{"id":"cbc7f3be-d25b-48ec-92f8-a67a75c3a720","arxiv_id":"2607.15998","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"Conditional on GRH and modularity, the m-th moment of analytic rank of elliptic curves over a degree-d number field is bounded by O((9dm/2)^m), with a corresponding exponential tail bound.","lead":"This paper gives conditional upper bounds, assuming GRH and modularity of all elliptic curves, for every moment of the analytic rank of elliptic curves over number fields of degree d. A reader might care because this is the first all-moments statement over arbitrary number fields and includes a purported tail bound that improves on known results over the rationals.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Tail Theorem 1.3 is invalid as written: the odd-s_i error in Lemma 6.2 is B^{(3jν/2−1/(3d))}, and at j=2m, ν=2/(9dm) this is B^{1/(3d)}, so (7.6) fails.","rationale":"The reader's verdict of REJECT is supported by the same mechanical error term that I identify: Lemma 6.2, at the parameter values used in Theorem 1.3, contributes an error B^{1/(3d)} log^{2m}, invalidating equation (7.6). I agree with the reader's rationale, though the reader's stated weakest_assumption was the modularity hypothesis rather than this internal gap. The modularity hypothesis is a genuine conditional assumption but is explicitly stated in the theorems and is not, by itself, a correctness flaw. The internal error is the load-bearing concern because it breaks the proof of one of the paper's two advertised results. The moment theorem is plausibly repairable via ν < 2/(9dm) and a limiting argument, but the tail theorem is not obviously repairable without a new estimate for the odd-s_i contributions at j = 2m. Therefore the reader's REJECT verdict should stand.","tokens_in":26109,"tokens_out":27633,"duration_ms":256871,"concrete_test":"Recompute the proof of Lemma 6.2 for j = 2m, isolating Piece (iii) as in (6.20)–(6.25). Apply Proposition 5.1(iii) to the dominant partition s_1 = ... = s_{2m} = 1 and sum over distinct prime ideals with q_i ≤ B^ν. Verify whether the resulting error is B^{3jν/2−1/(3d)} log(B)^j or whether a cancellation reduces it to o(B^{1/(3d)}). If no cancellation is identified, plug the B^{1/(3d)} log^{2m} term into (7.5) and observe that (7.6) cannot hold for any fixed β, m, d.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central defect is in the proof of Theorem 1.3, not in the conditional assumptions. Lemma 6.2 bounds S1(j, φ, B) for even j by a main term plus an error containing B^{3jν/2−1/(3d)} log(B)^j. This error is not a typographical artifact: it comes from Piece (iii), partitions with some s_i odd, via Proposition 5.1(iii), as stated in equation (6.25). In the proof of Theorem 1.3 the authors take j = 2m and ν = 2/(9dm). Then 3jν/2 − 1/(3d) = 3·2m·(2/(9dm))/2 − 1/(3d) = 2/(3d) − 1/(3d) = 1/(3d) > 0. Therefore S1(2m, φ, B) has an error term B^{1/(3d)} log(B)^{2m}. Substituting this into (7.5) gives an upper bound whose error term is (constant)^{2m} B^{1/(3d)}, which diverges as B → ∞. Hence the equality in (7.6) is unjustified and the advertised tail bound does not follow from the written derivation. The moment theorem (1.1) may survive because there j ≤ m, so the same exponent is ≤ 0, and a strict-inequality limiting argument can repair it; but Theorem 1.3 uses exactly j = 2m, so this repair is unavailable. This is a load-bearing gap in one of the paper's two headline claims.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves conditional upper bounds for limsup moments of analytic ranks of elliptic curves over a number field K, assuming GRH for elliptic curve L-functions and their symmetric squares, and assuming modularity of all elliptic curves over K. The proof uses an explicit formula to bound the analytic rank by a smoothed sum over prime ideals, averages the resulting prime sums over the family of elliptic curves of bounded height, and estimates the averaged sums via the first author's counting theorems for elliptic curves with prescribed local conditions. As an application, the paper claims a conditional exponential upper bound for the density of elliptic curves with analytic rank at least β.","tokens_in":26519,"tokens_out":12787,"duration_ms":144872,"significance":"If correct, the moment bound would generalize results of Cho–Jeong (K=Q) and the first author (first moment over number fields), with explicit, parameter-free constants. The paper is transparent about its assumptions, and the reliance on independent published counting theorems is a strength. However, the tail theorem is a separate headline claim that improves on Heath-Brown's bound, and its proof contains a load-bearing technical gap. The dependence on modularity over arbitrary number fields is also a very strong hypothesis, though it is explicitly stated.","major_comments":[{"comment":"The proof of Theorem 1.3 applies Lemma 6.2 with j=2m and ν=2/(9dm). The second error term in Lemma 6.2 is B^{3jν/2−1/(3d)} log(B)^j. Substituting j=2m and ν=2/(9dm) gives exponent 3·(2m)·(2/(9dm))/2 − 1/(3d) = 1/(3d), so the error is B^{1/(3d)} log(B)^{2m}. After the division by log(B)^{2m} in (7.5), this remains B^{1/(3d)}, which diverges as B→∞. Thus the equality in (7.6) is not justified, and Theorem 1.3 and its asymptotic (1.5) do not follow from the written derivation. This is a load-bearing error.","section":"§7, Eqs. (7.5)–(7.6)"},{"comment":"Immediately after (6.27), the proof asserts that for odd j, Lemma 6.1 together with ν≤2/(9dm) implies the odd-j terms are o(1). When m is odd, j=m, and ν=2/(9dm), Lemma 6.1 gives S1(m) ≪ 1 + log(B)^m, so the term in (6.27) is only O(1) after division by log(B)^m, not o(1). Consequently the asymptotic in (6.26) is not established as written for m odd. This may be repairable by taking ν strictly smaller than 2/(9dm), but the stated theorem and proof need correction.","section":"§6, proof of Theorem 6.3"},{"comment":"The passage from the pointwise bound (3.99) to the averaged binomial expansion (3.108) is not fully justified for odd m. The expansion contains terms with signs (−2)^j for odd j, and replacing S1(j) by an upper bound is only legitimate for an upper bound of the whole expression after controlling absolute values. The boundary case in Theorem 6.3 shows that this matters: an O(1) contribution from an odd j term of either sign is not negligible. The proof should either use absolute values or prove nonnegativity of the relevant combinations.","section":"§3, Eqs. (3.102)–(3.108)"}],"minor_comments":[{"comment":"In the proof of Lemma 6.2, the phrase 'same argument as in Proposition 6.1' refers to a non-existent proposition; it should be 'Lemma 6.1'.","section":"§6, Lemma 6.2"},{"comment":"Typo: 'de green prime ideal' should presumably be 'degree one prime ideal' or similar.","section":"§4, after Prop. 4.2"},{"comment":"The notation δ_E and δ'_E is introduced with similar symbols; a short glossary would improve readability.","section":"Throughout"}],"recommendation":"reject","confidential_remarks":"The manuscript has a plausible and likely repairable moment theorem, but Theorem 1.3 is a headline application and its proof has a concrete asymptotic error that cannot be fixed by a local modification of the given argument. In my view this warrants rejection of the current version. If the authors were to remove or substantially weaken the tail theorem and repair the odd-j boundary issue in Theorem 6.3, a future version could be reconsidered."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things worth knowing before you open it. The moment bound over number fields is genuinely new — it generalizes Cho–Jeong from Q and Phillips's first moment to all moments, and the PCM factorization in Lemma 3.10 is a clean way to handle CM curves. But the paper's second advertised result, the tail bound of Theorem 1.3, does not follow from the proof as written. When they take j=2m and ν=2/(9dm), the odd-part error term in Lemma 6.2 is B^{1/(3d)} log^{2m} B. Substituted into (7.5), that error term grows with B, so the equality in (7.6) is unjustified and the exponent −2β/(9d) is unsupported.\n\nWhat is good: Sections 2–5 are coherent. The counting estimates are cited properly from the first author's earlier paper. The paper is honest about its conditional assumptions; the modularity assumption for arbitrary number fields is strong, but it is stated plainly and is not hidden. No fitting or constant tuning.\n\nThe soft spots are concentrated in Sections 6–7. In Theorem 6.3, the claim that odd-j terms are o(1) is false at the boundary j=m when ν=2/(9dm): Lemma 6.1 gives O(log^m B). That part is probably repairable by taking ν slightly below 2/(9dm) and then taking a limit, since the main term tends to the desired constant. The tail theorem is worse: the same limiting trick does not work, because for j=2m the error only becomes o(1) when ν < 1/(9dm), which changes the final optimization. So the advertised density bound needs a genuinely different error analysis or a revised statement.\n\nI would not desk-reject this. The moment theorem is a real step forward and the method is worth a referee's time. But I would send it back with a request to fix the tail section. The paper is for people in arithmetic statistics and low-lying zeros; its main value is the moment framework over number fields. My recommendation: send to peer review, with clear instructions that Theorem 1.3 as stated is not justified.","headline":"New moment bounds over number fields, but the tail theorem's proof has a divergent error term at the chosen parameters, so the advertised density exponent doesn't follow as written.","tokens_in":26993,"tokens_out":7329,"would_cite":true,"duration_ms":72058,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["11G05","11G40","11M41"],"pacs":[],"model":"deepseek-v4-flash","headline":"For every number field, the moments of analytic ranks of elliptic curves are bounded by (9dm/2)^m, conditionally.","keywords":["analytic rank","elliptic curves","number fields","moments","1-level density","explicit formula","symmetric square L-function","Frobenius traces"],"falsifier":"Compute, for a number field K of degree d, the unconditionally averaged m-th power of analytic ranks up to height B and show it eventually exceeds the bound (9dm/2)^m; even a single m and K violating the bound would refute the theorem's conclusion. A practical starting point would be the first moment over an imaginary quadratic field, which the bound caps at (9d+1)/2.","tokens_in":25975,"feed_emoji":"🧮","tokens_out":4967,"duration_ms":49183,"temperature":0.7,"pith_summary":"This paper establishes, under the Riemann Hypothesis for elliptic curve L-functions and their symmetric squares (plus modularity of all curves over the field), explicit upper bounds for every moment of the analytic rank in the family of elliptic curves over a number field of degree d. The m-th moment is bounded by a finite sum that is asymptotically (9dm/2)^m. As a consequence, the paper bounds the proportion of curves of analytic rank at least β, giving decay like β^{-2β/(9d)+o(β)}. These are the first such universal moment bounds over general number fields, and they unify and extend earlier conditional results for the rational field.","feed_headline":"Rank moments capped for elliptic curves over any number field","feed_subtitle":"Conditional proof bounds all moments by (9dm/2)^m and shows very high ranks are extraordinarily rare.","key_machinery":"The 1-level density D1(E/K, φ)=Σ_γ φ(γ log X/(2π)) of the low-lying zeros, together with the explicit formula expressing it as a main term plus a smoothed sum over prime ideals. A nonnegative test function φ lets the paper majorize analytic rank by D1/φ(0). The k=2 prime-power contribution is handled through the symmetric square L-function, whose possible pole at s=2 arises exactly for curves with CM defined over K. The Frobenius trace formula (Proposition 5.1) computes averages of products of traces over the height-ordered family, with odd products vanishing and even products exhibiting a leading term; the final test function φ(y)=(sin(πν y)/(2πy))^2 gives explicit constants.","core_discovery":"The central discovery is a reduction of moment bounds for analytic ranks to estimates for averaged products of Frobenius traces. Using the explicit formula, analytic rank is majorized by a smoothed sum over prime ideals attached to the curve's L-function; averaging over the height-ordered family, the odd moment contributions vanish and the even contributions factor into a leading term governed by a simple integral of the chosen test function. The main term yields the normal-moment expression, and the error terms are controlled by asymptotics for the number of curves with prescribed local conditions. The paper's theorem states that, conditionally, E_K[r_an^m] ≤ sum_{k=0}^{floor(m/2)} m!/((m-2","pith_inferences":["If modularity were established for elliptic curves over arbitrary number fields, the moment bounds would become unconditional in the GRH sense; the proof's chief obstruction is the modularity input, not the analytic machinery.","The coincidence between the moment formula and the moments of a normal distribution with mean (9dm+1)/2 and variance 1/3 suggests the 1-level density may itself be asymptotically Gaussian, and a proof via Isserlis' theorem could simplify the computation.","The reliance on GRH for symmetric square L-functions could be relaxed by averaging over the family (as the paper notes), potentially yielding unconditional-in-GRH moment bounds with weaker constants.","The tail exponent -2β/(9d) implies that in high-degree number fields, large ranks remain more probable; this strong dependence on d suggests rank distribution may look quite different over large-degree fields."],"forward_implications":["For any number field of degree d, all moments of analytic rank are finite and grow at most like (9dm/2)^m under the stated hypotheses.","The proportion of curves with analytic rank at least β decays like β^{-2β/(9d)+o(β)} for large β, faster than any fixed power of β.","The moment bound implies strong Markov-type tail decay, quantitatively improving earlier conditional estimates over the rational field.","The method works uniformly over all number fields, depending only on the degree d and the local-counting input, not on special arithmetic features of the base field.","The bounds are consistent with the minimalist philosophy that rank 0 and 1 exhaust almost all curves, though they do not prove that distribution."],"fun_headline_variants":["Elliptic curve rank moments bounded over all number fields","New conditional cap on rank moments for elliptic curves","High analytic ranks are rare: moment bound proves it","Bounding moments of elliptic curve analytic ranks","Elliptic curve rank moments: conditional upper bound"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The entire argument requires that every elliptic curve over the arbitrary number field K be modular; for general number fields, especially non-totally-real ones, this is unproven and is a substantially stronger input than the Riemann hypothesis assumptions, so if modularity fails the explicit formula used as the starting point is not available.","fun_headline_variants_meta":{"raw":{"variants":["Elliptic curve rank moments bounded over all number fields","New conditional cap on rank moments for elliptic curves","High analytic ranks are rare: moment bound proves it","Bounding moments of elliptic curve analytic ranks","Elliptic curve rank moments: conditional upper bound"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000557,"raw_usage":{"total_tokens":2385,"prompt_tokens":543,"completion_tokens":1842,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":287,"completion_tokens_details":{"reasoning_tokens":1770}},"tokens_in":287,"tokens_out":1842,"duration_ms":15542,"temperature":1.0,"reasoning_tokens":1770,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T21:41:51.704127+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute, for a number field K of degree d, the unconditionally averaged m-th power of analytic ranks up to height B and show it eventually exceeds the bound (9dm/2)^m; even a single m and K violating the bound would refute the theorem's conclusion. A practical starting point would be the first moment over an imaginary quadratic field, which the bound caps at (9d+1)/2.","supporting_citations":[],"review_version":1}