{"id":"8c53385d-91c0-4a90-89cc-aa37d6ce9fee","arxiv_id":"1908.00586","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For undamped motion in 1D and 2D single-well power-law potentials, random hard velocity reversals do not change the stationary position and velocity distributions, which equal the deterministic time averages.","lead":"This paper shows that randomly reversing a particle's velocity at random times in a single-well potential leaves the long-time probability distributions of position and velocity unchanged, matching the deterministic motion. It derives explicit density formulas for 1D and 2D power-law wells and verifies them with simulations.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified: the 2D equivalence follows from radial reduction plus phase mixing, so the dense-orbit/heavy-tailed concern is not load-bearing for the central claim.","rationale":"The reader's conditional verdict is reasonable as a request for additional rigor and typo corrections, but the specific weakest assumption identified—dense orbits in 2D—does not carry the weight assigned to it. The p(r) result is a 1D radial-reduction statement and does not require denseness. The equality between deterministic and randomized distributions holds on the closure of any orbit, dense or closed, because velocity reversals preserve the invariant set and the stroboscopic chain mixes on it for any waiting-time density. The α<1 discrepancy in Figs. 6-7 is expected at t=1000: with α=0.1 the expected number of reversals by t=1000 is only about two, so the no-reversal deterministic peak still dominates. The proposed long-time test would settle whether the peak decays as P(τ_1>t), which is the correct asymptotic signature. I therefore find no load-bearing flaw in the central claim, while agreeing that the presentation would be improved by a precise ergodicity condition on p(τ) and by fixing the noted typographical errors.","tokens_in":13928,"tokens_out":22888,"duration_ms":269226,"concrete_test":"Run the 2D n=4 randomized simulation with α=0.1 waiting times to t = 10^6 and t = 10^8, instead of t = 1000, and compare the empirical p(r) with Eq. (D.12). Measure both the total variation distance to the predicted density and the weight of the residual deterministic peak. If the distance decays toward zero and the peak weight tracks the no-reversal probability P(τ_1>t) ~ t^{-α}, then the α<1 discrepancy is purely finite-time and the asymptotic insensitivity claim is confirmed; if the distance saturates above zero, the infinite-mean case would be a genuine counterexample to the universal claim.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that hard velocity reversals at random times asymptotically reproduce the deterministic occupation measure, for essentially any waiting-time density. The reader's flagged assumption—that the 2D deterministic orbit must be dense on its energy-angular-momentum surface—is not actually needed for the main equality. In a central potential, the radial motion with fixed E and L is a 1D oscillator in the effective potential U_eff(r) = L^2/(2mr^2) + κr^n. A velocity reversal sends L to -L, but U_eff depends only on L^2, so the radial process is exactly the 1D randomized system studied in Sec. 3.1. Hence the radial density p(r) ∝ 1/sqrt(E - U_eff(r)) follows from the same 1D argument, with or without orbit density. For the marginal densities p(x), p(y), p(phi), the deterministic and randomized processes converge to the same invariant measure on the closure of the orbit: for a generic non-closed orbit this closure is the full torus (dense), while for a closed rational orbit both processes remain on the same closed curve. In either case they agree asymptotically, so the insensitivity claim survives. The apparent failure for α<1 at t=1000 is a finite-time effect: before the first reversal the state is exactly deterministic, and P(τ_1>t) ~ t^{-α} decays to zero, while after any positive number of reversals the stroboscopic chain on the compact energy surface mixes if p(τ) has a density. Thus the heavy-tailed simulations are consistent with slow, not absent, convergence. The paper would benefit from stating the mild ergodicity condition on p(τ) and from correcting the typos in Appendix B and the rmin/rmax swap, but these do not threaten the central result.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper studies deterministic Newtonian motion in single-well potentials of the form V(x)=κ|x|n in one and two dimensions and compares the long-time statistical properties of the deterministic orbit with those of a randomized process in which the velocity is reversed at random time instants (hard velocity reversals). The authors derive exact expressions for the stationary position and velocity densities in 1D for arbitrary n (Eqs. (5)-(6) and Appendix A), and an exact formula for the radial density in 2D via the effective potential (Eq. (D.12)). They present Monte Carlo simulations showing that the asymptotic densities are insensitive to the distribution of waiting times between reversals, both in 1D and for the marginal densities in 2D, and that randomization of initial conditions instead leads to modified densities. The paper also discusses transient effects for heavy-tailed waiting-time distributions.","tokens_in":14228,"tokens_out":11027,"duration_ms":109580,"significance":"The main contribution is the explicit confirmation that energy-conserving velocity reversals at random times do not alter the long-time distribution of a conservative system in a single-well potential, generalizing earlier results of Dybiec et al. [5] to arbitrary exponent n and to the 2D radial density. The derivations are elementary and transparent, and the numerical simulations agree with the analytical predictions in the parameter regimes studied. The paper provides a clean example of the equivalence of time and ensemble averaging under a specific randomization, and it offers useful formulas for researchers working on Lévy-walk-type models in external potentials. The novelty is moderate, but the results are solid and clearly presented.","major_comments":[],"minor_comments":[{"comment":"The formula for \\tilde{p}(E,v) contains x^2 in several places where v^2 must appear, e.g., \"1 - m x^2/(2E)\" and \"2E/(m x^2)\", while the stated condition of validity is E - m v^2/2 > 0. This inconsistency makes the printed expression unusable; please correct the variable name throughout the equation.","section":"Appendix B, Eq. (B.1)"},{"comment":"The text reads \"Distances from the origin to apocenters (rmin) and pericenters (rmax)\" and later \"the apocenter (rmin) and the pericenter (rmin)\"; the labels are switched. The apocenter is the largest radius (rmax) and the pericenter is the smallest radius (rmin).","section":"Section 3.2.1, paragraph on Fig. 5"},{"comment":"The extrapolation to t→∞ for heavy-tailed waiting times with α<1 is based on simulations at t=1000 that still show a residual deterministic peak. A brief explanation of why the peak is expected to vanish (e.g., the probability of no reversal decays as t^{-α}) would make the asymptotic claim more persuasive.","section":"Section 3.2.2, Figs. 6-7"},{"comment":"The abstract states a general insensitivity of probability distributions to velocity reversals in 2D. While the radial density p(r) follows exactly from the 1D reduction in Appendix D, the corresponding statement for px(x), py(y), and p(ϕ) is supported in the manuscript only by finite-time numerics for n≠2. A sentence clarifying which 2D results are proven and which are numerical would bring the wording in line with the evidence presented.","section":"Abstract and Section 3.2.2"}],"recommendation":"minor_revision","confidential_remarks":"The manuscript is a modest extension of the authors' previous work [5], but the technical content is sound and the presentation is clear. The issues identified are local typos and a desire for a more explicit statement of the 2D asymptotic claim's basis. If the authors fix the typographical errors and add a brief justification or caveat regarding the 2D marginals, the paper would be suitable for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nA quick read on 1908.00586. The paper does what it says: it writes down exact stationary densities for a particle in a monomial single-well potential, first for 1D time averages, then shows that random hard velocity reversals asymptotically reproduce those same densities. The new items are the general-n formulas (5)-(6) for p(x) and p(v), and the 2D radial density (D.12) obtained from the effective potential. Both follow from textbook mechanics, but they are cleanly derived and the simulations match. The comparison of the three protocols (long deterministic trajectory, ensemble over the energy surface, velocity-reversal ensemble) is pedagogically nice.\n\nThe soft spots are minor. Appendix B's Eq. (B.1) for the velocity density contains x^2 where v^2 belongs; the text swaps rmin and rmax when describing apocenter/pericenter; and the abstract's blanket statement about insensitivity to p(τ) is stronger than the finite-time numerics show for heavy-tailed waiting times with α<1, where the deterministic peak persists past t=1000. That last point is only a finite-time effect; the stress-test is right to say the 2D radial reduction makes the dense-orbit assumption unnecessary for the main p(r) result, and the heavy-tailed decay is slow but not absent. Still, the paper would be improved by a sentence stating the ergodicity/mixing condition on p(τ) and by a careful pass over the typos.\n\nNo code or data are included, but the derivations are reproducible by hand, and the equations are internally consistent. The citation pattern is straightforward, mostly building on the authors' own earlier paper [5], with no sign of inflated claims beyond the abstract wording.\n\nWho is this for? Anyone working on Lévy walks, random velocity flips, or occupation measures in conservative systems. It is a modest advance, not a breakthrough, but it is a solid one. I would send it to peer review: a serious referee can fix the typos and tighten the asymptotics statement in about an hour, and the exact formulas are worth having on record.\n\nVerdict: deserves a referee.","headline":"Correct, modest extension of the authors' earlier work; the main claim survives the flagged 2D-orbit objection, but the abstract overstates heavy-tailed convergence.","tokens_in":14830,"tokens_out":2905,"would_cite":false,"duration_ms":28776,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["05.40.Fb","05.10.Gg","02.50.-r","02.50.Ey"],"model":"deepseek-v4-flash","headline":"This paper establishes that random hard velocity reversals leave the asymptotic position and velocity statistics of a particle in a single-well potential exactly equal to the deterministic orbit's statistics, regardless of the…","keywords":["single-well potential","velocity reversal","energy-conserving randomization","stationary probability density","waiting-time distribution","central potential","quasi-periodic motion","time-average versus ensemble-average"],"falsifier":"Run the randomized one-dimensional dynamics to a very large observation time with a fixed initial condition and a waiting-time distribution with finite mean, such as an exponential distribution, and check whether the measured p(x) and p(v) converge to Eqs. (5) and (6) with the prescribed normalization; a persistent mismatch would falsify the central claim. For heavy-tailed waiting times with α<1, simulate to times far beyond t=1000 and test whether the remnant deterministic peak decays to the smooth density, and in 2D with n=4 compare the long-run normalized p(r) histogram against Eq. (D.12) for several initial conditions.","tokens_in":13705,"feed_emoji":"🔁","tokens_out":7045,"duration_ms":73217,"temperature":0.7,"pith_summary":"The paper studies an undamped particle in a one-dimensional single-well potential V(x)=κ|x|^n whose deterministic motion is periodic. It shows that if the motion is interrupted at random instants by a hard reversal of velocity, an operation that conserves total energy, then in the long-time limit the probability densities of position and velocity are exactly the same as those recorded along one long deterministic trajectory, and are insensitive to how the reversal times are chosen. The explicit densities are u-shaped and blow up at the turning points where the velocity vanishes. The same insensitivity is demonstrated for two-dimensional central-potential wells, where the key radial density is p(r) ∝ 1/√(E - L²/(2mr²) - κr^n). Randomizing the initial conditions instead produces genuinely different densities, obtained as mixtures of the fixed-energy densities, so the insensitivity is specific to the velocity-reversal randomization.","feed_headline":"Random velocity flips don't change long-run statistics","feed_subtitle":"In single-well potentials, asymptotic position and velocity densities match the deterministic orbit for any reversal-time distribution.","key_machinery":"The load-bearing object is the constant-energy orbit and the period T of the deterministic motion. The density construction starts from the time-averaging relation p(x)dx = (2/T)dx/v, which says that the probability of finding the particle near x is proportional to the time it spends there, and the energy relation E = mv²/2 + κ|x|^n supplies v as a function of x. A hard velocity reversal maps the state (x,v) to (x,-v), preserving both position and total energy, so the randomized process never leaves the same energy surface and its stationary measure must be the same invariant measure as the deterministic flow; combining the period with the energy relation gives the explicit formulas. In two dimensions, the same logic is carried by the effective radial potential U_eff(r) = L²/(2mr²) + κr^n and the radial energy relation (m/2)ṙ² = E - U_eff(r), which yields p(r) ∝ 1/√(E - U_eff(r)).","core_discovery":"For a particle of mass m in a one-dimensional potential V(x)=κ|x|^n with a fixed initial condition and total energy E, the long-observation probability density of position is p(x) = (2/T)[(2/m)(E - κ|x|^n)]^{-1/2} and the density of velocity is p(v) = (2m/(Tκn))[(1/κ)(E - mv²/2)]^{1/n-1}, where T is the deterministic period. Interrupting the deterministic motion by hard velocity reversals, v(t_i)→-v(t_i), at random times drawn from any one-sided waiting-time distribution p(τ) produces, asymptotically, exactly these same densities, so the asymptotic statistics of the randomized process coincide with those of the deterministic trajectory and are independent of p(τ). In two dimensions with a central potential V(r)=κr^n, the same asymptotic insensitivity holds for marginal densities; the radial density is p(r) ∝ 1/√(E - L²/(2mr²) - κr^n), where L is the conserved angular momentum and the effective potential U_eff(r)=L²/(2mr²)+κr^n bounds the radial motion. The randomization thus makes ensemble averages interchangeable with time averages along a single constant-energy trajectory, while randomization of initial conditions breaks this equivalence and yields a mixture of fixed-energy densities.","pith_inferences":["The same asymptotic equivalence may hold for any bounded conservative one-dimensional system whose deterministic orbit is a single closed loop, because the construction uses only the period and the constant-energy curve, not the specific monomial form of the potential.","The paper's two-dimensional result for non-closed orbits implicitly assumes that the quasi-periodic trajectory samples the accessible annulus densely enough for time averages to converge to the steady densities; a rigorous ergodic statement, or a counterexample with a resonant orbit, would sharpen the claim.","A natural extension is to replace hard reversals by other energy-preserving random kicks, such as rotating the velocity vector in 2D; if the invariant measure on the energy surface is unique, the same asymptotic densities should appear.","The slow convergence observed for reversal-time distributions with diverging mean suggests that in practical finite-time experiments the reversal-time distribution still matters; a quantitative estimate of the decay rate of the initial-condition peak would be useful for applications to Lévy-walk-type models."],"forward_implications":["For a fixed initial condition, long time averages and ensemble averages over the randomized process give identical asymptotic densities, so simulations with velocity reversals can be used interchangeably with long deterministic trajectories to estimate p(x) and p(v).","The exact waiting-time distribution p(τ) only shapes transients; all choices with finite mean lead to the same stationary densities, so measured asymptotic statistics cannot be used to infer the reversal-time statistics.","Randomizing initial conditions changes the story: the resulting densities are mixtures of the fixed-energy densities and can be far from the arcsine-like shapes, with closed-form hypergeometric expressions for uniform energy distributions.","In two dimensions, exact marginal densities are available for the harmonic well (n=2), and the radial density p(r) ∝ [E - L²/(2mr²) - κr^n]^{-1/2} holds generally, with hard velocity reversals not altering these asymptotics.","For heavy-tailed waiting times with diverging mean (α<1), the stationary limit is approached only slowly; at finite observation time a peak corresponding to the initial condition remains, so the insensitivity is an asymptotic rather than a finite-time statement."],"supporting_citations":[{"why":"Introduces the hard-velocity-reversal randomized model and provides semi-analytical and numerical densities for selected exponents n, which the present paper's analytical derivation extends.","marker":"[5]"},{"why":"Supplies the period formula for motion in V(x)=κ|x|^n that is used to normalize the densities p(x) and p(v).","marker":"[28]"},{"why":"Provides the ergodic observation principle that a single long deterministic trajectory yields the probability densities of finding the system in a given state.","marker":"[6]"},{"why":"Gives Bertrand's theorem, used to identify which two-dimensional central potentials have closed orbits and to motivate the non-closed-orbit analysis for n≠2.","marker":"[33]"},{"why":"Defines the Lévy-walk process with random velocity flips whose spatiotemporal coupling motivates the randomized motion studied here.","marker":"[18]"},{"why":"Provides the comparison model of hard spheres in a square-well potential whose velocity probability densities resemble those of the present model.","marker":"[27]"}],"fun_headline_variants":["Random velocity flips? Long-run stats stay put","Time randomization doesn't shift asymptotic densities","In single wells, random reversals mimic deterministic orbit","Reordering velocities preserves asymptotic densities","Randomization of flips: no effect on long-time PDFs"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The broadest load-bearing premise is that in the long-time limit the randomized trajectory with velocity reversals explores the entire constant-energy, and in 2D constant-angular-momentum, available region in the same statistical proportions as the deterministic orbit, so that time and ensemble averages coincide; for 2D non-closed orbits and for reversal times with diverging mean this is verified numerically at finite time rather than proven.","fun_headline_variants_meta":{"raw":{"variants":["Random velocity flips? Long-run stats stay put","Time randomization doesn't shift asymptotic densities","In single wells, random reversals mimic deterministic orbit","Reordering velocities preserves asymptotic densities","Randomization of flips: no effect on long-time PDFs"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000268,"raw_usage":{"total_tokens":1630,"prompt_tokens":967,"completion_tokens":663,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":583,"completion_tokens_details":{"reasoning_tokens":591}},"tokens_in":583,"tokens_out":663,"duration_ms":7373,"temperature":1.0,"reasoning_tokens":591,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T15:45:37.414944+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the randomized one-dimensional dynamics to a very large observation time with a fixed initial condition and a waiting-time distribution with finite mean, such as an exponential distribution, and check whether the measured p(x) and p(v) converge to Eqs. (5) and (6) with the prescribed normalization; a persistent mismatch would falsify the central claim. For heavy-tailed waiting times with α<1, simulate to times far beyond t=1000 and test whether the remnant deterministic peak decays to the smooth density, and in 2D with n=4 compare the long-run normalized p(r) histogram against Eq. (D.12) for several initial conditions.","supporting_citations":[{"cited_title":"Dybiec, K","cited_arxiv_id":null,"evidence_quote":"Introduces the hard-velocity-reversal randomized model and provides semi-analytical and numerical densities for selected exponents n, which the present paper's analytical derivation extends."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the period formula for motion in V(x)=κ|x|^n that is used to normalize the densities p(x) and p(v)."},{"cited_title":"Walters,An Introduction to Ergodic Theory(Springer Verlag, Berlin, 1982)","cited_arxiv_id":null,"evidence_quote":"Provides the ergodic observation principle that a single long deterministic trajectory yields the probability densities of finding the system in a given state."},{"cited_title":"Goldstein, C","cited_arxiv_id":null,"evidence_quote":"Gives Bertrand's theorem, used to identify which two-dimensional central potentials have closed orbits and to motivate the non-closed-orbit analysis for n≠2."},{"cited_title":"Zaburdaev, S","cited_arxiv_id":null,"evidence_quote":"Defines the Lévy-walk process with random velocity flips whose spatiotemporal coupling motivates the randomized motion studied here."},{"cited_title":"Scalas, A","cited_arxiv_id":null,"evidence_quote":"Provides the comparison model of hard spheres in a square-well potential whose velocity probability densities resemble those of the present model."}],"review_version":1}