{"id":"cd520b3a-b6e5-4eee-8a71-e5a3a222d289","arxiv_id":"2412.07930","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"A lattice implementation of the center-symmetric Landau gauge yields a local order parameter: the link average obeys predicted center-symmetry constraints below the deconfinement temperature and violates them above it.","lead":"The authors implement a specially designed gauge, the center-symmetric Landau gauge, on the lattice and show that the average gluon link signals the transition between confinement and deconfinement. This provides a local alternative to the nonlocal Polyakov loop as an order parameter for center symmetry in pure Yang-Mills theory.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The production algorithm's equivariance under the center transformation is unproven, but the direct numerical comparisons of E_gf with its tilde-g transform largely support the needed invariance; verdict unchanged.","rationale":"The reader's weakest_assumption is exactly the approximate tilde-g invariance of the gauge-fixed ensemble in the confined phase, and I agree that this is the most load-bearing point. The paper mitigates it with direct numerical evidence: the comparison of E_gf with its tilde-g transforms in Fig. 4 and the configuration-level tests in Sec. VI C show that the production algorithm produces an approximately invariant ensemble below T_c. The missing formal proof for FASD equivariance is a genuine rigor gap, but the empirical check is targeted and convincing at the present statistics. A benchmark with the proven equivariant algorithm would settle the residual doubt, but it is a strengthening rather than a prerequisite for the stated conclusion. The two-temperature test is limited (no continuum extrapolation, no transition scan), but the theoretical constraint is parameter-free and the observed low-T compliance and high-T violation are exactly as predicted. I therefore do not change the reader's ACCEPT verdict.","tokens_in":34476,"tokens_out":14365,"duration_ms":148024,"concrete_test":"Implement the exactly equivariant basic steepest-ascent algorithm of App. D on the same 64^3 x 6 and 64^3 x 8 ensembles and recompute the normalized temporal link average Eq. (83). If the low-temperature result still matches diag(e^{-i2pi/(3L4)}, e^{i2pi/(3L4)}, 1) and the high-temperature result still deviates, the unproven FASD equivariance is not load-bearing. In parallel, quantify FASD equivariance directly: for pairs of starting configurations related by tilde g, run FASD on both and compare the outputs; the average distance should vanish with increasing statistics below T_c.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim, Eq. (83), follows from Eq. (60) only if the gauge-fixed ensemble is approximately invariant under the center transformation tilde g in the confined phase (Sec. IV D). The paper proves this invariance for the basic steepest-ascent algorithm (App. D), but the production runs use Fourier accelerated steepest descent (Sec. V D), for which no equivariance proof is given. The 'grouping' argument in Sec. IV E is heuristic: a non-equivariant copy-selection map can in principle destroy the symmetry even if the unfixed ensemble is symmetric. The numerical checks are direct and encouraging—Fig. 4 shows that E_gf and its tilde-g transforms overlap below T_c, and Sec. VI C shows the diagonal-component relations (77)-(79) are satisfied—but they use only 100 configurations at a single lattice spacing and test the same symmetry constraints that Eq. (83) predicts. They therefore do not fully exclude a copy-selection bias of FASD that artificially produces the low-temperature pattern. If such a bias existed, the identification of deviations from Eq. (83) with deconfinement would not be established for the algorithm actually used.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a lattice implementation of the center-symmetric Landau gauge for SU(3) pure gauge theory at finite temperature. It derives symmetry constraints for the temporal link average; in particular, Eq. (83) states that in the confined phase the ratio M = <U4>/(det<U4>)^{1/3} must equal the fixed diagonal matrix diag(e^{-i2π/(3L4)}, e^{i2π/(3L4)}, 1). The paper tests this prediction in Monte Carlo simulations at beta=6.0 on 64^3 x 8 (below Tc) and 64^3 x 6 (above Tc) lattices, finding agreement with the predicted pattern below Tc and violations above Tc. The authors also discuss the role of Gribov copies, show that a basic steepest-ascent algorithm preserves the center transformation, and provide numerical evidence that the production algorithm approximately does so.","tokens_in":34685,"tokens_out":7026,"duration_ms":70231,"significance":"If the central claim holds, this is a genuine advance: it provides a local, gauge-dependent order parameter for center symmetry that can be computed with standard lattice techniques and that may be useful in continuum frameworks where the Polyakov loop is difficult to handle. The derivation of Eq. (83) is elegant, and the predicted phase pattern exp(±i 2π/(3L4)) is parameter-free and falsifiable. The paper is transparent about the Gribov-copy caveats and includes an explicit equivariance proof for the basic algorithm (App. D). The numerical test is a direct and encouraging confrontation of the theory with data, albeit with limited statistics and no continuum extrapolation.","major_comments":[{"comment":"The equivariance proof in App. D covers only the basic steepest-ascent algorithm, whereas the production runs use Fourier accelerated steepest descent as stated in Sec. V D. The grouping argument in Sec. IV E is heuristic and does not exclude the possibility that a non-equivariant copy-selection map artificially produces the symmetric-phase pattern. Since the derivation of Eq. (83) depends on the approximate invariance of the gauge-fixed ensemble under tilde-g, this gap is load-bearing for the numerical verification. I recommend either proving that the production algorithm preserves tilde-g equivariance (or a closely related variant), or providing a dedicated numerical test with more configurations, a second beta value, and a comparison of full configuration-level distributions, not only the averages.","section":"Sec. IV E and App. D"},{"comment":"The numerical evidence is based on 100 configurations per ensemble at a single lattice spacing (beta = 6.0), with no continuum extrapolation. The agreement with Eq. (83) below Tc is encouraging, but the statistical and systematic errors are large enough that the claim that 'any deviation from the RHS signals the dynamical breaking of center-symmetry' is not yet a robust, quantitative statement. A scan over beta or an explicit test of the L4 dependence at fixed T would substantially strengthen the paper.","section":"Sec. V D"}],"minor_comments":[{"comment":"Footnote 13 contains the typo 'Strictily', and Sec. III D contains 'reasonning'; both should be corrected.","section":"Sec. IV D"},{"comment":"Some matrix entries in Eqs. (C2), (C5), and (C6) contain notations like '0.0522(21) i' and '0.02(21)2' that appear to be formatting errors; these should be cleaned up.","section":"Appendix C"},{"comment":"The statement that the LHS of Eq. (83) 'does not require renormalization' would benefit from a one-sentence justification, since the unrenormalized product may have operator-mixing subtleties that are worth addressing.","section":"Sec. V C"},{"comment":"The caption does not define which ensemble is which color; consider adding a legend or explicit color labels.","section":"Fig. 4"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within the scope of the journal and the research is conceptually sound. The main weakness is the mismatch between the proven equivariance (basic algorithm) and the production algorithm used for the numerical results. This is a correctable issue through additional proof or more extensive numerical checks. I do not see any citation or novelty concerns."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"We have a real lattice test of the center-symmetric Landau gauge here, and that is the headline. The continuum proposal in [16,17] is now implemented on the lattice, with explicit symmetry constraints for the link average solved up to a prefactor, and the resulting parameter-free prediction Eq. (83) is checked below and above Tc. At L4=8 the diagonal elements of the normalized link average sit on the predicted phases; at L4=6 they clearly do not. That is a solid, non-circular numerical test, and the derivation is clean.\n\nWhat is new: the lattice formulation itself, the symmetry constraints for links (not just continuum gauge fields), and the first numerical evidence that a local gauge-dependent link average works as an order parameter. The paper also does useful extra work: the twisted-link formulation and the improved estimators for the gluon field, where the NLO and NNLO estimates move from -3.51 toward -4.19 (expected -4π/3). That is a nice consistency check.\n\nSoft spots are mostly about evidence strength, not logic. 100 configurations per ensemble, a single beta, no continuum extrapolation. The link correlator part of the claim is only theoretical here. The one genuinely delicate point is the invariance of the gauge-fixed ensemble under the center transformation in the confined phase. The paper proves equivariance for the basic steepest-ascent algorithm but not for the Fourier accelerated steepest descent used in production; it falls back on numerical comparison. The stress-test concern is legitimate but I think it lands less hard than it might: Fig. 4 directly compares E_gf with its tilde-g transforms below Tc and they overlap, and the diagonal-component relations (77)-(79) are shown to hold. That is a direct check of the needed property, even if it is at 100 configurations and one lattice spacing. The paper is honest about the limitation and does not oversell.\n\nSo the central claim holds up: within this gauge and copy-selection procedure, the temporal link average is a good order parameter. Minor caveats aside, the paper is well written, the citations to the continuum program are appropriate, and the numerical data are consistent with the theory. This deserves a serious referee. My recommendation would be accept after a light revision that acknowledges the equivariance caveat more prominently and ideally adds at least one more beta or more statistics, but neither is a blocker for the central result.","headline":"First lattice test of the center-symmetric Landau gauge: a clean symmetry derivation and a parameter-free two-temperature check that supports the temporal link average as a local order parameter.","tokens_in":35225,"tokens_out":2474,"would_cite":true,"duration_ms":25309,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["11.15.Ha","12.38.Aw","12.38.Gc"],"model":"deepseek-v4-flash","headline":"In the center-symmetric Landau gauge, a single temporal link average becomes a local order parameter: in the confined phase it must equal $\\mathrm{diag}(e^{-i2\\pi/(3L_4)}, e^{i2\\pi/(3L_4)}, 1)$, and any deviation signals deconfinement.","keywords":["center symmetry","confinement-deconfinement transition","center-symmetric Landau gauge","lattice gauge theory","Polyakov loop","order parameter","Gribov copies","SU(3) Yang-Mills"],"falsifier":"Run the gauge fixing at $\\beta=6.0$ on a $64^3\\times 8$ lattice using the basic, provably $\\tilde g$-symmetric steepest-ascent algorithm of Appendix D with high statistics: if the normalized temporal link average deviates from $\\mathrm{diag}(e^{-i2\\pi/(3L_4)}, e^{i2\\pi/(3L_4)}, 1)$ beyond errors, the claimed constraint (60) does not force Eq. (83), and the central claim collapses. Conversely, the same run above $T_c$ ($L_4=6$) must show the deviation in one of the three center sectors; if the constraint still holds there, the link average is not tracking center-symmetry breaking. An independent check is a direct overlap measurement of the ensembles $E_{gf}$ and $E^{\\tilde g}_{gf}$ below $T_c$: if they are not approximately equal, the derivation of the order-parameter property fails at its stated premise.","tokens_in":34286,"feed_emoji":"🔗","tokens_out":14130,"duration_ms":112102,"temperature":0.7,"pith_summary":"This paper puts the continuum center-symmetric Landau gauge onto the lattice and shows that, in that gauge, the average of a single temporal link — a local quantity — is an order parameter for the confinement-deconfinement transition of SU(3) Yang-Mills theory. In the confined phase, an unbroken center symmetry forces the normalized temporal link average to equal the fixed matrix $\\mathrm{diag}(e^{-i2\\pi/(3L_4)}, \\, e^{i2\\pi/(3L_4)}, \\, 1)$, so that any statistically significant deviation is a direct signal of deconfinement. This matters because the standard order parameter, the Polyakov loop, is nonlocal and awkward in continuum calculations, whereas the link average is local and renormalization-free. The subtlety, examined at length, is that this property holds only when both the gauge and the selection of Gribov copies keep the gauge-fixed ensemble invariant under the center transformation; Monte Carlo data at $\\beta=6.0$ comply below $T_c$ ($64^3\\times 8$) and show the predicted violation above $T_c$ ($64^3\\times 6$).","feed_headline":"Deconfinement read off a single lattice link average","feed_subtitle":"A local gauge-fixed observable does the work of the nonlocal Polyakov loop, confirmed below and above Tc.","key_machinery":"The load-bearing object is the gauge-fixing functional $\\tilde F[U]=\\sum_{n,\\mu}\\mathrm{Re}\\,\\mathrm{tr}\\, g_c^\\dagger(\\hat\\mu)U_\\mu(n)$ with the fixed background $g_c(\\hat\\mu)=e^{-i(4\\pi/3)(\\hat\\mu_4/L_4)\\lambda_3/2}$, together with the center transformation $\\tilde g(n)=e^{i\\pi\\lambda_4/2}e^{i\\pi\\lambda_1/2}e^{-i(n_4/L_4)\\pi(\\lambda_3+\\lambda_8/\\sqrt3)}$ that leaves it invariant. Because links transform linearly under $\\tilde g$, the average obeys the exact identity (41) linking the ensembles $E_{gf}$ and $E^{\\tilde g}_{gf}$; when those ensembles coincide approximately, that identity becomes a constraint fixing the matrix structure of $\\langle U_\\mu(n)\\rangle$ down to a single real number $\\eta$. The constraint is solved by decomposing the average along the weight vectors of $\\mathrm{SU}(3)$ and using the Weyl-reflection form of $\\tilde g$; the result is $\\langle U_\\mu(n)\\rangle=\\eta\\,g_c(\\hat\\mu)$, whose temporal component yields Eq. (83). The paper proves in Appendix D that the basic steepest-ascent algorithm preserves the required ensemble invariance, and relies on numerical checks for the accelerated production algorithm.","core_discovery":"The central claim, stated on the paper's own terms, is that in the center-symmetric Landau gauge the link average $\\langle U_\\mu(n)\\rangle$ is a local order parameter for center symmetry. The gauge is defined by maximizing $\\tilde F[U]=\\sum_{n,\\mu}\\mathrm{Re}\\,\\mathrm{tr}\\, g_c^\\dagger(\\hat\\mu)U_\\mu(n)$ with a fixed center-symmetric background $g_c(\\hat\\mu)=e^{-i(4\\pi/3)(\\hat\\mu_4/L_4)\\lambda_3/2}$, a functional left invariant by the specific center transformation $\\tilde g$ of Eq. (49). In the symmetric phase, where the gauge-fixed ensemble obeys $E_{gf}^{\\tilde g}\\simeq E_{gf}$, the transformation law (41) becomes the constraint $\\langle U_\\mu(n)\\rangle = \\tilde g(n)\\langle U_\\mu(n)\\rangle \\tilde g^\\dagger(n+\\hat\\mu)$, and combining it with global color and charge-conjugation invariance yields the unique solution $\\langle U_\\mu(n)\\rangle = \\eta\\, g_c(\\hat\\mu)$ with $\\eta$ real. For temporal links this is Eq. (83): the average normalized by its determinant equals $\\mathrm{diag}(e^{-i2\\pi/(3L_4)}, e^{i2\\pi/(3L_4)}, 1)$. The numerical data support the prediction: below $T_c$ the normalized average matches the required matrix, while above $T_c$ it deviates in three distinct ways that correspond to the three center sectors of the broken phase.","pith_inferences":["If the link average is a true order parameter, a fine scan of the normalized average across $T_c$ (several $L_4$ values at fixed $\\beta$) should show it locking onto the Eq. (83) form exactly in the confined phase and leaving it sharply at $T_c$; the paper's two-point comparison ($L_4=6$ vs $8$) does not yet establish the sharpness of the transition this observable sees.","The empirical unitarity of $M$ in the deconfined phase, if generic, suggests the vanishing of color-off-diagonal components of the link average is stronger than statistical — it may follow from the combined color-and-$\\hat C$ symmetry structure, a line the paper leaves open.","A natural extension the authors mention but do not pursue on the lattice: with dynamical quarks, center symmetry is explicitly broken and the transition becomes a crossover; a local order-parameter-like quantity with a well-defined crossover temperature could be useful in finite-density simulations where the Polyakov loop is sign-problematic.","The twisted-link formulation of Sec. VI A implies a particularly clean test: in the confined phase $\\langle\\hat U_\\mu(n)\\rangle = \\eta\\,\\mathbb{1}$, so components such as $\\mathrm{Im}\\langle\\hat U_4^{3}\\rangle$ and $\\mathrm{Re}\\langle\\hat U_4^{8}\\rangle$ vanish; counting how often individual configurations deviate from this pattern gives a per-configuration order-parameter diagnostic that the pape"],"forward_implications":["The normalized temporal link average of Eq. (83) supplies a local, renormalization-free proxy for the Polyakov loop: any deviation of the matrix elements from $e^{\\pm i2\\pi/(3L_4)}$ and $1$ signals deconfinement, observable on a single site or averaged over the lattice.","The same mechanism extends to two-link correlators and, in the continuum limit, to the gluon and ghost propagators in this gauge, giving momentum-space quantities that behave as order parameters — quantities continuum methods can compute.","Each center sector of the deconfined phase is tagged by which diagonal element of the link average sits on $e^{i2\\pi k/3}$, so the gauge-fixed link average labels the broken-symmetry ensembles just as the Polyakov loop phase does.","The data reveal an unexplained fact: in the deconfined phase the normalized link average $M$ is unitary, since one diagonal element is of modulus one and charge conjugation forces the other two to have equal modulus; the paper derives $\\det M=1$ and $|M_{11}|=|M_{22}|=|M_{33}|=1$ from this, but not the origin of the modulus-one element."],"supporting_citations":[{"why":"Introduces the continuum center-symmetric Landau gauge and the gauge-field average as an order parameter; the lattice functional (47) is inferred from it.","marker":"[16]"},{"why":"Proves generally that averages and correlators in center-symmetric gauges are local order parameters, and supplies the Weyl-transformation identities used to solve the constraint (60).","marker":"[17]"},{"why":"Prior continuum test of two-point correlators as order parameters within the Curci-Ferrari model, the expectation this paper tests on the lattice.","marker":"[18]"},{"why":"Proposal to use a fixed center-symmetric background, with [16], as the basis of the gauge choice.","marker":"[7]"},{"why":"Establishes order parameters from Landau-gauge propagators over distinct center sectors; the complementary approach the paper contrasts with its own.","marker":"[31]"},{"why":"The simulation toolkit used for the gauge-fixing runs that produce the numerical confirmation.","marker":"[33]"}],"fun_headline_variants":["Local link average detects deconfinement without Polyakov loop","Center-symmetric gauge yields local order parameter for confinement","Single lattice link average replaces Polyakov loop as transition probe","Gauge-fixed link average: a local stand-in for Polyakov loop"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The order-parameter property rests on the gauge-fixed ensemble being approximately invariant under the center transformation $\\tilde g$ in the confined phase; this is proven only for the basic steepest-ascent algorithm, while the production algorithm inherits it by assumption, supported by numerical checks.","fun_headline_variants_meta":{"raw":{"variants":["Local link average detects deconfinement without Polyakov loop","Center-symmetric gauge yields local order parameter for confinement","Single lattice link average replaces Polyakov loop as transition probe","Gauge-fixed link average: a local stand-in for Polyakov loop"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000611,"raw_usage":{"total_tokens":2859,"prompt_tokens":977,"completion_tokens":1882,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":593,"completion_tokens_details":{"reasoning_tokens":1811}},"tokens_in":593,"tokens_out":1882,"duration_ms":14251,"temperature":1.0,"reasoning_tokens":1811,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T18:24:38.500196+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the gauge fixing at $\\beta=6.0$ on a $64^3\\times 8$ lattice using the basic, provably $\\tilde g$-symmetric steepest-ascent algorithm of Appendix D with high statistics: if the normalized temporal link average deviates from $\\mathrm{diag}(e^{-i2\\pi/(3L_4)}, e^{i2\\pi/(3L_4)}, 1)$ beyond errors, the claimed constraint (60) does not force Eq. (83), and the central claim collapses. Conversely, the same run above $T_c$ ($L_4=6$) must show the deviation in one of the three center sectors; if the constraint still holds there, the link average is not tracking center-symmetry breaking. An independent check is a direct overlap measurement of the ensembles $E_{gf}$ and $E^{\\tilde g}_{gf}$ below $T_c$: if they are not approximately equal, the derivation of the order-parameter property fails at its stated premise.","supporting_citations":[{"cited_title":"Fister and J","cited_arxiv_id":null,"evidence_quote":"Introduces the continuum center-symmetric Landau gauge and the gauge-field average as an order parameter; the lattice functional (47) is inferred from it."},{"cited_title":"Perturbative aspects of the deconfinement transition – Physics beyond the Faddeev-Popov model,","cited_arxiv_id":null,"evidence_quote":"Proves generally that averages and correlators in center-symmetric gauges are local order parameters, and supplies the Weyl-transformation identities used to solve the constraint (60)."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Prior continuum test of two-point correlators as order parameters within the Curci-Ferrari model, the expectation this paper tests on the lattice."},{"cited_title":"Celik, J","cited_arxiv_id":null,"evidence_quote":"Proposal to use a fixed center-symmetric background, with [16], as the basis of the gauge choice."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The simulation toolkit used for the gauge-fixing runs that produce the numerical confirmation."}],"review_version":1}