{"id":"10cc821f-dd12-4128-9405-6d1afb2edc8d","arxiv_id":"2412.10306","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A Metropolis Monte Carlo algorithm is shown analytically to reproduce stochastic relativistic viscous hydrodynamics in the Density Frame, with extension to Bjorken and general coordinates.","lead":"Physicists propose a new way to simulate random, fluctuating motion in hot nuclear matter by alternating steps of ideal fluid flow with random momentum swaps between neighboring cells, accepting or rejecting each swap using entropy. The method targets relativistic heavy-ion collisions, where small particle numbers make thermal fluctuations important.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central claim is not yet secured: Sec. IV derives the mean stress from a single corner update and asserts that the full corner loop reproduces the continuum stochastic equations, but neither the convergence of the sequential update nor the noise correlations of Eq.","rationale":"The paper's analytic work is substantial and internally coherent: the Density Frame constitutive relation is derived from both frame transformations and kinetic theory, and the Metropolis mean-stress argument is plausible. The reader's weakest-assumption analysis identified the discrete-to-continuum convergence of the corner-based update as the key unproven step, and I agree. The central claim has two ingredients, a mean viscous stress and a stochastic noise with FDT correlations; only the first is derived, and even that derivation stops at one corner plus an assertion about the rest of the loop. The noise is inherited from the proposal distribution, but the acceptance step and the sequential multi-corner sweep could modify the effective noise, and this is not checked. Because the paper explicitly disclaims numerical implementation, the stochastic central claim is currently unsupported by simulation. This does not make the claim wrong, but it makes the conditional verdict appropriate: the derivation should be followed by a numerical test of the mean stress, the noise covariance, and convergence in Delta t and V0.","tokens_in":21967,"tokens_out":21492,"duration_ms":199959,"concrete_test":"Implement the algorithm in 2+1D Cartesian coordinates with periodic boundary conditions and a uniform equilibrium background. Measure the covariance of the accepted momentum transfers across a given cell face over many Metropolis steps and check that it equals 2T kappa^{ijmn}/(Delta t V0) with kappa from Eq. (55), and that the mean transfer vanishes. Then initialize a small-amplitude shear perturbation and compare the measured damping rate with the linearized Density Frame dispersion relation, while also checking that the noise variance is unchanged by the acceptance step. Repeat for several values of Delta t and cell size to confirm first-order convergence. A mismatch by a factor of two or a nonzero spurious drift would indicate that the corner-splitting factors in Eqs. (102)-(104) or the sequential corner loop do not reproduce Eq. (90).","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's strongest claim is that the Metropolis algorithm reproduces stochastic relativistic viscous hydrodynamics in the Density Frame, including both the mean stress and the FDT noise of Eq. (90). The analytic derivation of the mean stress is careful, but the passage from a single corner update to the full discrete evolution contains a load-bearing gap. In Sec. IV A, Eqs. (107)-(109) compute the accepted-mean stress from one corner, and Eq. (110) states that this yields \"exactly half\" of the cell update, with the remaining half coming from another corner; the paper then asserts that looping over all corners reproduces the divergence structure of the continuum viscous stress. For general coordinates, Sec. IV B and Appendix A make the analogous assertion after introducing parallel transport and covariant derivative approximations (Eqs. (128)-(131), (A33)-(A36)). What is not shown is that these discrete updates converge to the continuum stochastic PDE: that the sequential corner updates have no additional bias or cross-correlations at finite lattice spacing, that the discrete covariant derivatives in Eqs. (106) and (129) give the correct continuum divergence, and that the accepted-noise variance equals Eq. (92) after all corners are visited. The paper explicitly states it has not implemented the algorithm numerically (Sec. I B and Sec. V), so the stochastic component of the central claim rests on an unverified discretization assumption. The companion paper [43] tests only deterministic Density Frame hydrodynamics in 1+1D, not the stochastic Metropolis algorithm.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript proposes a Metropolis algorithm for stochastic relativistic viscous hydrodynamics in the Density Frame. Each step combines an ideal hydrodynamic update with random spatial momentum transfers between fluid cells, accepted or rejected according to the entropy change. The paper derives the Density Frame constitutive relations, evaluates the shear and bulk noise kernels κ^{ijmn}, shows how the Density Frame emerges from relativistic kinetic theory, and extends the algorithm to general coordinates including Bjorken coordinates. The central claim is that the full algorithm reproduces the Density Frame mean viscous stress −T κ^{ijmn} ∇_{(m} β_{n)} and the fluctuation-dissipation noise with variance 2T κ^{ijmn}. The paper explicitly states that the algorithm has not yet been implemented numerically.","tokens_in":22199,"tokens_out":6518,"duration_ms":63404,"significance":"The analytic Density Frame machinery is a valuable contribution: it provides a first-order-in-time relativistic viscous formulation with no additional dynamical fields and only shear viscosity, bulk viscosity, and the equation of state as input. The derivation of κ^{ijmn} from the Landau frame, Eqs. (38) and (54)-(58), is careful and is checked against Armas and Jain. The proposed Metropolis scheme is conceptually elegant and, if correct, would be practically important for heavy-ion and critical-point simulations. However, the headline claim that the algorithm reproduces stochastic viscous hydrodynamics is supported only by a single-corner analytic calculation and is not demonstrated numerically, so the significance is conditional on closing that gap.","major_comments":[{"comment":"The mean-stress derivation is performed for one corner update: Eq. (107) computes the accepted mean of ξ from a single corner, and Eq. (110) states that this gives exactly half of the cell momentum update, with the remaining half coming from another corner. The paper then asserts, without proof, that looping over all corners with a shuffled order reproduces the full discrete update (99) and hence the continuum divergence structure (98). What is missing is a demonstration that the sequential corner updates do not introduce bias or cross-correlations at finite lattice spacing, that the stencil (106) approximates ∂_{(m} β_{n)} to the required order, and that the factors of 1/2 accumulate to the correct total update. This is load-bearing because the central claim is that the full algorithm, not a single corner, reproduces stochastic viscous hydrodynamics.","section":"§IV.A, Eqs. (107)-(110)"},{"comment":"The noise part of the claim is not verified. Eq. (102) sets the proposal variance to 2T κ^{ijmn}/(Δt V0), and the paper states that the accepted noise implements the fluctuation-dissipation result (92). However, only the first moment of the accepted proposals is computed (Eqs. (107)-(109)); the second moment of the accepted moves after the accept/reject filter, which involves averages such as ⟨θ(ΔS)ξijξmn⟩ and ⟨θ(−ΔS)e^{2ΔS}ξijξmn⟩, is not evaluated. Consequently the noise variance after the complete corner loop is not shown to equal 2T κ^{ijmn}, even to leading order in Δt. Since the stochastic character of the algorithm is half of the paper's central claim, this gap needs to be closed.","section":"§IV.A, Eqs. (92), (102), (107)-(109)"},{"comment":"The generalization to general coordinates is asserted by the same pattern: the mean stress is computed from one corner (Eqs. (130)-(131) and (A33)-(A34)), and then the mean update of cell A after all eight corners is stated to reproduce Eq. (118) or Eq. (A25). The discrete parallel-transport increments (123)-(124), the discrete covariant-derivative approximation (129), and the factors 1/8 and 1/4 are not shown to be consistent with the continuum equations to the relevant order. In particular, the accumulation or cancellation of the extrinsic-curvature term (125) after visiting all corners should be checked. As in the Cartesian case, an explicit finite-lattice consistency check is required.","section":"§IV.B and Appendix A, Eqs. (123)-(129), (A33)-(A36)"},{"comment":"The paper explicitly states that the Metropolis algorithm has not been implemented (Sec. I.B) and that numerical simulation is left for future work (Sec. V). The companion paper [43] tests only deterministic Density Frame hydrodynamics in 1+1D, and the earlier stochastic study [29] treats advection-diffusion rather than the Navier-Stokes system. Thus the central claim that the proposed algorithm reproduces stochastic relativistic viscous hydrodynamics is not supported by numerical evidence. A minimal test—for example, the decay of linearized sound modes or the static structure factor in a 2+1D box—would substantiate the claim and is within the scope of the present manuscript.","section":"§I.B and §V"}],"minor_comments":[{"comment":"The text after Eq. (A10) says 'capitol letter'; this should be 'capital letter'.","section":"Appendix A.1"},{"comment":"The notation δ_{tt'}δ_{rr'} discretizes the continuum white-noise correlator (92); the paper should state the intended continuum limit and the stochastic convention (Ito versus Stratonovich) used when Δt and the cell volume are sent to zero.","section":"§IV.A, Eq. (102)"},{"comment":"The discrete derivative ∂xβy is introduced without explaining why this particular stencil is the natural approximation of the continuum ∂_{(x} β_{y)}; a short comment would improve readability.","section":"§IV.A, Eq. (106)"},{"comment":"The notation contrasts βi and ⃗βi (e.g., Eqs. (A6) and (A32)); a table or a sentence defining both consistently would help the reader.","section":"Appendix A.1"},{"comment":"Reference [21] gives only the journal and year; full page numbers would be useful for the reader.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"To the editor: The paper is honest about not implementing the algorithm, and the companion deterministic paper is a useful complement. My main concern is that the discrete corner-loop convergence is asserted rather than demonstrated; this is fixable but needs either a proof or a numerical test. The manuscript fits the journal's scope, and the analytic parts are sound."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a solid theory paper that extends the Metropolis approach to full relativistic viscous hydrodynamics in the Density Frame and to general coordinates, with a clean analytic derivation of the mean stress and noise. The catch is exactly what the authors admit: they have not implemented the algorithm numerically, and the step from a single corner update to the full lattice loop is asserted rather than proved. The central claim should be read as a well-motivated proposal, not a demonstrated result.\n\nWhat's actually new: prior Metropolis-based stochastic fluid work (including the authors' own advection-diffusion paper and the non-expanding critical-point simulations) are simpler. Here they handle the full Navier-Stokes system in the Density Frame, derive the explicit shear and bulk viscous tensors from the Landau frame, show how the Density Frame arises from kinetic theory, and extend the algorithm to Bjorken/general coordinates via parallel transport. The derivation of the mean accepted stress in Sec. IV A is internally consistent: expand exp(ΔS) to first order, use the symmetry of the proposal distribution, and out pops the Density Frame stress and FDT noise. The general-coordinate version is more involved but the logic is the same.\n\nSoft spots: the biggest is the unverified discretization. Equations (107)-(109) compute the mean stress from one corner and state the rest comes from the opposite corner; the paper then asserts that looping over all corners gives the continuum divergence structure and noise correlations. That is a load-bearing assumption. At finite lattice spacing there could be bias or cross-correlations between sequential corner updates, and the discrete covariant derivatives in Eqs. (106) and (129) are not shown to converge to the continuum divergences. The paper is upfront about having no numerical implementation, and the companion paper only tests deterministic Density Frame hydrodynamics in 1+1D. So the stochastic claim rests on analytic argument plus hope. That doesn't kill the paper—it's a theory proposal—but it does mean the abstract's 'reproduces' is stronger than what is demonstrated.\n\nAlso minor: the paper's own framing in Sec. V says 'we have stopped short of actually simulating,' which is honest. The references to their own prior work are appropriate, not padding.\n\nWho this is for: someone working on stochastic hydrodynamics for small systems or critical-point searches will find the Density Frame derivation and the algorithm proposal useful, especially the general-coordinate extension. For a referee: yes, send it out. The derivation deserves careful checking, and a referee should push for either a numerical test in revision or a clear statement that the convergence is an open question. My own verdict would be conditional acceptance.","headline":"Clean analytic proposal for Metropolis-based stochastic relativistic viscous hydrodynamics in the Density Frame, honestly flagged as numerically untested—worth refereeing, but the abstract's 'reproduces' overstates what is shown.","tokens_in":22764,"tokens_out":3874,"would_cite":true,"duration_ms":572284,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["47.75.+f","05.40.-a","25.75.-q"],"model":"deepseek-v4-flash","headline":"Each entropy-weighted momentum swap between fluid cells reproduces stochastic relativistic viscous hydrodynamics.","keywords":["stochastic relativistic hydrodynamics","Metropolis algorithm","Density Frame","viscous hydrodynamics","fluctuation-dissipation theorem","Bjorken coordinates","first-order hydrodynamics","heavy-ion collisions"],"falsifier":"Implement the algorithm in 2+1 dimensions with a known equation of state and viscosities, and measure the mean stress and noise correlations in a controlled shear flow such as a sinusoidal velocity profile. If the measured mean stress deviates from $-T \\kappa^{ijmn}\\partial_{(m}\\beta_{n)}$ at order $\\Delta x$, or the noise variance is not $2T \\kappa^{ijmn}$ per cell, then the Metropolis update does not yield exactly the claimed stochastic hydrodynamics on a finite lattice.","tokens_in":21737,"feed_emoji":"🎲","tokens_out":5541,"duration_ms":51208,"temperature":0.7,"pith_summary":"The paper proposes a simulation scheme for relativistic fluids that keeps the thermal fluctuations ordinary viscous hydrodynamics discards. In each update the fluid first takes one ideal (non-viscous) step, then random amounts of spatial momentum are proposed between neighboring cells; each proposal is accepted or rejected using the entropy change as a statistical weight. The paper argues that the average effect of these swaps is exactly the viscous stress of the Density Frame, a first-order-in-time formulation of relativistic hydrodynamics, and that the accept/reject scatter automatically supplies the fluctuation-dissipation noise thermodynamics requires. If correct, the scheme offers a practical way to simulate stochastic fluid dynamics in heavy-ion collisions, including small systems and near-critical fluctuations, using only the shear viscosity, bulk viscosity, and equation of state.","feed_headline":"Random momentum swaps reproduce noisy relativistic hydrodynamics","feed_subtitle":"A Metropolis accept-reject step yields viscous stress and thermal noise from just shear and bulk viscosity plus the equation of state.","key_machinery":"The central object is the Density Frame constitutive relation, with viscous stress $\\Pi_{ij} = -T \\kappa^{ijmn} \\partial_{(m} \\beta_{n)}$ and its covariant generalization, where the same tensor $\\kappa^{ijmn}$ fixes both the mean dissipative stress and the variance of the proposed noise. The argument's engine is the Metropolis acceptance rule: proposals with positive entropy change are always kept, those with negative entropy change are kept with probability $e^{\\Delta S}$; since $\\Delta S = -\\Delta t\\, V_0\\, \\xi_{ij}\\partial_{(i}\\beta_{j)}$, the accept/reject bias gives the proposed noise a nonzero mean equal to the viscous stress while the remaining scatter is exactly the fluctuation-dissipation noise. The paper also derives $\\kappa^{ijmn}$ from the Landau frame by shifting $\\beta^\\mu$ using susceptibility tensors, and shows that the same tensor follows from relativistic kinetic theory with a shifted viscous correction $\\delta f$.","core_discovery":"The central claim is that each Metropolis step, consisting of an ideal hydrodynamic update followed by proposed random spatial momentum transfers accepted with the entropy weight $e^{\\Delta S}$, reproduces stochastic relativistic viscous hydrodynamics in the Density Frame. Concretely, the accepted transfers produce a mean viscous stress $\\langle \\xi_{ij}\\rangle = -T \\kappa^{ijmn} \\nabla_{(m} \\beta_{n)}$ and noise variance $2T \\kappa^{ijmn} \\delta(t-t')\\delta^2(r-r')$, where $\\kappa^{ijmn}$ is the Density Frame viscosity and noise tensor built from the shear and bulk viscosities and the equation of state. Because the Density Frame evolution equations are first order in time and involve only the energy and momentum densities on a single spatial slice, no auxiliary dynamical fields such as $\\Pi^{\\mu\\nu}$ are needed. The paper extends the construction to general coordinates through the $3+1$ split of general relativity; the only nontrivial complication is that proposed momentum transfers must be parallel transported from cell faces to cell centers, which generates the energy-changing viscous work terms that appear, for example, in Bjorken flow.","pith_inferences":["Inference: because the algorithm is local and needs no extra parameters, it is a plausible basis for coupling stochastic fluid dynamics to critical-point models; the paper notes this possibility but does not implement it.","Inference: the mean stress is derived at first order in $\\Delta S$ from a single lattice corner, so finite-lattice corrections might enter at order $\\Delta x$; a numerical convergence test measuring stress and noise across resolutions would sharpen the regime of validity.","Inference: the same accept/reject construction could apply to other conserved-charge diffusion problems, not only energy and momentum, whenever a Density-Frame-type gradient expansion holds."],"forward_implications":["Stochastic viscous relativistic hydrodynamics can be simulated with only the shear viscosity, bulk viscosity, and equation of state; no relaxation times or additional dynamical fields are required.","The algorithm satisfies the fluctuation-dissipation theorem by construction, because the same entropy weight produces both dissipation and noise.","The scheme extends to Bjorken and general curved coordinates via the $3+1$ split; parallel transport of momentum transfers yields the energy-changing viscous terms of an expanding system.","Combined with ideal hydrodynamics steps, the Metropolis update gives a first-order-in-time evolution suitable for small collision systems and near-critical dynamics where $1/N$ fluctuations matter.","The paper itself does not implement the algorithm numerically, but it provides the formulation and points to a companion numerical study of the deterministic Density Frame dynamics as the next step."],"supporting_citations":[{"why":"Supplies the Density Frame formulation of viscous hydrodynamics whose constitutive relation and noise tensor the algorithm reproduces.","marker":"[42]"},{"why":"The authors' previous Metropolis algorithm for the relativistic stochastic advection-diffusion equation, which the present momentum-swap update directly generalizes.","marker":"[29]"},{"why":"A prior stochastic-fluid simulation near a critical point using the same Metropolis principle, anchoring the approach in existing practice.","marker":"[28]"},{"why":"Companion numerical study of deterministic Density Frame hydrodynamics that supports the practicality of the formulation.","marker":"[43]"},{"why":"Landau-Lifshitz constitutive relation from which the Density Frame viscous tensor is obtained by a frame shift.","marker":"[47]"},{"why":"Provides considerations on hydrodynamics without an entropy current and the integrability structure used in deriving frame transformations.","marker":"[48]"}],"fun_headline_variants":["Metropolis swaps reproduce viscous fluid noise","Entropy-accepted transfers yield viscous hydrodynamics","Viscous hydrodynamics sans auxiliary fields via Metropolis","Stochastic fluid dynamics with only viscosity and EoS","Metropolis algorithm simulates viscous relativistic flows"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the discrete, corner-by-corner Metropolis update, whose mean stress is computed from a single corner at first order in the entropy change, reproduces the exact divergence structure and noise correlations of the continuum Density Frame equations once all corners are visited; the paper argues this but does not prove convergence at finite lattice spacing.","fun_headline_variants_meta":{"raw":{"variants":["Metropolis swaps reproduce viscous fluid noise","Entropy-accepted transfers yield viscous hydrodynamics","Viscous hydrodynamics sans auxiliary fields via Metropolis","Stochastic fluid dynamics with only viscosity and EoS","Metropolis algorithm simulates viscous relativistic flows"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000179,"raw_usage":{"total_tokens":1290,"prompt_tokens":927,"completion_tokens":363,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":543,"completion_tokens_details":{"reasoning_tokens":292}},"tokens_in":543,"tokens_out":363,"duration_ms":4187,"temperature":1.0,"reasoning_tokens":292,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T15:58:02.210419+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Implement the algorithm in 2+1 dimensions with a known equation of state and viscosities, and measure the mean stress and noise correlations in a controlled shear flow such as a sinusoidal velocity profile. If the measured mean stress deviates from $-T \\kappa^{ijmn}\\partial_{(m}\\beta_{n)}$ at order $\\Delta x$, or the noise variance is not $2T \\kappa^{ijmn}$ per cell, then the Metropolis update does not yield exactly the claimed stochastic hydrodynamics on a finite lattice.","supporting_citations":[],"review_version":1}