{"id":"6d685ed1-f18c-4d4d-8ba2-9c3ca8815a81","arxiv_id":"2502.08797","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A Ward identity analysis provides the first non-perturbative estimates of the O(a) improvement coefficient g_S for the flavour singlet scalar density in three-flavour QCD.","lead":"This lattice QCD paper reports the first non-perturbative estimates of a small correction coefficient, g_S, used to remove discretisation errors from computer simulations of the strong force. The values agree in size with perturbation theory, and the paper explains how to turn them into a non-perturbative determination of the related coefficient b_g.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Reported g_S is defined without the clover-term contribution (eq. 14) needed for the b_g relation; the O(a) ambiguity this creates is not quantified, so the 'first non-perturbative estimates' are conditional.","rationale":"The reader's weakest assumption concerns the completeness of the improvement pattern in eq. (8) and the relation of the extracted g_S to the coefficient needed in applications, particularly b_g. I agree that this is the most load-bearing issue: the paper itself discloses that the b_g relation requires an additional clover-term contribution (eq. 14) which is not yet implemented. However, I do not see this as invalidating the central claim that non-perturbative estimates of g_S have been obtained for the first time under a well-defined, though incomplete, discretisation. The same argument would also apply to the use of g_S for sigma terms, where any symmetric local discretisation of Tr[F F] with the appropriately tuned coefficient should improve the operator; the missing clover term introduces an O(a) ambiguity rather than an error at leading order. The problem is presentation: the abstract's b_g claim is stronger than what the data support, and the magnitude of the ambiguity is unquantified. This warrants a conditional verdict (as the reader already gave), but not rejection. The concrete test - re-evaluating with the clover term and comparing shifts - would settle whether the current numbers are preliminary or final for the advertised purposes. No other concern appears to be more load-bearing; the Ward identity derivation is standard, the cancellation of the e_S term is plausible, and the numerical setup follows previous work by the same groups.","tokens_in":10785,"tokens_out":22936,"duration_ms":222499,"concrete_test":"Implement the clover-term diagrams of eqs. (15)-(17) and re-evaluate g_S on the same five ensembles and with the same T/3 and T/4 choices, replacing {eTr[FF]}_{Sg} in eq. (10) by the full expression in eq. (14). If the resulting g_S values differ from Fig. 1 by less than one standard deviation at each coupling, the reported estimates are stable under the discretisation choice and the central claim stands as stated. If the shift is larger than the quoted errors, the current numbers should be re-reported and the abstract's b_g connection should be softened until the reanalysis is complete.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim rests on a specific discretisation of the improvement term in eq. (8): only the gauge-action-based {eTr[FF]}_{Sg} is used in the Ward identity (10). The paper itself states (Section 5, eq. (14)) that for the relation b_g = 2 g0^2 g_S to hold, an additional clover-term contribution proportional to a g0^2 (dc_sw/dg0^2) must be included in the definition of the field-strength insertion. This missing term is not merely a different renormalisation convention: at the finite lattice spacings used here (a about 0.06-0.1 fm), the O(a) ambiguity in g_S induced by this omission could be comparable to the statistical errors and to the small extracted values near the chiral point. The abstract's claim that the relation to b_g 'can also be established, allowing for its non-perturbative extraction' is not realised by the presented numbers, which the authors themselves say cannot be used for b_g. What is lacking is a quantitative estimate of the difference between the two definitions; without it, the reader cannot tell whether the reported g_S values are the improvement coefficient relevant for sigma terms (where any symmetric local discretisation should suffice) or only an intermediate quantity tied to the gauge action. The paper plans to reevaluate g_S with the clover term, but the current numbers are preliminary under a definition that is not clearly the physically relevant one.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports a first non-perturbative determination of the O(a) improvement coefficient g_S of the flavour singlet scalar density in Nf=3 lattice QCD with Wilson-clover fermions and tree-level Symanzik-improved gauge action. The authors derive a lattice Ward identity (eq. (10)) from two subtracted chiral Ward identities, using Schrödinger functional boundary conditions and previously determined renormalisation/improvement parameters Z, r_m, and c_A. The identity is solved on five CLS ensembles with g0^2 in [1.5,1.77], yielding g_S estimates that approach zero at small coupling and are compatible with zero at the two smallest couplings. The paper also discusses the relation b_g = 2 g0^2 g_S, pointing out that an additional clover-term contribution (eq. (14)) is required for this relation and that the current results cannot yet be used to extract b_g.","tokens_in":11060,"tokens_out":11839,"duration_ms":114803,"significance":"If the results hold, they provide the first direct non-perturbative estimates of g_S in the coupling range used in large-volume CLS simulations, which is relevant for O(a) improvement of meson and baryon sigma terms. The derivation is careful and the numerical analysis uses standard tools (Gamma-method errors, trivial-topology projection, LCP ensembles with nearly massless quarks). The paper is transparent about the missing clover term and clearly states that the b_g extraction is future work. The main value is a proof-of-principle and a set of finite-lattice-spacing estimates that can guide sigma-term calculations, though continuum extrapolation and the clover-term correction remain to be completed.","major_comments":[{"comment":"The reported g_S is defined without the clover-term contribution that eq. (14) shows is necessary for the b_g relation. The paper itself calls this \"merely an O(a) ambiguity\" (Section 4). At the lattice spacings used (a ~ 0.06-0.1 fm), an O(a) ambiguity can be comparable to the quoted statistical errors and to the small extracted values near the chiral point, so the numbers in Fig. 1 may not be the g_S needed for sigma-term improvement if a different discretisation of the field strength is ultimately adopted. A quantitative estimate of the difference between the two definitions (e.g., a perturbative estimate or an evaluation of the clover correlation functions on one ensemble) should be provided, or the abstract should make clear that the presented g_S is an intermediate, definition-dependent quantity and that the b_g extraction is not yet realised.","section":"Sections 4 and 5, eq. (14)"},{"comment":"The claim that this is the first non-perturbative determination of g_S needs qualification with respect to Ref. [7], which determines b_g non-perturbatively in the neighbouring coupling range g0^2 in [0.4,1.5]. Since b_g = 2 g0^2 g_S holds when the clover term of eq. (14) is included, Ref. [7] effectively provides a non-perturbative g_S in a different scheme/coupling range. The paper should explicitly state why the present determination is novel (direct Ward identity, different definition, different coupling range) or soften the \"first time\" wording to avoid an overstated novelty claim.","section":"Introduction and Section 5"}],"minor_comments":[{"comment":"The phrase \"a relation to b_g ... can also be established, allowing for its non-perturbative extraction as well\" is forward-looking; the abstract should make clear that the extraction of b_g is not performed in this paper and that the presented g_S cannot yet be used for that purpose.","section":"Abstract"},{"comment":"The figure caption notes that the perturbative prediction is not directly comparable to the data, but the main text would benefit from repeating this caveat when the data are discussed, to prevent the reader from inferring a discrepancy.","section":"Section 4, Fig. 1"},{"comment":"The sign convention g_S = -d_S with respect to Ref. [7] is only given in a footnote; it should also appear in the main text where g_S is first introduced, because the comparison with perturbative expressions and with future work depends on this sign.","section":"Footnote 1 and Section 4"},{"comment":"The column heading \"Nsep [MDU]\" is ambiguous; the text explains it, but the table would be clearer with a heading such as \"Nsep [MDU]\" followed by a footnote, or an explicit statement that Nsep is the separation in molecular dynamics units.","section":"Table 1"},{"comment":"The definitions of the connected correlation functions are deferred to Ref. [11]; since this is a proceedings paper, a brief restatement or at least a clear mapping of the notation (e.g., \"con\", \"disc\") to the definitions in the appendix of [11] would improve readability.","section":"Section 4, eqs. (11)-(12)"}],"recommendation":"minor_revision","confidential_remarks":"The paper is a proceedings contribution and its central derivation appears sound. The main risk is the O(a) ambiguity from the missing clover term, which the authors acknowledge but do not quantify; the editor may want to encourage them to add a quantitative estimate or a more explicit caveat. The novelty claim relative to Ref. [7] should also be scrutinised, as Ref. [7] already contains non-perturbative b_g information that implies a non-perturbative g_S in a different scheme. For a proceedings, minor revision seems appropriate."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The one-line take: this is a genuine first non-perturbative estimate of g_S in the CLS coupling range, and the Ward identity trick to cancel the e_S divergence is clean, but the reported numbers are not yet the b_g-relevant coefficients, and the size of the O(a) ambiguity is not quantified.\n\nWhat is actually new: figure 1 gives g_S at five couplings, g0^2 in [1.5,1.77], from three-flavour Wilson-clover ensembles on an LCP, using a new Ward identity that subtracts two singlet identities to eliminate the cubic divergence. The g_S values are small, consistent with zero at the two smallest couplings, matching the historical expectation that they can be neglected in sigma term calculations. Previous knowledge was one-loop perturbation theory and a b_g determination at lower couplings. The paper is honest that these numbers come with a caveat: relation (14) shows that to extract b_g via b_g=2 g0^2 g_S, an additional clover-term contribution must be included in the field-strength insertion, and the current analysis omits it. So the abstract's statement that the relation to b_g can be established is a plan, not something the present numbers deliver.\n\nThe strengths: the derivation of eq. (10) is careful; the inputs Z, r_m, c_A come from independent non-singlet determinations in their earlier papers, not from a fit to the same data; the error analysis uses the Gamma method and trivial-topology projection; the LCP tuning is good. The paper also clearly explains the additional diagrams needed for the clover term, which will allow a straightforward follow-up.\n\nSoft spots: this is a proceedings paper, so no continuum extrapolation, correlation function definitions deferred to ref [11], and the data are not shipped. More substantively, the current g_S belongs to a specific discretisation of the field strength (gauge-action-based). The missing clover term is an O(a) ambiguity, and at a ~ 0.06-0.1 fm it could be comparable to the statistical errors and to the small values near the chiral point. Without a quantitative estimate of the shift, a reader cannot tell whether these g_S values are the ones needed for sigma terms (where any local symmetric discretisation should work) or only intermediate quantities. The paper's own discussion acknowledges this, so it is a conditional, not fatal, weakness.\n\nWho it's for: lattice QCD practitioners computing sigma terms with Wilson fermions in the CLS range, and anyone working on b_g. It deserves serious refereeing as the first non-perturbative determination of g_S; I'd send it out, with the expectation that the authors either quantify the O(a) ambiguity or wait for the clover-term version.","headline":"First non-perturbative g_S estimates in the CLS range, clean Ward identity, but the b_g connection is still a plan and the omitted clover term leaves the O(a) ambiguity unquantified.","tokens_in":11621,"tokens_out":2896,"would_cite":true,"duration_ms":25458,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["11.15.Ha","12.38.Gc"],"model":"deepseek-v4-flash","headline":"The paper reports the first non-perturbative estimates of the O(a) improvement coefficient $g_S$ of the flavour singlet scalar density, finding values compatible with zero at the smallest couplings.","keywords":["O(a) improvement","flavour singlet scalar density","g_S","Ward identity","Wilson fermions","lattice QCD","b_g","sigma terms"],"falsifier":"Evaluate the extended Ward identity with the clover-term replacement of eq. (14) on the same five ensembles: if the resulting $g_S$ differs from the values in Fig. 1 by more than the quoted errors, the current estimates are only an intermediate quantity, and if the values remain zero-compatible after the clover term, the conclusion that $g_S$ is small in this coupling range is confirmed.","tokens_in":10540,"feed_emoji":"🧮","tokens_out":18140,"duration_ms":150417,"temperature":0.7,"pith_summary":"The paper aims to determine, for the first time, the non-perturbative value of $g_S$, the O(a) improvement coefficient of the flavour singlet scalar density in lattice QCD with Wilson-clover fermions. It derives a lattice Ward identity in which $g_S$ is the only unknown and evaluates it on five ensembles with $N_f=3$ mass-degenerate quarks at couplings $g_0^2 \\in [1.5,1.77]$, the range used in large-volume simulations. The resulting estimates decrease toward zero at small couplings and are compatible with zero within one standard deviation at the two smallest couplings. This supports the common practice of neglecting $g_S$ in meson and baryon $\\sigma$-term calculations, and the paper also shows how the same identity must be extended with a clover-term insertion before it can deliver the gauge-coupling improvement coefficient $b_g$.","feed_headline":"First non-perturbative g_S values hug zero","feed_subtitle":"A Ward-identity extraction at five lattice spacings backs neglecting g_S in sigma terms and paves the way to b_g.","key_machinery":"The central object is the O(a)-improved flavour singlet scalar density $(S_I)^0$, with improvement pattern containing a power-divergent term proportional to $e_S$ and an O(a) term proportional to $g_S$ times the field-strength insertion $e\\mathrm{Tr}[F_{\\mu\\nu}F_{\\mu\\nu}]$, discretised in the local, site-symmetric form $\\{e\\mathrm{Tr}[FF]\\}_{S_g}$. The argument is carried by a Ward identity obtained by subtracting two axial Ward identities, one with the operator $S_0 O_{ext}$ and one with $1\\cdot O_{ext}$; this subtraction cancels the divergent $e_S$ term and leaves $g_S$ as the only unknown once the renormalisation factors $Z$, $r_m$ and the improvement coefficient $c_A$ are supplied. The route to $b_g$ runs through $b_g = 2g_0^2 g_S$, which requires replacing $\\{e\\mathrm{Tr}[FF]\\}_{S_g}$ by $\\{e\\mathrm{Tr}[FF]\\}_{S_g} + a\\,g_0^2\\,(\\partial c_{sw}/\\partial g_0^2)\\,(i/2)O_{clover}$ in the Ward identity.","core_discovery":"On its own terms, the paper claims: non-perturbative estimates for $g_S$ have been determined for the first time. The extraction uses the chiral Ward identity (10), solved on five ensembles at $g_0^2 \\in [1.5,1.77]$ with nearly massless, O(a)-improved Wilson fermions and the tree-level Symanzik improved gauge action. The estimates, shown in Fig. 1 for two choices of the time interval, approach zero for small couplings and are compatible with zero within 1-$\\sigma$ at the two smallest couplings, confirming that $g_S$ is small in this range. The paper further establishes that the relation $b_g = 2g_0^2 g_S$ holds only after an additional, clover-term-like contribution is included in the Ward identity, and it outlines the implementation of that contribution as the next step toward non-perturbative $b_g$.","pith_inferences":["If the missing clover-term contribution shifts $g_S$ by an amount comparable to or larger than the quoted errors, earlier sigma-term analyses that neglected $g_S$ would need a revised O(a) uncertainty estimate; the paper does not quantify that shift.","The same Ward identity could be evaluated at additional lattice spacings to test whether the trend toward zero continues monotonically; a non-monotonic behaviour would signal that the remaining O(a) ambiguity is not negligible.","The stochastic-estimator technology for the disconnected diagrams transfers to other flavour-singlet quantities, such as the topological-charge density or the singlet axial current, where analogous improvement coefficients appear.","A non-perturbative $b_g$ obtained through $b_g = 2g_0^2 g_S$ would also test the one-loop assumption currently used in the scale and mass settings of large-volume Wilson-fermion simulations."],"forward_implications":["Sigma-term determinations with Wilson fermions at these couplings can continue to set $g_S=0$ without introducing a bias larger than the present statistical uncertainty.","The clover-term extension of the Ward identity gives a concrete route to non-perturbative $b_g$ values in the coupling range $g_0^2 \\in [1.5,1.77]$, where only one-loop perturbation theory is currently available.","Because the $g_S$ estimates are compatible with zero at the two finest lattices, the perturbative expectation of a small $g_S$ appears adequate near the continuum limit in this range.","The correlation functions and disconnected-diagram estimators developed here can be reused for other flavour-singlet Ward identities on the same ensembles."],"supporting_citations":[{"why":"defines $g_S$, the improvement pattern of the flavour singlet scalar density, and the relation $b_g = 2g_0^2 g_S$.","marker":"[1]"},{"why":"derives the clover-term correction to the field-strength discretisation that the $g_S$-$b_g$ relation requires.","marker":"[7]"},{"why":"supplies the renormalisation factors $Z$ and $r_m$ used to solve the Ward identity.","marker":"[11]"},{"why":"supplies the non-perturbative axial-current improvement coefficient $c_A$ appearing in the identity.","marker":"[12]"},{"why":"provides the ensembles on the line of constant physics used for the extraction.","marker":"[13]"},{"why":"supplies the non-perturbative clover coefficient $c_{sw}$ of the improved action.","marker":"[15]"},{"why":"defines the gradient-flow coupling used to tune the line of constant physics.","marker":"[16]"},{"why":"provides the frequency-splitting estimator used for single-propagator traces in disconnected diagrams.","marker":"[17]"},{"why":"justifies projecting the gauge ensembles onto the trivial topological sector for the Ward identity.","marker":"[19]"}],"fun_headline_variants":["First non-perturbative g_S estimates near zero","g_S measured non-perturbatively for first time","Near-zero g_S from five lattice spacings","First lattice QCD g_S paves way to b_g","Ward identity yields tiny g_S then b_g"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The result rests on the assumption that the field-strength term used to improve the singlet scalar density, built only from the gauge action, is the complete O(a) improvement, even though the paper itself notes that a clover-term piece must be added before $g_S$ can be translated into $b_g$; if that missing piece moves the numbers by more than the quoted errors, the values reported here are not the final $b_g$-relevant ones.","fun_headline_variants_meta":{"raw":{"variants":["First non-perturbative g_S estimates near zero","g_S measured non-perturbatively for first time","Near-zero g_S from five lattice spacings","First lattice QCD g_S paves way to b_g","Ward identity yields tiny g_S then b_g"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001063,"raw_usage":{"total_tokens":4446,"prompt_tokens":924,"completion_tokens":3522,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":540,"completion_tokens_details":{"reasoning_tokens":3443}},"tokens_in":540,"tokens_out":3522,"duration_ms":23073,"temperature":1.0,"reasoning_tokens":3443,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T23:37:43.890327+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Evaluate the extended Ward identity with the clover-term replacement of eq. (14) on the same five ensembles: if the resulting $g_S$ differs from the values in Fig. 1 by more than the quoted errors, the current estimates are only an intermediate quantity, and if the values remain zero-compatible after the clover term, the conclusion that $g_S$ is small in this coupling range is confirmed.","supporting_citations":[{"cited_title":"$\\mathcal{O}(a)$ improved quark mass renormalization for a non-perturbative matching of HQET to three-flavor QCD","cited_arxiv_id":"1811.02591","evidence_quote":"provides the ensembles on the line of constant physics used for the extraction."}],"review_version":1}