{"id":"a674d8a2-6027-4a88-acf5-3128c007ec53","arxiv_id":"1908.10990","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":7,"one_line_summary":"Finite-size scaling of wrapping probabilities gives Tc(XY)=2.2018441(5), Tc(Villain)=0.33306704(7), (t/U)c=0.0597291(8), nu=0.67183(18), and eta=0.03853(48) for the 3D U(1) universality class.","lead":"Using worm-type Monte Carlo simulations, this paper extracts the critical points of three models in the 3D U(1) universality class with record precision. It also shows that a wrapping-probability derivative has tiny finite-size corrections, yielding a sharp correlation-length exponent.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Quoted systematic errors rely on assumed correction exponents (omega2=1.77) and a neglected analytic background in Tw; a robustness check against free exponents is needed.","rationale":"The reader's weakest-assumption analysis correctly identifies the fixed correction exponents and the neglected analytic background in Tw as the most fragile point of the error-bar construction. My independent reading of the paper reaches the same conclusion: the fits are internally consistent and the central values agree with prior literature, but the quoted uncertainties are conditional on an assumed correction spectrum that is only partially verified. This is a genuine concern about the reliability of the error bars, but it does not by itself overturn the paper's central claims. The authors provide multiple cross-checks (Lmin variation, with/without b2 terms, fixed-nu consistency checks) that suggest the corrections are small, and the omega1 estimates from their own data are consistent with the assumed value. The concrete test I propose would settle whether the unverified omega2 and analytic background actually shift the results beyond the quoted errors; until such a test is performed, the concern remains a caveat rather than a demonstrated flaw. Therefore the appropriate verdict is unchanged: accept the paper as a high-precision study, while noting that the error bars should be interpreted with this caveat in mind.","tokens_in":27974,"tokens_out":8888,"duration_ms":97711,"concrete_test":"Have the authors re-analyze their raw per-L data at Tc for Tw and GRxE with (i) omega2 left free, (ii) an explicit analytic-background term (c L^{-(2-eta)} for Tw; c L^{-1/nu} for GRxE), and (iii) both. Additionally, re-fit the critical-point data with omega1 and omega2 both free. If the central values of Tc, nu, and eta shift by more than the quoted 1-sigma errors (5e-7, 0.00018, 0.00048, respectively), the systematic errors are underestimated; if the shifts are smaller, the current error bars are validated.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's headline precision (Tc(XY)=2.2018441(5), Tc(Villain)=0.33306704(7), nu=0.67183(18), eta=0.03853(48)) is obtained from finite-size scaling fits that fix omega1=0.789 and omega2=1.77, taken from RG/literature, for all observables including the new wrapping-probability derivatives and the Bose-Hubbard world-line data. The authors verify omega1 only partially (Tables VII and VIII; estimates 0.77(13) and ~0.7), but omega2=1.77 is never independently checked for these observables. Moreover, Eq. (32) for Tw neglects an analytic background, with the statement that it is 'effectively higher-order' asserted but not demonstrated. For Tw, the analytic-background exponent relative to the leading term is 2-eta ~ 1.961, only slightly larger than omega2=1.77, so if its amplitude is not tiny it can bias eta by more than the quoted 0.00048. For GRxE, the conclusion that leading corrections vanish is reached within a model that assumes corrections only at omega1=0.789; an analytic background (exponent ~1.488 when factored as L^{-1/nu}) or a subleading term at 1.77 could shift nu beyond the stated 0.00018 if amplitudes are moderate. Because all critical-point fits share the same fixed omega values, a systematic error in these exponents would shift Tc estimates coherently across observables, and the 'conservative' error obtained from the spread of fits would not cover the shift. The central numerical results are likely correct, but the quoted error bars are only as secure as the assumed correction spectrum.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents a worm-type Monte Carlo study of three models in the three-dimensional U(1) universality class: the classical XY model, the Villain model, and the two-dimensional Bose-Hubbard model with unitary filling. From finite-size scaling analyses of wrapping probabilities and the superfluid stiffness, the authors determine the critical points Tc(XY)=2.2018441(5), Tc(Villain)=0.33306704(7), and (t/U)c=0.0597291(8). They further measure the correlation-length exponent nu=0.67183(18) from the temperature derivative of a wrapping probability, which is found to have negligible leading corrections, and the exponent eta=0.03853(48) from a susceptibility-like quantity Tw. The paper introduces wrapping-probability derivatives as a high-precision observable for U(1) criticality and reports universal critical wrapping probabilities.","tokens_in":28288,"tokens_out":13953,"duration_ms":134154,"significance":"If the quoted precision is substantiated, these results provide the most accurate numerical benchmarks for the 3D U(1) universality class to date, improving on the best existing Monte Carlo estimates. The method of using GRxE, the temperature derivative of a wrapping probability, for determining nu is a useful new technique that avoids reliance on correction-to-scaling amplitudes. The paper is exemplary in its transparency: the finite-size scaling fits are documented in full detail, with chi2/dof, Lmin stability checks, and error bars derived from the spread of many fit choices. The consistency of the universal amplitudes across the XY and Villain models (Table V) is a persuasive check. The main weakness is that the quoted systematic errors depend on assumptions about correction-to-scaling exponents and the absence of analytic backgrounds that are only partially tested; this is the focus of the major comments.","major_comments":[{"comment":"The assertion that the analytic background in Tw is 'effectively higher-order' is not demonstrated, and the same issue applies to GRxE. For Tw, the analytic-background exponent relative to the leading behavior is 2-eta about 1.961, which is close to the assumed omega2=1.77; for GRxE, an analytic background would contribute with relative exponent 1/nu about 1.488. If the corresponding amplitude is not tiny, it can bias eta or nu by more than the quoted uncertainties (0.00048 and 0.00018, respectively). I request a quantitative check, for example by adding a term proportional to L^{-(2-eta)} or L^{-1/nu} in Eqs. (32) and (33) and reporting the fitted amplitude, or an equivalent argument showing that the background is negligible.","section":"Secs. III C and IV B (Eq. (32) and Table VI)"},{"comment":"All finite-size scaling analyses fix the correction exponents omega1=0.789 and omega2=1.77. While omega1 is partially tested (Tables VII and VIII yield omega1 about 0.77(13) and about 0.7), omega2 is never independently verified for the wrapping-probability observables or for the Bose-Hubbard world-line data. Since all fits share the same omega2, a systematic error in this exponent would shift the quoted Tc, nu, and eta coherently, and the spread-based error estimates would not cover the shift. I ask for fits with omega2 treated as free, or with a different subleading correction structure, for at least the Villain-model observables that determine the headline results, to demonstrate that the quoted precision is robust.","section":"Secs. III C and IV A (Tables III, IV, VII, IX, XI)"}],"minor_comments":[{"comment":"It would be helpful to state explicitly that the correction amplitudes b_m are assumed independent of temperature, so that the derivative with respect to T in Eq. (26) does not act on them; this assumption underlies the claimed negligibility of corrections for GRxE and should be stated in the text.","section":"Eq. (24) and Eq. (26)"},{"comment":"The y-axis label 'Rx-b1L^{-omega1}-b2L^{-omega2}' is somewhat ambiguous; please define in the caption that the plotted quantity is the wrapping probability with the fitted correction terms subtracted.","section":"Fig. 4"},{"comment":"The header of Table VII lists columns b1, omega1, and b2, but the text says 'omega2=1.77 is adopted'; a note clarifying that omega2 is fixed while omega1 is sometimes free (and sometimes fixed at 0.789) would improve readability.","section":"Table VII"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a strong numerical contribution and the data presentation is of high quality. My main reservation is that the headline precision claims rest on fixed correction exponents (omega2=1.77) and on an asserted but unquantified analytic background for Tw. I would support acceptance after the authors demonstrate robustness to these assumptions, for example by adding fits with free omega2 or explicit background terms. This is a standard request for a paper whose central contribution is precision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Punchline: this is a paper to take seriously. It delivers the most precise Monte Carlo critical points to date for the 3D XY, Villain, and 2D Bose-Hubbard models in the 3D U(1) class, and it introduces wrapping probabilities to directed-flow and world-line representations, which is a genuine methodological step. The claim that the T-derivative of a wrapping probability has negligible leading corrections holds up as far as the fits show; the b1 amplitudes in Table VI are consistent with zero and the nu estimates are stable across Lmin.\n\nWhat's good: the worm algorithm is described carefully, the FSS fits are transparent, and the errors are conservatively taken as the spread over fits with different Lmin and correction combinations. The critical points—Tc(Villain)=0.33306704(7), Tc(XY)=2.2018441(5), (t/U)c=0.0597291(8)—are consistent with earlier estimates and improve precision dramatically. The universal wrapping probabilities are new and provide a cross-check of universality between XY and Villain. The nu=0.67183(18) is comparable to the best MC results and rules out the space-shuttle experimental value, which is worth saying.\n\nSoft spots: the analysis fixes omega1=0.789 and omega2=1.77 for all observables, including the new wrapping-probability derivatives. The paper partially tests omega1 (Tables VII/VIII give ~0.77(13) and ~0.7), but omega2 is never independently checked. The stress-test note about the analytic background in Tw is fair: Eq. (32) drops it with the statement that it is effectively higher-order, and since the analytic-background exponent 2-eta ~ 1.96 is close to omega2=1.77, a moderate amplitude could bias eta by more than the quoted 0.00048. That said, the fits with and without b2 terms are stable, and the quoted errors already come from the spread across fits, so I would not call this a fatal flaw. The coherent shift of Tc across observables that the stress-test worries about would require a correction spectrum substantially different from what the data support; the paper's internal consistency between multiple observables and models is evidence against it. Still, a robustness check with omega2 free would be a reasonable request from a referee.\n\nWho this is for: Monte Carlo practitioners working on O(2)/U(1) critical phenomena and anyone who needs benchmark critical points. I would cite it. The lack of raw data/code is a minor inconvenience, not a blocker.\n\nVerdict: accept—send it to a serious referee. The method is new, the numbers are benchmark-quality, and the soft spots are manageable.","headline":"High-precision benchmark critical points for the 3D U(1) class, with a clever new use of wrapping probabilities; the error bars rest partly on assumed correction exponents, but the central results look solid.","tokens_in":28907,"tokens_out":2217,"would_cite":true,"duration_ms":20250,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["82B20","82B27","82B80"],"pacs":["05.10.Ln","64.60.Fr"],"model":"deepseek-v4-flash","headline":"Using worm-type Monte Carlo on a torus, this paper reports that wrapping probabilities of directed flows and particle world-lines locate the critical points of the 3D XY, Villain, and Bose-Hubbard models at…","keywords":["3D XY model","Villain model","Bose-Hubbard model","wrapping probability","worm algorithm","finite-size scaling","critical exponents","U(1) universality class"],"falsifier":"At the Villain critical point, refit the $G_{R_x E}$ data to $L^{1/\\nu}(Q_0 + b_1 L^{-\\omega_1})$ with $b_1$ and $\\omega_1$ left free and see whether the resulting $\\nu$ stays within $0.671\\,83(18)$ when the smallest included lattice size $L_{\\min}$ is raised; a stable shift beyond one quoted error would falsify the claim of negligible leading corrections. A second, model-level check is to repeat the Bose-Hubbard critical-point analysis with a different inverse-temperature contour, say $\\beta = 4L$ instead of $2L$; if $(t/U)_c$ moves outside $0.059\\,729\\,1(8)$, the quantum-to-classical scaling assumption would need revision.","tokens_in":27730,"feed_emoji":"🌀","tokens_out":12090,"duration_ms":108115,"temperature":0.7,"pith_summary":"Using worm-type Monte Carlo on a torus, this paper argues that wrapping probabilities—whether the directed flow of the classical XY or Villain models, or the particle world-lines of the 2D Bose-Hubbard model, winds around the lattice—are the most reliable high-precision observables for the three-dimensional U(1) universality class. From finite-size scaling of these dimensionless quantities, it determines the critical points $T_c(\\mathrm{XY}) = 2.201\\,844\\,1(5)$, $T_c(\\mathrm{Villain}) = 0.333\\,067\\,04(7)$, and $(t/U)_c(\\mathrm{BH}) = 0.059\\,729\\,1(8)$, improving on earlier estimates significantly. At the Villain critical point, the derivative of the one-direction wrapping probability with respect to temperature shows negligible leading finite-size corrections, which yields the correlation-length exponent $\\nu = 0.671\\,83(18)$; a susceptibility-like worm-return-time quantity gives $\\eta = 0.038\\,53(48)$. A sympathetic reader would care because these numbers are benchmarks that can be tested against conformal bootstrap calculations, other Monte Carlo methods, and experiments such as the superfluid transition of helium.","feed_headline":"Wrapping counts pin down three critical points in the 3D U(1) class","feed_subtitle":"Monte Carlo winding statistics give Tc benchmarks and nu=0.67183(18) for direct comparison with bootstrap and helium experiments.","key_machinery":"The central object is the dimensionless wrapping probability $R_\\kappa$, equal to one when a directed flow (XY and Villain models) or a particle world-line (Bose-Hubbard model) winds around the torus in direction $\\kappa$. Its temperature derivative $G_{R_\\kappa E} = \\mathrm{d}R_\\kappa/\\mathrm{d}T$ is estimated from the covariance of $R_\\kappa$ with the energy and obeys the finite-size scaling form $G_{R_\\kappa E} = L^{1/\\nu}(Q_0 + \\sum_m b_m L^{-\\omega_m})$. The simulations use worm-type updates, which sample the directed-flow or world-line configurations efficiently and let the wrapping numbers be read from the movement of the two defect points; the Bose-Hubbard model is simulated in the imaginary-time world-line representation with $\\beta = 2L$. The load-bearing feature is that the leading correction amplitude $b_1$ for $R_x$ is extremely small, so $\\nu$ can be determined without fitting that correction amplitude.","core_discovery":"The central discovery is that the topology of Monte Carlo configurations—measured by which directions the directed flows or world-lines wrap around the periodic boundaries—carries the critical information of the U(1) transition more cleanly than conventional observables. In the Villain model at criticality, $G_{R_x E} = dR_x/dT$, computed as a covariance of the wrapping indicator $R_x$ with the energy, scales as $L^{1/\\nu}$ with a leading correction amplitude consistent with zero, while $G_{R_a E}$, $G_{R_2 E}$, and the stiffness derivative carry visible corrections. This lets $\\nu = 0.671\\,83(18)$ be extracted from fits that omit the leading correction term, avoiding the corresponding parameter uncertainty. The same wrapping observables, together with the superfluid stiffness measured through winding-number fluctuations, locate the three critical points with uncertainty at or below $10^{-7}$, and the critical wrapping probabilities $R_x^c = 0.3787(2)$, $R_a^c = 0.6889(4)$, and $R_2^c = 0.2640(3)$ are found to be universal between the XY and Villain models. In addition, the mean worm return time $T_w$, scaling as $L^{2-\\eta}$, gives $\\eta = 0.038\\,53(48)$.","pith_inferences":["A natural test is whether $G_{R_x E}$ keeps its near-zero $b_1$ in other realizations of the (2+1)-dimensional U(1) class, such as the quantum rotor model or compact lattice gauge theory; if it does, the route to $\\nu$ becomes model-independent and could exceed the current precision.","The small corrections may be a feature of the directed-flow representation rather than of the universality class; simulating the XY model in its spin representation with the same wrapping criterion would separate representation effects from universal ones.","The reported universal values $R_x^c$, $R_a^c$, and $R_2^c$ could be sharpened by adapting high-precision cluster algorithms to the directed-flow picture, providing independent cross-checks.","The same covariance technique could be applied to other geometric observables, such as the probability of double winding or the distribution of winding numbers, to see whether similarly small corrections persist and to yield independent estimates of $\\omega_1$."],"forward_implications":["If correct, $T_c(\\mathrm{XY})=2.201\\,844\\,1(5)$, $T_c(\\mathrm{Villain})=0.333\\,067\\,04(7)$, and $(t/U)_c=0.059\\,729\\,1(8)$ supersede previous critical-point estimates for all three models.","$\\nu = 0.671\\,83(18)$ is consistent with the earlier Monte Carlo values $0.6717(1)$ and $0.6717(3)$, and the paper's numbers make the space-shuttle helium value $\\nu = 0.6709(1)$ unlikely.","The critical wrapping probabilities $R_x^c$, $R_a^c$, and $R_2^c$ reported for the XY and Villain models provide new universal dimensionless numbers for this universality class.","For the Bose-Hubbard model, the quantum critical point estimate improves on prior results by more than a factor of 40 in precision.","The negligible leading corrections in $G_{R_x E}$ provide a practical route to $\\nu$ with fewer fitting parameters, useful for other models in the same universality class."],"supporting_citations":[{"why":"Introduces the worm algorithm for classical statistical models; the paper's XY and Villain simulations build on it.","marker":"[22]"},{"why":"Supplies the previous most precise exponents $\\nu=0.6717(1)$, $\\eta=0.0381(2)$ and the verification of $\\omega_1=0.785(20)$, the benchmark this work improves and checks.","marker":"[11]"},{"why":"Gives the renormalization-group estimate $\\omega_1=0.789(11)$ and the subleading value $\\omega_2=1.77$ used to fix the correction exponents in the fits.","marker":"[56]"},{"why":"Provides an earlier high-precision Monte Carlo $\\nu=0.6717(3)$ and the critical scaled stiffness $0.5160(6)$ that the paper's $\\rho_s^c L$ result is compared against.","marker":"[12]"},{"why":"Gives the conformal bootstrap estimates $\\nu=0.6719(11)$ and $\\eta=0.03852(64)$ that the new exponents are tested against.","marker":"[17]"},{"why":"Defines the superfluid stiffness through winding-number fluctuations, used to sample $\\rho_s$ and its critical value.","marker":"[53]"},{"why":"Gives the strong-coupling-expansion benchmark $(t/U)_c=0.05974(4)$ that the Bose-Hubbard quantum critical point determination improves upon.","marker":"[28]"},{"why":"Gives the quantum Monte Carlo benchmark $(t/U)_c=0.05974(3)$ for the Bose-Hubbard quantum critical point.","marker":"[29]"},{"why":"Presents an explicit worm algorithm for the Villain/quantum-rotor model that the paper adapts for its Villain simulations.","marker":"[34]"}],"fun_headline_variants":["Wrapping topology nails 3D U(1) critical points","Ultra-precise Tc and nu from Monte Carlo wraps","Universal wrap probabilities set U(1) benchmarks","Clean nu from wrapping: 0.67183(18) in U(1)","Three models, one class: MC wraps define U(1)"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The fits fix the two correction exponents $\\omega_1 = 0.789$ and $\\omega_2 = 1.77$ from earlier renormalization-group and Monte Carlo work and assume they also apply to the new wrapping observables and to the Bose-Hubbard world-line data; the second of these is not independently verified here, and the analysis also assumes that no smooth, non-scaling background term enters at the fitted orders.","fun_headline_variants_meta":{"raw":{"variants":["Wrapping topology nails 3D U(1) critical points","Ultra-precise Tc and nu from Monte Carlo wraps","Universal wrap probabilities set U(1) benchmarks","Clean nu from wrapping: 0.67183(18) in U(1)","Three models, one class: MC wraps define U(1)"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000366,"raw_usage":{"total_tokens":2086,"prompt_tokens":1183,"completion_tokens":903,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":799,"completion_tokens_details":{"reasoning_tokens":814}},"tokens_in":799,"tokens_out":903,"duration_ms":9446,"temperature":1.0,"reasoning_tokens":814,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:27:39.012639+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"At the Villain critical point, refit the $G_{R_x E}$ data to $L^{1/\\nu}(Q_0 + b_1 L^{-\\omega_1})$ with $b_1$ and $\\omega_1$ left free and see whether the resulting $\\nu$ stays within $0.671\\,83(18)$ when the smallest included lattice size $L_{\\min}$ is raised; a stable shift beyond one quoted error would falsify the claim of negligible leading corrections. A second, model-level check is to repeat the Bose-Hubbard critical-point analysis with a different inverse-temperature contour, say $\\beta = 4L$ instead of $2L$; if $(t/U)_c$ moves outside $0.059\\,729\\,1(8)$, the quantum-to-classical scaling assumption would need revision.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the worm algorithm for classical statistical models; the paper's XY and Villain simulations build on it."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the previous most precise exponents $\\nu=0.6717(1)$, $\\eta=0.0381(2)$ and the verification of $\\omega_1=0.785(20)$, the benchmark this work improves and checks."},{"cited_title":"Observation of correlated particle-hole pairs and string order in low-dimensional mott insulators,","cited_arxiv_id":null,"evidence_quote":"Gives the renormalization-group estimate $\\omega_1=0.789(11)$ and the subleading value $\\omega_2=1.77$ used to fix the correction exponents in the fits."},{"cited_title":"High-precision measurement of the thermal ex- ponent for the three-dimensional xy universality class,","cited_arxiv_id":null,"evidence_quote":"Provides an earlier high-precision Monte Carlo $\\nu=0.6717(3)$ and the critical scaled stiffness $0.5160(6)$ that the paper's $\\rho_s^c L$ result is compared against."},{"cited_title":"Estimates of the critical temperatures for the 3D X Y and Villain models and the critical hopping amplitude for th e two- dimensional (2D) unitary-filling BH model","cited_arxiv_id":null,"evidence_quote":"Gives the conformal bootstrap estimates $\\nu=0.6719(11)$ and $\\eta=0.03852(64)$ that the new exponents are tested against."},{"cited_title":"The dy- namics of quantum criticality revealed by quantum monte car lo and holography,","cited_arxiv_id":null,"evidence_quote":"Defines the superfluid stiffness through winding-number fluctuations, used to sample $\\rho_s$ and its critical value."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the strong-coupling-expansion benchmark $(t/U)_c=0.05974(4)$ that the Bose-Hubbard quantum critical point determination improves upon."},{"cited_title":"Albeit these estimates are all based on Monte Carlo simula- tions, they are not completely consistent with each other","cited_arxiv_id":null,"evidence_quote":"Gives the quantum Monte Carlo benchmark $(t/U)_c=0.05974(3)$ for the Bose-Hubbard quantum critical point."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Presents an explicit worm algorithm for the Villain/quantum-rotor model that the paper adapts for its Villain simulations."}],"review_version":1}