{"id":"d2ed1578-4aba-45cc-a3c7-b73b7228f919","arxiv_id":"2506.18707","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"A saddle-point expansion around the true, interacting-theory saddle point gives an analytic inverse transform of the QCD resummed form factor that matches exact numerical inversion, unlike the standard Taylor expansion around the free saddle point.","lead":"This paper develops an analytic approximation for converting a QCD resummed form factor from moment space back to physical space, using the true saddle point of the interacting theory instead of the free-theory saddle point. The method agrees with exact numerical inversion to percent level, while the traditional Taylor-based inversion deviates beyond its perturbative uncertainty.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq (19) as printed is internally inconsistent: h_y''>0 would make the Gaussian integral diverge; the saddle-point denominator requires h_y''<0. This sign error is in the central formula and must be fixed before the claim can be assessed.","rationale":"The reader's verdict is conditional, centering on the lack of formal control of the saddle-point expansion. I agree that is a gap: App A's frozen-coupling argument does not strictly transfer to the running-coupling case, and the paper itself states in Sec. 3 that no proof exists for arbitrary y. However, I see a more immediate defect: the central formula as written has the wrong sign for h_y''. This is not a subtle asymptotics issue; it changes the Gaussian from convergent to divergent. The numerical plots may well be correct after an implicit sign fix, but with the equations as printed a reader cannot reproduce them. This is the single most load-bearing concern because it sits in the exact result the paper claims. I therefore keep the conditional verdict: the paper should correct the sign, re-derive the anharmonic coefficients, and ideally provide code or data. The expansion-control issue remains secondary; the reported agreement and small-looking sextic corrections are suggestive but not a proof, and the frozen-coupling limit provides only partial justification for the running-coupling case.","tokens_in":13356,"tokens_out":13359,"duration_ms":146975,"concrete_test":"Evaluate the saddle-point second derivative at a representative point, e.g. the regularized LL thrust case at τ=0.01, using f~1 from Eq (42). Compute g_y''(N) numerically; if g_y''>0, then Eq (20) should read h_y''(0)=-g_y''(N)<0. Recompute Σ from Eq (19) with denominator sqrt(-2π h_y''(0)) and compare with the exact numerical inversion for τ∈[0.002,0.05]. If the corrected formula reproduces the reported blue curves while the printed formula gives imaginary or divergent values, the sign inconsistency is confirmed and must be fixed.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In Sec. 3, Eqs (15)-(20), the paper defines h_y(ν)=g_y(N+iν) and writes Σ=(1/2π)∫dν exp[h_y(ν)]. For the integral to converge, the exponent must have a maximum at ν=0, hence h_y''(0)<0. The paper instead states in Eq (20) that h_y''(0)=-g_y''(N)>0 and inserts this positive value into the denominator of Eq (19). With h_y''>0 the Gaussian integrand grows away from ν=0, so the basic saddle-point formula is not defined. The correct expression is exp[h_y(0)]/sqrt(2π |h_y''(0)|), or equivalently sqrt(2π g_y''(N)) when g_y''(N)>0 at the saddle. The same sign enters the anharmonic coefficients c_k in Eqs (21)-(23), so the claimed expansion is not well-defined as written. Because Eq (19) is the paper's central result, this inconsistency is load-bearing: the numerical agreement in Figs 2-4 cannot be checked against the printed formula, and no code or data are provided. A sign typo is plausible, but it must be stated explicitly and the corrected formula re-derived.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a fully analytic saddle-point inversion formula for the resummed QCD form factor of event-shape distributions, expanding the exponent of the inverse-transform integrand around the solution of the full (interacting-theory) saddle-point equation rather than around the free-theory point N = 1/y. The saddle point is obtained numerically and by an analytic recursion, and the resulting form factor is compared with the exact numerical inverse transform for thrust at LL, NLL, and NNLL accuracy, finding percent-level agreement while the standard CTTW Taylor expansion deviates by more than the estimated perturbative uncertainty. A frozen-coupling limit is used to exhibit a formal large expansion parameter.","tokens_in":13640,"tokens_out":9259,"duration_ms":89792,"significance":"If the central formula is corrected, the paper would provide a useful analytic route from N-space resummed form factors to momentum-space event-shape predictions, and it would independently confirm the authors' earlier numerical finding that standard Taylor-based analytic inversions can differ substantially from exact numerical inversion. The recursive saddle-point solution and the explicit regularized form-factor expressions are concrete and testable. However, the printed derivation of the central formula contains a load-bearing sign inconsistency that prevents the claim from being assessed as written.","major_comments":[{"comment":"The printed saddle-point formula is internally inconsistent. Eq. (20) asserts h''_y(0) = -g''_y(N) > 0, but for the integral in Eq. (15) to converge the exponent must have a maximum at ν = 0, which requires h''_y(0) < 0; indeed Eq. (21) uses e^{-x^2}, which corresponds to the opposite sign. With h''_y(0) > 0 the Gaussian integrand grows away from ν = 0, and the denominator sqrt(2π h''_y(0)) in Eq. (19) is not real. The correct expression is Σ ≃ exp[h_y(0)]/sqrt(-2π h''_y(0)) = exp[g_y(N)]/sqrt(2π g''_y(N)) when g''_y(N) > 0. The same sign error propagates into the change of variable x = ν sqrt(h''_y(0)/2) and into the coefficients c_k in Eq. (23), so the stated expansion is not well-defined as written. Because Eq. (19) is the paper's central result, the authors must correct the sign, re-derive Eqs. (21)-(23), and confirm that the numerical curves in Figs. 2-4 were obtained with the corrected formula.","section":"Sec. 3, Eqs. (19)-(23)"},{"comment":"The treatment of anharmonic corrections is also not well-defined as printed. After the sign of h''_y is fixed, x must be defined with |h''_y(0)|, and the statement that 'anharmonic terms can be included only if c_k i^k < 0' is at most a necessary condition for an individual term; it does not guarantee convergence of the integral in Eq. (21) when several coefficients are present, because the large-x asymptotic behavior is controlled by the highest-order term retained. The paper should specify exactly which terms of K(x) are kept in the 'sextic' approximation, verify the convergence of the resulting integral, and state the numerical impact of the quartic and sextic terms separately.","section":"Sec. 3, Eqs. (21)-(23)"},{"comment":"The 'exact numerical inversion' used as the benchmark is not sufficiently specified. The contour, integration method, numerical precision, and the value of the regularization parameter r in Eq. (42) (in particular the scale at which α_S is evaluated) are not given. Since the central evidence is the agreement between the saddle-point approximation and this exact inversion, the numerical procedure must be described in enough detail for the comparison to be reproducible and independently checkable.","section":"Sec. 4, Figs. 2-4"}],"minor_comments":[{"comment":"The sentence stating that 'the function h_y(ν) has its minimum at ν = 0' should read 'maximum' once the sign of h''_y(0) is corrected; the maximum is required for the Laplace-type approximation used here.","section":"Sec. 3, after Eq. (20)"},{"comment":"The statement that the saddle-point method is 'fully consistent with the numerical inversion within the Minimal Prescription' for the unregularized form factor is not accompanied by any plot, table, or quantitative estimate; adding this comparison would strengthen the claim.","section":"Sec. 4, final paragraph"},{"comment":"The notation F(k)(α_S, ℓ) for the k-th derivative with respect to ℓ should use a superscript F^{(k)} to avoid confusion with an argument k, and ℓ = ln(1/y) should be explicitly distinguished from L = ln N throughout.","section":"Sec. 2, Eq. (8)"},{"comment":"The convergence of the recursion u_{n+1} = Φ(u_n) is demonstrated for a single value y = 0.01 at LL accuracy; the authors should state the conditions under which this fixed-point iteration is expected to converge and comment on how the number of iterations varies with y and with the logarithmic order.","section":"Sec. 3.1"},{"comment":"The claim that the frozen-coupling limit provides a good approximation to the running-coupling case because the analytic coupling saturates at π/β_0 is heuristic; the relation between the value π/β_0 ≃ 1.6 and the artificially chosen A = 0.2 in Eq. (46) is not explained.","section":"App. A, final paragraph"}],"recommendation":"major_revision","confidential_remarks":"The sign error in Eq. (20) is central but appears to be a correctable typo, since Eq. (21) and the reported numerical agreement use the opposite sign. The paper should be carefully revised, with the corrected formula and a full description of the numerical benchmark, before it can be considered for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nYou should know two things about arXiv:2506.18707. First, it does something genuinely new: it inverts the resummed event-shape form factor by expanding around the true saddle point of the interacting theory, rather than around the free-theory expansion point N = 1/y used by CTTW and successors. Second, the numerical evidence is strong: with an analytic coupling that removes the Landau pole, their saddle-point inversion reproduces the exact numerical Laplace inversion at the percent level from tau ~ 0.05 down to ~0.002, while the CTTW Taylor inversion misses by more than the perturbative uncertainty band over most of that range. That difference matters, because it is the same approximation that feeds alpha_s determinations from thrust.\n\nWhat the paper does well: it sets up the saddle-point equation, solves it by a simple recursion that converges quickly, regularizes the Landau singularity with an analytic coupling, and checks the approximation at LL, NLL, and NNLL, plus a frozen-coupling limit where a large parameter (ell = ln 1/tau) formally justifies the expansion. The authors are also honest that no such large parameter can be extracted for the running-coupling case, and they give a plausibility argument rather than a proof.\n\nThe soft spot is real and in the central formula. Eq (15) defines the form factor as ∫ dnu exp[h_y(nu)], and for that integral to converge h_y must have a maximum at nu=0, i.e. h_y''(0) < 0. But Eq (20) states h_y''(0) = -g_y''(N) > 0, and the Gaussian denominator in Eq (19) uses that positive value. With h_y''>0, the Gaussian integrand grows and the integral is not defined; the change of variables in Eq (21) then produces e^{-x^2} only if the second derivative is negative. This is almost certainly a sign typo — the correct denominator is sqrt(2 pi g_y''(N)) — but it is load-bearing as written: without correcting it, the central formula cannot be checked. The authors should re-derive Eqs (19)-(23) with explicit signs.\n\nBeyond that, the lack of code or data is a minor weakness; and the convergence of the recursion is only shown for one point, but that is easy to extend. The paper is for resummation practitioners, especially anyone using analytic momentum-space formulas for event shapes or threshold resummation, and for those studying analytic inversion of Laplace integrals in QCD.\n\nMy recommendation: send it to peer review. The method is credible, the numerics are convincing, and the sign issue appears fixable. A referee should ask for the corrected formula, a clarification of the sign convention, and ideally the code or numbers behind the figures. After those revisions, I would take the result seriously.","headline":"A genuinely new saddle-point inversion for event-shape resummed form factors that agrees with exact numerics far better than CTTW, but the printed central formula has a sign error that needs a fix before the claim is fully assessable.","tokens_in":14149,"tokens_out":5735,"would_cite":false,"duration_ms":52122,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper claims that analytic inversion of resummed QCD form factors succeeds when the expansion is made around the true saddle point of the interacting theory, matching exact numerical inversion within perturbative uncertainty where the…","keywords":["saddle-point method","Sudakov resummation","event shape distributions","thrust","form factor inversion","Landau singularity","analytic QCD coupling","true saddle point"],"falsifier":"Evaluate the inverse-transform integral in Eq. (15) numerically to high precision for the regularized thrust form factor at a fixed perturbative value such as $\\tau = 0.01$, and compare it with the closed Gaussian formula in Eq. (19) while checking that the anharmonic coefficients $c_k$ of Eq. (23) form a convergent, subleading series: the central claim requires percent-level agreement and a small convergent anharmonic correction, so a deviation of several percent, or a non-convergent anharmonic series, would settle the claim negatively.","tokens_in":13149,"feed_emoji":"⚛️","tokens_out":15544,"duration_ms":136318,"temperature":0.7,"pith_summary":"This paper takes on a long-standing technical problem in perturbative QCD: converting an all-orders resummed form factor from the moment space where it factorizes into the physical space where event-shape data are collected. The authors present a fully analytic inversion formula obtained by expanding the exponent of the Laplace integral around the saddle point of the full interacting theory instead of the free-theory saddle point used by the classical Taylor-based inversion. For the thrust distribution at leading, next-to-leading, and next-to-next-to-leading logarithmic accuracy, the formula agrees with the exact numerical inverse transform at the percent level down to $\\tau \\sim 0.002$, while the standard analytic inversion deviates outside its perturbative uncertainty band. If correct, the result supplies an accurate analytic route to momentum-space Sudakov resummation and confirms, in a fully analytic framework, the earlier numerical finding that the standard momentum-space formula biases the extraction of $\\alpha_S(m_Z)$ from thrust.","feed_headline":"True saddle point gives accurate analytic QCD inversion","feed_subtitle":"Expanding around the interacting saddle point matches exact numerics; the standard Taylor inversion does not.","key_machinery":"The load-bearing mechanism is the saddle-point method applied to the Laplace-inversion exponent $g_y(N)$. Its characteristic move is to expand around the 'true' saddle point $\\bar N(y)$ — the stationary point of the full interacting theory — rather than around the free-theory point $N = 1/y$ used by the classical inversion, and the paper reduces the resulting transcendental equation to the nested-logarithm fixed-point problem $u = \\Phi(u)$ for $u = y\\bar N$, solved analytically by recursion $u_{n+1} = \\Phi(u_n)$ with rapid convergence. The Gaussian evaluation of the integral then produces the closed form-factor formula, with the Landau singularity handled beforehand by replacing the running coupling with a Landau-free analytic coupling, and with the frozen-coupling limit supplying the formal control parameter $\\ell = \\ln(1/y)$ for the expansion.","core_discovery":"On the paper's own terms, the central claim is that the inverse transform $\\Sigma(y,\\alpha_S)$ of the resummed form factor is dominated by the saddle point $\\bar N(y)$ of the interacting theory — the solution of $g'_y(\\bar N)=0$ for the full exponent $g_y(N) = yN - \\ln N + \\ln N\\, f_1(\\lambda) + \\sum_{n\\ge 0}(\\alpha_S/\\pi)^n f_{n+2}(\\lambda)$ — and that a quadratic expansion around it yields the closed analytic form $\\Sigma(y,\\alpha_S) \\simeq \\exp[h_y(0)]/\\sqrt{2\\pi h''_y(0)}$. The saddle point is obtained either numerically or by iterating the fixed-point equation $u = \\Phi(u)$ with $u = y\\bar N$, starting from the free-theory value $u_0 = 1$; the recursion converges to better than $0.01\\%$ within a handful of steps. The paper shows that this Gaussian formula tracks the exact numerical inversion of the Landau-regularized form factor within perturbative uncertainty for $\\tau$ down to about $0.002$, at LL, NLL, and NNLL accuracy, whereas the classical Taylor expansion of the exponent around the free-theory point $N = 1/y$ leaves the uncertainty band. The same analytic construction reproduces the authors' earlier numerical conclusion that the standard momentum-space formulation yields an $\\alpha_S(m_Z)$ from thrust inconsistent with the world average.","pith_inferences":["A refit of existing thrust data with the saddle-point-inverted form factor — which this paper does not perform — is the direct test of the implied claim that the extracted $\\alpha_S(m_Z)$ moves from the low value produced by the standard momentum-space formula toward the world average.","The convergence of the fixed-point iteration $u_{n+1} = \\Phi(u_n)$ is demonstrated numerically, not proven; establishing a contraction bound on $\\Phi$ would turn the recursion into a rigorous existence proof for the true saddle point in the running-coupling case.","Repeating the saddle-point-versus-exact comparison under alternative Landau-singularity regularizations, beyond the minimal prescription, would separate regularization dependence from inversion-method error in the final form factor.","The critical $\\tau$ where the saddle point disappears is effectively a parameter-free estimate of the non-perturbative boundary, and comparing that boundary across event shapes and collision energies would test whether it scales as $\\Lambda_{\\mathrm{QCD}}/Q$."],"forward_implications":["Momentum-space Sudakov resummation for event shapes becomes analytically reliable without numerical integration: the closed saddle-point formula reproduces exact inversions at the percent level throughout the perturbative two-jet region.","Because the exponent structure of the form factor is universal, the same true-saddle-point construction carries over to other event-shape variables, such as heavy-jet mass and the C-parameter, and to threshold resummation, with only the coefficient functions changing.","The discrepancy between the standard Taylor-based momentum-space resummation and the exact inversion is reproduced in a fully analytic setting, tying the thrust $\\alpha_S(m_Z)$ bias seen in the earlier numerical study to the choice of the free-theory expansion point.","The method sets its own validity boundary: no saddle point exists below roughly $\\tau \\sim 0.004$–$0.01$, so the disappearance of the saddle point marks the scale $\\tau \\sim \\Lambda_{\\mathrm{QCD}}/Q$ at which non-perturbative corrections take over."],"supporting_citations":[{"why":"Defines the classical Taylor-expansion inversion in momentum space that serves as the baseline and is found wanting.","marker":"[1]"},{"why":"The authors' earlier numerical study of thrust resummation whose discrepancy findings the present analytic method confirms.","marker":"[10]"},{"why":"Supplies the Minimal Prescription used for exact numerical inversion of the unregularized form factor and for the earlier comparison of strong-coupling determinations.","marker":"[12]"},{"why":"Provides the analytic, Landau-singularity-free coupling and the 'tilde' regularization of the resummation functions used in all numerical comparisons.","marker":"[17]"},{"why":"Gives the explicit leading-logarithmic form of the saddle-point equation that the recursion solves.","marker":"[22]"},{"why":"Underpins the analytic QCD coupling construction on which the regularization procedure of the paper is based.","marker":"[23, 24]"}],"fun_headline_variants":["Saddle point beyond free theory fixes QCD inversion","Interacting saddle point yields exact form factor inversion","Analytic inversion matches numerics via true saddle point","Saddle-point recursion beats Taylor in QCD resummation","Accurate analytic QCD form factor from interacting saddle"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the inversion integrand decays fast enough around the true saddle point for the Gaussian term to dominate: in the running-coupling case the authors point out that no large control parameter can be extracted from the integral, so the accuracy of the closed formula rests on the unproven supposition that the anharmonic corrections remain small.","fun_headline_variants_meta":{"raw":{"variants":["Saddle point beyond free theory fixes QCD inversion","Interacting saddle point yields exact form factor inversion","Analytic inversion matches numerics via true saddle point","Saddle-point recursion beats Taylor in QCD resummation","Accurate analytic QCD form factor from interacting saddle"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000377,"raw_usage":{"total_tokens":2022,"prompt_tokens":972,"completion_tokens":1050,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":588,"completion_tokens_details":{"reasoning_tokens":973}},"tokens_in":588,"tokens_out":1050,"duration_ms":7542,"temperature":1.0,"reasoning_tokens":973,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T18:44:50.181574+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Evaluate the inverse-transform integral in Eq. (15) numerically to high precision for the regularized thrust form factor at a fixed perturbative value such as $\\tau = 0.01$, and compare it with the closed Gaussian formula in Eq. (19) while checking that the anharmonic coefficients $c_k$ of Eq. (23) form a convergent, subleading series: the central claim requires percent-level agreement and a small convergent anharmonic correction, so a deviation of several percent, or a non-convergent anharmonic series, would settle the claim negatively.","supporting_citations":[{"cited_title":"A model for next-to-leading order threshold resummed form factors","cited_arxiv_id":"hep-ph/0407225","evidence_quote":"Provides the analytic, Landau-singularity-free coupling and the 'tilde' regularization of the resummation functions used in all numerical comparisons."}],"review_version":2}