10. A more load-bearing issue is that the proof of the main theorem itself contains an invalid final contradiction: the sign of the exponent in the choice of ε0 is reversed, so the displayed argument does not actually produce the required contradiction. There is also a missing factor of 2 in the SSV-case exponent α, which undermines Proposition 3.20's claimed quantitative dependency. These are internal inconsistencies, not disagreements with consensus, and they sit at the heart of every Favard upper bound in the paper. I do not see a reason to reject the theorem outright: the errors appear to be repairable by choosing ε0 < 1/α² and setting α = 2(c1+C1) in the SSV case, and the overall strategy is coherent. But as written the central proof is not complete, so the conditional verdict is appropriate. I partially agree with the reader because the unpublished-reference concern is real but secondary; the more urgent check is the proof's own algebra.","tokens_in":42502,"tokens_out":10192,"duration_ms":114495,"concrete_test":"Re-derive the contradiction step with explicit algebra: from (3.16) and (3.17), cancel K and L^n to obtain the exact inequality in N and m. Then test both possible choices: (a) ε0 > 1/α² gives N^{-c} ≲ m, which is true for all large N; (b) ε0 < 1/α² gives N^{c} ≲ m, which is false for large N. Also recompute (4.16) using L^{-2C1m}ψ(m)² = N^{-2(C1+c1)√ε0} and check whether α = 2(c1+C1) makes the lower bound match. If the corrected '<' choice and corrected α restore a genuine contradiction, the main theorem is salvageable; if not, the central proof has a deeper gap.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The mechanism that proves every Favard upper bound in the paper is the contradiction at the end of the proof of Theorem 2.4. After comparing (3.16) and (3.17), the author obtains\n\n  N^{1-α√ε0} ≲ m.\n\nThe text then says: 'we choose ε0 > 1/α² so that N^{α'} ≲ log N, for some α'>0. This is clearly false.' But if ε0 > 1/α², then α√ε0 > 1, so 1-α√ε0 < 0, and the left-hand side is N^{-c} with c>0. The inequality N^{-c} ≲ m is true for large N, not contradictory. To obtain a contradiction one needs 1-α√ε0 > 0, i.e. ε0 < 1/α²; then the left side is N^{+c}, which cannot be bounded by m ≈ log N. As written, the proof has the sign reversed and does not rule out H¹(E_{N,K}) > K^{-β}.\n\nThere is a second algebraic slip in the same chain. In Proposition 4.5, the SSV case gives L^{-2C1m}ψ(m)² = L^{-2C1m}L^{-2c1m} = N^{-2(C1+c1)√ε0}, so the required α is 2(c1+C1), not c1+C1 as stated in (3.18). This does not destroy the existence argument if corrected, but it invalidates the quantitative claim in Proposition 3.20 as written. Since the final contradiction is the load-bearing step for all of Theorem 2.4/2.16, the central proof is not rigorous as it stands, even though the error appears repairable.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a higher-dimensional version of the Nazarov–Peres–Volberg / Bond–Laba–Volberg method for upper bounds on the Favard length of small neighbourhoods of self-similar sets. It claims power-law decay for the 2^d-corner Cantor set in R^d (Theorem 1.2) and, more generally, for rational product Cantor sets whose digit sets satisfy one of three conditions: cardinality at most 10, accumulating cyclotomic divisors with at most two prime factors, or existence of a fibered subset (Theorems 2.4, 2.10, 2.16). The proof introduces a counting function, a set of low-multiplicity directions E_{N,K}, and reduces the desired Favard bound to estimating H^1(E_{N,K}). That estimate is obtained by comparing upper and lower bounds for a Riesz product over an almost-unit interval, using SSV and SLV properties. The fibered case relies on a lower bound for mask polynomials from the author's unpublished preprint [17].","tokens_in":43054,"tokens_out":9889,"duration_ms":106308,"significance":"If correct, the main theorem is a genuine advance: it provides the first non-trivial Favard-length upper bound for purely 1-unrectifiable sets in R^d, d≥3, whose H^1 measure is not concentrated on a two-dimensional affine subspace. The paper also introduces a useful modular framework, separating the SSV/SLV estimates from the combinatorial counting arguments, and it makes explicit the dependence of the final exponent on the SSV and SLV constants. The new fibered digit-set class is potentially interesting even in the plane. However, the proof of the central contradiction currently contains a sign error and an incorrect exponent, and the new fibered case rests on an unpublished preprint. The result is likely repairable, but the manuscript as written is not rigorous.","major_comments":[{"comment":"The final contradiction has the sign reversed. After comparing (3.16) and (3.17), the author obtains N^{1-α√ε0} ≲ m and then chooses ε0 > 1/α², claiming that N^{α'} ≲ log N for some α'>0. But if ε0 > 1/α², then 1-α√ε0 < 0, so the left-hand side is N^{-c} with c>0, and the inequality N^{-c} ≲ m is true for all large N. A contradiction requires 1-α√ε0 > 0, i.e. ε0 < 1/α². This is not cosmetic: it is the exact step that rules out H^1(E_{N,K}) > K^{-β}, and therefore Theorems 2.4 and 2.16 are not proved as written. The argument is repairable by choosing ε0 < 1/α², but the stated proof is invalid.","section":"§3.4, proof of Theorem 2.4, Eqs. (3.16)–(3.19)"},{"comment":"The explicit value of α is wrong by a factor of 2 in the SSV case. In the SSV branch of the proof of Proposition 4.5, ψ(m)=L^{-c1m}, so L^{-2C1m}ψ(m)^2 = L^{-2(C1+c1)m} = N^{-2(C1+c1)√ε0}. Therefore the correct α is 2(c1+C1), not c1+C1 as stated in (3.18) and (4.16). This error affects the allowable range of ε0 and, together with the sign error above, invalidates the quantitative characterization in Proposition 3.20 and the claimed recovery of an ε>1/4 exponent in Remark 1.3 and §3.5. The existence argument can survive after redefining α, but the quantitative claims as written are unsupported.","section":"§4.3, Prop. 4.5, Eq. (4.16); §3.4, Prop. 3.19, Eq. (3.18)"},{"comment":"The new fibered-digit-set case is conditional on Proposition 2.11, a size estimate #A ≥ min_σ FIB(S_A^{(2)},σ) taken verbatim from the author's unpublished preprint [17]. The construction of the single-scale SLV set in Section 5.2 depends directly on this bound, including the strict inequality (5.15) used to match the λ parameters. Because [17] is not available in the manuscript and the result is not proved or even stated in sufficient detail here, the referee cannot verify the new theorem for fibered sets. This does not affect Theorem 1.2, which uses the #A_i≤10 condition, but it does mean Theorem 2.10 and the fibered part of Theorem 2.16 are not independently verifiable. Please either include a proof of Proposition 2.11 or state clearly that the result is conditional on [17] being accepted.","section":"§2.3, Prop. 2.11; §5.2, Lemma 5.2 and Prop. 5.9"}],"minor_comments":[{"comment":"The text says 'whereas, B is an (S_A^{(2)}, σ_p)-fibered set' for the set D; this should read 'whereas D is an (S_A^{(2)}, σ_p)-fibered set'.","section":"§2.3.1, Example 2.15"},{"comment":"The proof switches notation between C1 and C2: after defining δ(s,t) with constants c1,C1, the displayed computation uses (c1+C2)^2. This should be corrected to C1 throughout.","section":"§3.5, proof of Proposition 3.20"},{"comment":"The claimed bound Fav ≲ N^{-1/4+u} appears to have a sign error. With ρ>3 and β<1, the maximum of δ(1,0) is strictly less than 1/4, so the exponent should be of the form 1/4-u (u>0), not 1/4+u. This should be corrected or the claim removed.","section":"§3.5, after Proposition 3.20"}],"recommendation":"major_revision","confidential_remarks":"The central theorem is probably repairable: the sign error is fixed by choosing ε0<1/α², and the missing factor 2 is fixed by resetting α. I therefore recommend major revision rather than rejection. The editor should be aware that the new fibered case depends on the author's own preprint [17]; accepting the paper before that preprint is independently verified would make the new result conditional on unpublished work. The quantitative claims in Proposition 3.20 and Remark 1.3 should be revised or removed, since they are not supported even after the local algebraic corrections."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nYou should look at this paper for the results, not for the proof as printed. The main theorem is genuinely new: for the d-dimensional 2^d-corner Cantor set, and more generally for rational product Cantor sets satisfying one of three digit-set conditions, the paper claims Fav(N_{L^{-N}}(S^∞)) ≤ C N^{-ε(or ε/log log N)}. In d≥3 this is the first upper power law for a purely 1-unrectifiable set whose H^1 is not concentrated on a plane; the fibered digit-set class is new even in the plane. The high-dimensional adaptation of the Bond–Laba–Volberg machinery is done carefully, with the anisotropic rescaling issue handled explicitly.\n\nThe bad news: the proof of the central theorem contains a sign error in the final contradiction. After deriving N^{1-α√ε0} ≲ m, the text chooses ε0 > 1/α². That makes the left side N^{-c}, compatible with m≈log N. The contradiction requires ε0 < 1/α². Since every upper bound in the paper passes through this step, Theorem 2.4 and 2.16 are not proved as written. There's a second slip: in the SSV case, (4.16) gives L^{-2C1m}ψ(m)^2 = N^{-2(C1+c1)√ε0}, so α should be 2(c1+C1), not c1+C1; this invalidates the quantitative claims in Prop 3.20 (which also confusingly uses C2 in place of C1).\n\nBoth errors look easily repairable—the argument structure is standard and the intended choice is ε0<1/α² with α corrected. But until fixed, the paper cannot be cited for the power law. The fibered case also leans on the author's unpublished joint preprint [17] for the cardinality lower bound; that bound is not machine-checked and should be independently verified. The paper is honest about its limitations. It's a solid candidate for peer review conditional on substantial revision.\n\nWho is this for: people working in projection theory and fractal geometry, and anyone tracking quantitative Besicovitch–Federer questions. I'd send it to a serious referee, but I wouldn't cite it yet. Maybe discuss at reading group to see how fixable the sign error is.","headline":"The first Favard-length power law in d≥3 is real and worth refereeing, but the proof as written has a sign error in the load-bearing contradiction and a factor-2 slip in the exponent.","tokens_in":43506,"tokens_out":3841,"would_cite":false,"duration_ms":40609,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["28A78","28A80","11C08","42B05","11B75"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that the Favard length of neighbourhoods of the 2^d-corner Cantor set decays as a power law — the first nontrivial upper bound when d≥3.","keywords":["Favard length","rational product Cantor sets","four-corner Cantor set","cyclotomic polynomials","vanishing sums of roots of unity","Riesz products","purely unrectifiable sets","power law"],"falsifier":"Find a digit set A with a fibered subset whose cardinality is strictly less than min_σ FIB(S^{(2)}_A, σ); that would disprove the paper's key combinatorial input (Proposition 2.11) and collapse Theorem 2.10. Alternatively, numerically measure Fav(N_{2^{-dN}}(K_d^∞)) for d=3 and see whether it decays like N^{-ε} rather than N^{-ε/log log N}.","tokens_in":42442,"feed_emoji":"📐","tokens_out":8091,"duration_ms":80423,"temperature":0.7,"pith_summary":"The paper aims to show that the Favard length—the average length of the orthogonal projections of a set onto lines in every direction—decays as a power of the scale for certain self-similar dusts in R^d, d≥2. For the 2^d-corner Cantor set this yields C^{-1}N^{-1} ≤ Fav(N_{2^{-dN}}(K_d^∞)) ≤ C N^{-ε}, the first nontrivial upper bound of this kind for d≥3. The same machinery covers rational product Cantor sets whose digit sets are small, have accumulating cyclotomic zeroes with at most two prime factors, or contain a fibered subset; the fibered case is new even in the plane. A sympathetic reader should care because the result turns the qualitative Besicovitch-Federer projection theorem into a quantitative rate and shows that the arithmetic of vanishing sums of roots of unity controls a geometric-measure-theoretic quantity.","feed_headline":"Power-law decay proven for Favard length in every dimension","feed_subtitle":"The 2^d-corner Cantor set's projections shrink like N^-ε, a first for purely unrectifiable dust above the plane.","key_machinery":"The load-bearing object is the lacunary Riesz product ∏_{k=0}^{n-1} φ_t(L^k ξ), where φ_t(ξ)=φ_{A_1}(ξ)φ_{A_2}(t_1ξ)⋯φ_{A_d}(t_{d-1}ξ) and each φ_{A_i} is the normalized mask polynomial A_i(e^{2πiξ})/#A_i. The cyclotomic factorization A_i=A_i^{(1)}A_i^{(2)}A_i^{(3)}A_i^{(4)} separates the 'accumulating zeroes' S^{(2)}_{A_i}—roots of unity coprime to #A_i—from the tame factors. The proof needs two quantitative level-set properties: the SSV property bounds the set where the product is small, and the SLV property constructs a large set Γ whose difference set lies where the product is large; the Salem trick then forces the Riesz-product integral to be large, contradicting the L^2 bound on the co","core_discovery":"The central discovery is that the planar Buffon-needle method for the four-corner Cantor set transfers to all dimensions, provided the digit sets' mask polynomials are analyzed through their cyclotomic factorization. The paper's main theorem (Theorem 2.16) states that if each digit set A_i satisfies one of three conditions—#A_i ≤ 10, the lcm of its accumulating zeroes has at most two prime factors, or A_i admits a fibered subset—then the Favard length of the L^{-N}-neighbourhood of the rational product Cantor set S^∞ decays as N^{-ε} when all mask polynomials have only roots of unity, and as N^{-ε/log log N} otherwise. Theorem 1.2 is the special case A_i={0,2^d−1}, whose mask polynomial X^{2","pith_inferences":["A plausible takeaway the paper does not state: the real bottleneck for power laws is not the dimension but the arithmetic inequality #A ≥ min_σ FIB(S^{(2)}_A,σ); any new class of digit sets that verifies this inequality should inherit a Favard power law by the same proof.","Because Proposition 3.20's exponent function is dimension-free and depends only on SSV/SLV constants, numerical experiments on the 8-corner set in R^3 could directly probe whether the attainable exponent is close to the planar 1/4 or degrades with dimension.","The explicit link to Vitushkin's conjecture suggests fibered digit sets could serve as tractable test cases for quantitative analytic-capacity estimates, since the same Riesz-product level sets control both quantities."],"forward_implications":["For the 2^d-corner Cantor set, the Favard length of the 2^{-dN}-neighbourhood is squeezed between N^{-1} and N^{-ε} for all large N, pinning the decay rate to a power law.","In dimensions d≥3 this is the first nontrivial asymptotic upper bound for a purely 1-unrectifiable set whose H^1 measure is not concentrated in a two-dimensional affine subspace.","The exponent analysis (Proposition 3.20) gives a universal, dimension-free function δ(c1,C1) describing how the power depends on the SSV and SLV constants, and it recovers an exponent 1/4 for the planar four-corner set, improving the previously stated 1/6.","Any digit set satisfying one of the three conditions—small cardinality, two-prime accumulating lcm, or fibered subset—yields a power law, with a log-log loss only when the mask polynomial has irrational-phase roots.","The fibered condition produces power laws for a family of planar rational product Cantor sets that were not covered by earlier planar theorems."],"supporting_citations":[{"why":"Supplies the original planar power-law method and the explicit 1/6 exponent that this paper generalizes to d≥3.","marker":"[27]"},{"why":"Introduces the SSV/SLV machinery and proves planar power laws for rational product Cantor sets with max digits ≤6; supplies the set-of-small-values lemmas.","marker":"[3]"},{"why":"Extends the planar method to countably infinite digit sets via cyclotomic factorization; source of Lemma 5.1 and the #A≤10 / two-prime hypotheses.","marker":"[22]"},{"why":"Provides fibered sets, assignment functions, and the cardinality bound Proposition 2.11 on which the new fibered case rests.","marker":"[17]"},{"why":"Gives the matching lower bound C^{-1}N^{-1} used in Theorem 1.2.","marker":"[24]"},{"why":"Contains the combinatorial propagation and L2 estimates that Section 6 generalizes to higher dimensions.","marker":"[2]"},{"why":"Supplies the Poisson-localization estimate used to control the low-frequency integral in Proposition 4.3.","marker":"[6]"},{"why":"Standard reference for the Besicovitch-Federer theorem, self-similar sets, and the open set condition framing the problem.","marker":"[25]"}],"fun_headline_variants":["First non-trivial Favard decay bounds in d≥3","Power law for Favard length of higher-dimensional Cantor dust","Cyclotomic trick powers Favard decay in every dimension","Favard length power law extends beyond the plane"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The new fibered-digit result depends entirely on an unproved-in-this-paper combinatorial estimate, taken from the author's own preprint reference [17], that any digit set containing a fibered subset must have at least a certain explicit number of digits; if that estimate is false or has a gap, the construction of large-value sets in Section 5 fails and the fibered case collapses.","fun_headline_variants_meta":{"raw":{"variants":["First non-trivial Favard decay bounds in d≥3","Power law for Favard length of higher-dimensional Cantor dust","Cyclotomic trick powers Favard decay in every dimension","Favard length power law extends beyond the plane"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.002066,"raw_usage":{"total_tokens":7861,"prompt_tokens":712,"completion_tokens":7149,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":456,"completion_tokens_details":{"reasoning_tokens":7080}},"tokens_in":456,"tokens_out":7149,"duration_ms":55411,"temperature":1.0,"reasoning_tokens":7080,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T11:18:11.245712+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find a digit set A with a fibered subset whose cardinality is strictly less than min_σ FIB(S^{(2)}_A, σ); that would disprove the paper's key combinatorial input (Proposition 2.11) and collapse Theorem 2.10. Alternatively, numerically measure Fav(N_{2^{-dN}}(K_d^∞)) for d=3 and see whether it decays like N^{-ε} rather than N^{-ε/log log N}.","supporting_citations":[{"cited_title":"Nazarov, Y","cited_arxiv_id":null,"evidence_quote":"Supplies the original planar power-law method and the explicit 1/6 exponent that this paper generalizes to d≥3."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the SSV/SLV machinery and proves planar power laws for rational product Cantor sets with max digits ≤6; supplies the set-of-small-values lemmas."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Extends the planar method to countably infinite digit sets via cyclotomic factorization; source of Lemma 5.1 and the #A≤10 / two-prime hypotheses."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides fibered sets, assignment functions, and the cardinality bound Proposition 2.11 on which the new fibered case rests."},{"cited_title":"Mattila, Orthogonal projections, Riesz capacities, and Minkowski content, Indiana Univ","cited_arxiv_id":null,"evidence_quote":"Gives the matching lower bound C^{-1}N^{-1} used in Theorem 1.2."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Contains the combinatorial propagation and L2 estimates that Section 6 generalizes to higher dimensions."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Poisson-localization estimate used to control the low-frequency integral in Proposition 4.3."},{"cited_title":"Mattila: Geometry of Sets and Measures in Euclidean Spaces , Cambridge University Press, (1995)","cited_arxiv_id":null,"evidence_quote":"Standard reference for the Besicovitch-Federer theorem, self-similar sets, and the open set condition framing the problem."}],"review_version":1}