{"id":"48a7fccc-9dd9-4c67-9a57-41db201ffd40","arxiv_id":"2607.18207","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"The fluid-Ansatz and WallGo methods agree for mild phase transitions but diverge by tens of percent for strong ones, where the semi-classical approximation itself may break down.","lead":"Two standard computer methods for predicting how fast vacuum bubbles expand after a cosmological phase transition agree for weak transitions but diverge badly for strong ones. The finding matters because future gravitational-wave observatories such as LISA target exactly those strong transitions.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"O(δ²) claim is computed at κ=1 (3 moments) while the linear benchmarks it is meant to explain use κ=3 (10 moments); no check that the quadratic shift survives at the converged truncation.","rationale":"The paper is best read as a methodological benchmark: it compares two ways of solving the same semi-classical Boltzmann equation. The weak-regime agreement (α≲0.01) is well supported by internal consistency checks, including the equilibrium-only friction curves in Fig. 3, the fluid-order convergence in Fig. 4, and the WallGo grid-resolution scan in Sec. 4.4. The reader's chosen weakest assumption—the shared leading-log collision operator—is acknowledged by the authors and affects the physical interpretation, but it does not undermine the internal comparison of the two methods. The genuinely load-bearing weak point is the nonlinear O(δ²) analysis: it uses only a 3-moment truncation while the linear comparisons that motivate it use 10 moments. Since the linear solution is shown to be sensitive to truncation order up to D=10, the strong-regime nonlinear shift may be an artifact of the truncated basis rather than evidence about the converged fluid Ansatz. This does not invalidate the weak-regime benchmark, but it should be made a necessary condition for the strong-regime conclusions. The reader's CONDITIONAL verdict remains appropriate, so I leave it unchanged.","tokens_in":24467,"tokens_out":7410,"duration_ms":73450,"concrete_test":"Extend the nonlinear system of Secs. 5.2–5.3 from κ=1 (3 moments) to κ=2 (6 moments) and κ=3 (10 moments), retaining all quadratic kinetic and collision projections (eq. 43), and recompute the SMEFT v_w(M) curve of Fig. 6. If the O(δ²)-induced shift relative to the corresponding linear κ=3 result shrinks to <5% or reverses sign for α>0.1, the strong-regime conclusion is a low-order artifact; if the shift remains ~20–40% at the higher truncation, the claim survives.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing step is the nonlinear extension in Sec. 5. Equation (35) defines O(δ²) corrections, but Sec. 5.1 immediately truncates the momentum expansion at first order, reducing the fluctuation to δ = δμ + p^μ(...), i.e. only D=3 moments (eqs. 37–40). The entire O(δ²) analysis of Fig. 6 therefore compares a 3-moment linear solution with a 3-moment quadratic solution. The linear benchmark that produced the claimed 40–60% discrepancy in Figs. 1–3 used order 3 / D=10, and Fig. 4 shows the linear result changes significantly with truncation order up to D=10. Hence the observed nonlinear v_w shift could be an artifact of the poor low-order linear truncation rather than a property of the converged fluid Ansatz. This is especially important because the direct O(δ²) pressure ratio remains only ~4×10^-2 (Fig. 6), so the large velocity shift is entirely a backreaction of the truncated 3-moment system. The paper does not report a convergence study of the nonlinear system in κ. The reader's concern about shared leading-log collision uncertainties is real but external to the benchmark; the truncation issue threatens the paper's strong-regime interpretation directly.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper benchmarks two methods for computing the terminal bubble wall velocity in first-order cosmological phase transitions: the extended fluid Ansatz and the Chebyshev spectral method implemented in the public code WallGo. The comparison is performed in two models (SMEFT with a φ⁶ operator and a light-Higgs SM-like model) and at several truncation orders of the fluid momentum expansion. The linearized analysis finds good agreement between the two approaches for weak transitions, α≲0.01, especially when only top-quark annihilation is included, and finds discrepancies reaching O(40–60%) for α≳0.1. The paper then extends the fluid Ansatz to second order in the fluctuations and reports significant shifts in v_w for strong transitions even though the direct O(δ²) pressure contribution remains at the few-percent level. The authors interpret this as a breakdown of the linearized fluid Ansatz in the strong-transition regime and note that this regime also approaches the limits of the WKB approximation underlying the Boltzmann treatment.","tokens_in":24801,"tokens_out":4953,"duration_ms":49247,"significance":"If the main claims hold, the paper provides a useful benchmark of two widely used approaches and clarifies the regime in which their predictions can be trusted. The authors are careful in several respects: the same scattering amplitudes are used in both codes, the fluid Ansatz is tested at multiple truncation orders, an equilibrium-only sanity check is performed, and the leading-log collision model is presented explicitly. The conclusion that the two methods agree precisely where the WKB approximation is comfortable but disagree where they are most needed for gravitational-wave predictions is important and, if supported, would sharpen the case for going beyond current semi-classical methods. However, the nonlinear O(δ²) analysis is not carried out at the same truncation order as the linear benchmarks, which leaves the strong-regime conclusion unsupported as presented.","major_comments":[{"comment":"The O(δ²) analysis is performed with the momentum expansion truncated at order κ=1 (D=3 moments), as stated after Eq. (37) and implemented in Eqs. (38)–(40). The linear benchmarks that establish the 40–60% discrepancy and the convergence behaviour use κ=3 (D=10) — see §4.3 and Fig. 4, where the linear result is shown to change appreciably with truncation order up to D=10. The paper does not report a κ-convergence study for the nonlinear system. The large v_w shift in Fig. 6 may therefore be an artifact of comparing a 3-moment linear baseline with a 3-moment quadratic system rather than a property of the converged fluid Ansatz. This is especially concerning because the direct O(δ²) pressure ratio in Fig. 6 is only ~4×10⁻², so the velocity shift is dominated by the backreaction of the truncated 3-moment system. The strong-transition conclusion requires the nonlinear computation to be repea","section":"§5.1 and Fig. 6"},{"comment":"Both codes use the same leading-log collision operator with thermal-mass-regulated t- and u-channel amplitudes, Eqs. (65)–(67), and the paper cites ref. [30] for O(1) theoretical uncertainties in these collision terms. Because the two approaches share the same collision model, their agreement for α≲0.01 validates the internal consistency of the two implementations but does not validate the underlying physics: if the true collision rates differ from the leading-log model, both methods share the same systematic error. The paper acknowledges this only in the final paragraph of the conclusions. This caveat should appear at the point where the agreement is claimed (e.g., in §4.3 or the abstract), and ideally be complemented by a sensitivity test with respect to the thermal-mass regularization or collision-model parameters.","section":"§3.1, Appendix A.3, Conclusions"},{"comment":"The main comparison figures (Figs. 1–3) show single fluid-Ansatz curves without uncertainty bands. Section 4.4 demonstrates convergence for the fluid Ansatz and discusses WallGo's grid dependence, but the relative-error curves in Figs. 1–3 are computed pointwise with respect to a single WallGo configuration, and the text itself notes in §4.4 that WallGo's nominal error bars are not reliable predictors of convergence. As a result, the central quantitative claims — 'essentially the same wall velocity' at α≲0.01 and 'O(40%–60%) discrepancy' at α≳0.1 — are stated without propagated uncertainties. At minimum, the figures should display the truncation sensitivity of the order-3 fluid curves and the WallGo grid-convergence envelope.","section":"§4.3, §4.4"}],"minor_comments":[{"comment":"The sentence 'we obtain the nonlinear fluid fluctuation profiles q(z) shown in eq. (5)' appears to refer to Fig. 5, not Eq. (5), which is the macroscopic matching condition. Please correct the cross-reference.","section":"§5.3"},{"comment":"The symbol R is defined in Eq. (20) as a z-dependent condition, but later R is used to mean max|R|, including in figure captions and the text. Please distinguish the local profile from its maximum and state explicitly that the plotted quantity is max_z |R(z)|.","section":"Eq. (20) and Fig. 1 captions"},{"comment":"In Figures 1–3, the right axis is labeled 'Relative err.' but the caption does not always state that the error is defined as |v_w^{fluid(κ=3)} − v_w^{WallGo}|/v_w^{WallGo}. Please define this in every caption or in a common note.","section":"§4.3"},{"comment":"The statement that the numerical results agree with ref. [23] and ref. [43] 'in the Γ_ij entries where there is an expected overlap' would be more useful if it specified the comparison accuracy (e.g., number of matching digits) and whether the same thermal masses and amplitude normalizations were used.","section":"Appendix A.3"}],"recommendation":"major_revision","confidential_remarks":"The paper addresses a timely and practically important comparison, and the linear benchmark part is well executed. The main scientific risk is the unsupported O(δ²) conclusion: the nonlinear analysis is performed at κ=1 while the linear benchmarks use κ=3, and no convergence study for the nonlinear system is provided. If the authors can extend the nonlinear calculation to higher truncation order or otherwise show that the v_w shift is not an artifact of the low-order truncation, the paper would be suitable for publication. The shared leading-log collision uncertainty should also be more prominently framed."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This paper is genuinely useful: the first direct head-to-head between the fluid Ansatz and WallGo on the same models with the same leading-log collision terms, including sanity checks (equilibrium-only limit, parameter-number convergence, explicit WKB caveats). The linear-level result is clean and reproducible in spirit: for weak transitions (α ≲ 0.01) the methods agree well, and for strong transitions they diverge by 40–60%. That divergence is real and worth knowing, and the paper is appropriately cautious about WKB validity in that regime.\n\nThe soft spot is the nonlinear O(δ²) section, which carries the paper's most striking claim. The whole analysis in Sec. 5 truncates the momentum expansion at κ=1 (3 moments), while the linear benchmarks that produced the claimed discrepancies use κ=3 (10 moments). Figure 4 shows the linear result changes noticeably with truncation order up to D=10. So the observed large velocity shift from O(δ²) terms could be an artifact of the poor low-order linear truncation rather than a robust property of the converged fluid Ansatz. The authors do not report a convergence study of the nonlinear system in κ. They also note that the direct second-order pressure contribution stays small (≈4×10⁻²), which makes the claimed backreaction entirely dependent on the coupled low-order system. This doesn't invalidate the linear benchmark, but the 'significant shifts as α→1' conclusion needs to be qualified as preliminary until checked at higher κ.\n\nThe other issues are minor and addressable: the custom fluid-Ansatz code isn't released, the main comparison figures lack error bands on the fluid side, and both methods share the same leading-log collision model with O(1) theoretical uncertainties the authors themselves flag (Sec. 3.1, Appendix A.3, Conclusions). Those concerns are real but external to the benchmark's internal consistency.\n\nThis paper is for anyone computing v_w for gravitational-wave predictions. The linear benchmark is a solid reference; the strong-regime discussion is a useful warning. Worth sending to peer review. A good referee should ask for a κ-convergence study of the nonlinear system and ideally a public release of the fluid-Ansatz code. As is, the linear part is solid, and the O(δ²) part needs strengthening before it can be cited as evidence about strong transitions.","headline":"Useful first benchmark of two wall-velocity methods, but the O(δ²) strong-transition claim rests on an unconverged 3-moment truncation and should be treated as preliminary.","tokens_in":25249,"tokens_out":2189,"would_cite":true,"duration_ms":20836,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper claims that the two dominant semi-classical methods for computing cosmological bubble-wall velocities agree only for mild transitions (α≲0.01) and split by roughly 40–60% for the strong transitions gravitational-wave observatorie","keywords":["bubble wall velocity","cosmological phase transitions","fluid Ansatz","Chebyshev spectral method","Boltzmann equation","WKB approximation","thermal friction","gravitational waves"],"falsifier":"Take one strong-transition benchmark (for example the dimension-six model with α≈0.1), solve the Boltzmann equation with a full phase-space method that imposes no shape Ansatz, and compare the resulting terminal velocity with the two methods' 40–60% spread; also compute the largest fluctuation amplitude in the spectral solution to see whether the divergence point coincides with max|R|≈1.","tokens_in":24380,"feed_emoji":"🌊","tokens_out":6599,"duration_ms":64739,"temperature":0.7,"pith_summary":"The paper compares two schemes for computing the terminal velocity of a cosmological bubble wall from the semi-classical Boltzmann equation: the fluid Ansatz, which represents departures from equilibrium as momentum-polynomial fluctuations tied to fluid variables, and the spectral Chebyshev method used in a public companion code. Its central claim is that the two agree within errors for weak transitions (α≲0.01), match well when only top-quark annihilation is included, but disagree by roughly 40–60% once scattering processes are added and the transition is strong (α≳0.1). The paper also claims that within the fluid Ansatz the standard linearization fails in the strong regime: quadratic O(δ²) terms shift the terminal velocity significantly even though their direct pressure contribution stays small (≤ about 4%). Why a reader should care: strong first-order phase transitions are the main targets for planned gravitational-wave observatories, so the velocity that seeds their predicted spectra is least certain exactly where the signal is most interesting.","feed_headline":"Bubble wall speeds diverge 40–60% when transitions turn strong","feed_subtitle":"Methods agree for weak transitions, diverge by 40–60% exactly where strong signals would appear.","key_machinery":"The central comparison objects are (1) the fluid Ansatz, which postulates that the distribution has equilibrium Bose/Fermi form with an argument shifted by a fluctuation δ expanded in powers of four-momentum; taking moments reduces the Boltzmann equation to a linear ordinary-differential system solved by a Green's function method, with a new O(δ²) boundary-value-problem extension; and (2) the spectral method, which expands δf in restricted Chebyshev polynomials on compactified momentum and spatial coordinates and solves the Boltzmann equation on a discrete Chebyshev grid, turning it into an algebraic matrix problem. Both use a leading-log collision operator with thermal-mass-regulated t/u-ch","core_discovery":"The paper's discovery is a quantitative map of where two popular Ansätze leave the safe zone. At α≲0.01, the fluid Ansatz at third order in a momentum expansion and the spectral Chebyshev method return the same terminal wall velocity, with close agreement when only top-quark annihilation is retained; including top-quark scatterings widens the gap. For α≳0.1 the two terminal velocities differ by O(40–60%), and the divergence starts exactly where a diagnostic parameter measuring fluctuation size approaches one. Extending the fluid Ansatz to second order in fluctuations shows that, for strong transitions, the quadratic terms increase friction and lower the terminal velocity substantially relati","pith_inferences":["If the order-one collision-term uncertainties are real, the two methods share them; the 40–60% strong-transition gap should be treated as a lower bound on the total theoretical error in v_w, since both could be offset from the true value.","The R≈1 correspondence suggests a testable conjecture: any shape Ansatz that keeps the fluctuation amplitude small will reproduce near-equilibrium velocities, while strong transitions require either a fully non-linear Boltzmann solution or a non-Boltzmann treatment.","The velocity shift from small quadratic pressures implies gravitational-wave spectra, which are steep functions of v_w, could change by more than their nominal error bars even when pressure-based convergence tests look fine.","One can extend the comparison by implementing the same O(δ²) terms in the spectral Chebyshev framework; if it shows a similar velocity shift, the mismatch is physics, not a peculiarity of the fluid Ansatz."],"forward_implications":["Weak phase transitions (α≲0.01) have cross-validated velocity predictions: either scheme can be used there with the disagreement inside the estimated errors.","Strong phase transitions (α≳0.1) are not converged across methods: a 40–60% spread in v_w means spectra predicted for future gravitational-wave observatories carry a significant unquantified velocity uncertainty on top of collision-operator uncertainties.","Within the fluid Ansatz, neglecting O(δ²) in the distribution function is not a safe shortcut for strong transitions: even though the quadratic pressure is small, its backreaction substantially lowers the terminal velocity, and the shift worsens agreement with the spectral approach.","The WKB/semi-classical Boltzmann setup is itself strained in the same region (LwT≲3), so the discrepancy should not be read as one Ansatz being correct; high-precision predictions likely require going beyond WKB.","Order-one uncertainties in the leading-log collision terms (acknowledged in the paper) mean the weak-regime agreement demonstrates internal consistency of the two implementations, not validated plasma physics."],"fun_headline_variants":["Strong transitions split wall speed methods by up to 60%","Wall-speed approaches diverge sharply for strong phase transitions","Weak transitions agree, strong ones diverge 40–60%","Wall velocity predictions split as transitions get stronger"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The entire comparison rests on the leading-log collision operator with thermally regulated exchange amplitudes, which the paper itself says can carry order-one theoretical uncertainties; if that collision model is wrong, both methods inherit the same bias and their agreement in the weak regime would not establish the real wall velocity.","fun_headline_variants_meta":{"raw":{"variants":["Strong transitions split wall speed methods by up to 60%","Wall-speed approaches diverge sharply for strong phase transitions","Weak transitions agree, strong ones diverge 40–60%","Wall velocity predictions split as transitions get stronger"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000207,"raw_usage":{"total_tokens":1295,"prompt_tokens":860,"completion_tokens":435,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":604,"completion_tokens_details":{"reasoning_tokens":369}},"tokens_in":604,"tokens_out":435,"duration_ms":5033,"temperature":1.0,"reasoning_tokens":369,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T15:40:16.030263+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take one strong-transition benchmark (for example the dimension-six model with α≈0.1), solve the Boltzmann equation with a full phase-space method that imposes no shape Ansatz, and compare the resulting terminal velocity with the two methods' 40–60% spread; also compute the largest fluctuation amplitude in the spectral solution to see whether the divergence point coincides with max|R|≈1.","supporting_citations":[],"review_version":1}