{"id":"bd58b92d-0195-4b52-8661-1900286da030","arxiv_id":"1908.03499","paper_version":3,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"This paper derives exact finite-time maximum and minimum statistics for biased random walks, links the relaxation spectrum to the Marchenko-Pastur law of random matrices, and applies the results to molecular motors and sports scoring.","lead":"Scientists derived exact formulas for the largest and smallest excursions of a biased random walker, a model for molecular motors, extending earlier bounds on entropy production. The work also shows that the relaxation times of these extremes obey the same statistical law as the eigenvalues of random matrices, and illustrates the ideas with soccer and basketball scoring.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Motor-application claim rests on an underived effective-1D reduction (Eqs. 17-18) that fixes only the total rate, not the drift or noise of S(t); the exact 1D result is not the weak point.","rationale":"Reader's conditional verdict is appropriate. I independently re-derived the 1D generating-function route: the absorption probability identity, the Bessel first-passage density (A19), and the integral representation (B16) are consistent; Eq. (12) and the MP spectrum are a variable change rather than an independent approximation. The symmetry relation (10)-(11) follows from the first-passage duality and is supported by simulation. My main concern is the motor transfer. Eqs. (17)-(18) form a heuristic with no derivation from the 2D master equation, and the paper's wording itself calls it an approximation; a single-parameter Gillespie comparison cannot establish the broad abstract claim. A mismatch between σ_eff and σ_2D for general affinities is plausible because the effective parameters are chosen from a sum of cosh terms, not from the currents. This does not undermine the exact 1D result but does justify keeping the verdict CONDITIONAL pending a wider validation; no change from the reader's verdict.","tokens_in":26947,"tokens_out":16150,"duration_ms":170629,"concrete_test":"Using the Fig. 5 parameter values and Eqs. (17)-(18), compute σ_eff = A_eff(2ν_eff)sinh(A_eff/2) and compare with σ_2D = Σ_α 2ν_α A_α sinh(A_α/2), where A_α ∈ {A_m, A_c, A_m+A_c, A_c−A_m}. If the two differ by more than ~20%, the effective 1D walk has the wrong mean entropy drift and the agreement in Fig. 5b cannot be general; then rerun Gillespie at a second, far-from-fitted parameter set (e.g. Δμ=20 k_BT, f_ext near stall) and test whether Eq. (9) with A_eff,ν_eff still reproduces ⟨Smin(t)⟩ to the claimed accuracy.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The exact 1D extremal statistics are internally consistent: Eq. (9) reduces correctly at t→0 and t→∞, and the Marchenko-Pastur representation is a change of variables in Eq. (B16). The load-bearing weakness is the transfer to the 2D molecular-motor model. There S(t)=A_m X(t)+A_c Y(t) is a random walk with four distinct jump affinities A_m, A_c, A_m+A_c and A_c−A_m, not a single ±A_eff walk. Equations (17)-(18) only require k_eff^+ + k_eff^- = Σ_α(k_α^+ + k_α^-), i.e. they match the total jump rate. They impose no matching of the mean entropy current σ_2D = Σ_α 2ν_α A_α sinh(A_α/2) nor of the entropy diffusivity D_2D = Σ_α ν_α A_α^2 cosh(A_α/2). Since finite-time extrema are controlled by both drift and fluctuations, the effective model cannot be exact for generic parameters; for the Fig. 5 rates a direct evaluation of σ_eff versus σ_2D is not reported. Validation is a single Gillespie parameter set, so the abstract-level claim that the approach captures molecular-motor extreme events is only weakly supported. The sports section is explicitly an illustrative Poisson model and is less load-bearing for the paper's scientific claim.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript derives exact finite-time statistics of the extrema (maximum and minimum) of the position and of the entropy production of a one-dimensional continuous-time biased random walk. Starting from the first-passage-time distribution, the authors obtain generating functions and integral representations for the average extrema (Eqs. (9), (12), (14)), show that the relaxation spectrum is the Marchenko-Pastur law, and use this to propose Wishart/Laguerre random-matrix estimators. They then apply the 1D formulas to a two-dimensional stochastic model of a molecular motor through effective parameters (Eqs. (17)-(18)) and illustrate the relaxation-time spectra with scoring data from soccer, handball, and basketball.","tokens_in":27158,"tokens_out":7274,"duration_ms":78049,"significance":"If the 1D derivation is correct, this is a valuable exact contribution: it generalizes the infimum law for entropy production, establishes a symmetry between the distributions of minima and maxima (Eq. (11)), provides a finite-time supremum bound (Eq. (21)), and connects finite-time extreme-value statistics to random-matrix theory. The 1D derivations are largely self-contained and internally consistent; I verified the t->0 and t->infinity limits of Eq. (12), the change of variables leading to the Marchenko-Pastur integral in Appendix C, and the duality argument behind Eq. (11). The Gillespie simulations in Figs. 2 and 3 support the 1D results. The main weakness is the transfer to the 2D molecular-motor model: the effective-1D reduction is heuristic and is validated for a single parameter set, so the abstract-level claim that the approach captures key features of motor extreme events is not yet established. The sports section is explicitly illustrative and is not the main scientific claim.","major_comments":[{"comment":"The reduction of the 2D molecular-motor model to an effective 1D biased walk is not derived from the 2D master equation. In the 2D model S(t)=A_m X(t)+A_c Y(t) changes by four step affinities A_m, A_c, A_m+A_c, and A_c-A_m, with mean entropy current sigma_2D = sum_alpha 2 nu_alpha A_alpha sinh(A_alpha/2) and entropy diffusivity D_2D = sum_alpha nu_alpha A_alpha^2 cosh(A_alpha/2). Equations (17)-(18) match only the total jump rate k_eff^+ + k_eff^- = sum_alpha (k_alpha^+ + k_alpha^-); they impose no matching of sigma_2D or D_2D. Since finite-time extrema are controlled by both drift and fluctuations, the effective model cannot be exact for generic parameters. The validation in Fig. 5 is a single Gillespie parameter set, and no comparison of sigma_eff and D_eff with sigma_2D and D_2D is reported. I ask the authors to derive or justify the reduction, to report these cumulant comparisons for the parameters used, and to test the approximation over a range of affinities; without this, the abstract's claim about molecular-motor extreme events is not supported.","section":"IV, Eqs. (17)-(18)"},{"comment":"The distribution comparison in Fig. 5(c) uses Eq. (3), which is the infinite-time global minimum distribution, to validate simulation data at t=50 ms. The paper's central exact result is the finite-time statistics encoded in Eqs. (5) and (9); Fig. 5(b) tests only the average finite-time behavior, while Fig. 5(c) tests only the long-time limit. Please compare the motor simulations to the finite-time distribution (5) or the corresponding generating function, and state the relevant relaxation times for the motor parameters so the reader can judge whether t=50 ms is indeed in the asymptotic regime. As written, the finite-time distribution claim for the motor model is not validated.","section":"IV, Fig. 5(c)"}],"minor_comments":[{"comment":"The caption states that a single 64x64 random matrix is used, while the legend and text report m=22, 24, and 26; please reconcile these numbers.","section":"Fig. 4 caption"},{"comment":"The notation for the Kampe de Feriet function in Eq. (E5), written as '2+0F1+1', is very hard to parse; please give the explicit parameter vectors or a reference with the exact convention used.","section":"Appendix E, Eq. (E5)"},{"comment":"The inset shows curves for 0.5 <= A <= 5 with nu=1, but there is no legend identifying which curve corresponds to which A; add a legend or a color scale.","section":"Fig. 2 inset"},{"comment":"The sentence 'For women's soccer matches one that therefore does not expect a comeback' contains a typo ('one that' should be 'one does'), and 'the weaker time falls behind' should read 'the weaker team'; also, the sports conclusions should explicitly remind the reader that the model assumes fixed Poisson scoring rates with no correlations or non-stationarity, as this assumption is not tested against within-match data.","section":"Sec. V"}],"recommendation":"major_revision","confidential_remarks":"The exact 1D results appear sound and are a genuine contribution, but the title and abstract foreground molecular motors, and the motor section rests on an underived effective-1D approximation validated only for one parameter set. I recommend major revision: the authors should either derive the reduction, provide a substantially broader numerical validation including cumulant comparisons, or substantially soften the claims about molecular motors. The sports section is illustrative and should be presented as such."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The core of this paper is the exact finite-time generating function for the extrema of the 1D biased random walk, together with the Marchenko-Pastur relaxation spectrum and the max-min symmetry. Those are new, cleanly derived, and internally consistent. I checked the t→0 and t→∞ limits of Eq. (12), the change of variables to the MP density, and the duality argument behind Eq. (11) — all correct. The Gillespie simulations in Figs. 2–3 and 5–6 match the formulas, which is gratifying. The random-matrix estimators (Wishart and Laguerre) are a clever practical device and the convergence discussion in Appendix D is honest about the bias.\n\nThe soft spots are all at the application layer, and they are real but not fatal. The effective-1D reduction for the molecular motor (Eqs. 17–18) matches only the total jump rate. It does not match the mean entropy current or the entropy diffusivity of the 2D model, so there is no reason it should be exact for generic parameters. The authors know this—they call it an approximation—but the abstract says the approach “captures key features” of motor extreme events, and that claim is supported by a single parameter set. That needs either a derivation from the 2D master equation or a systematic scan over a range of affinities and rates. The sports section is explicitly illustrative and less load-bearing; still, the prescriptive comment about shortening women's matches overreaches, and the text in Sec. V muddles τ0 with the mean waiting time τ1. The global-infimum result is recycled from the authors' own Ref. [31], but that is properly cited and it is a component, not the contribution.\n\nWho gets value: stochastic thermodynamics researchers who want exact finite-time extrema statistics for biased random walks, and anyone interested in the random-matrix connection. The motor and sports applications should be read as examples, not as demonstrated universal claims. The paper deserves a serious referee — the maths is solid and the central result is publishable — but the motor-reduction claim needs to be tempered or strengthened before acceptance. I would send it to peer review, expecting major revision on the application sections.","headline":"The exact finite-time extrema statistics for the 1D biased random walk are the real prize here; the molecular-motor application rests on an under-validated effective reduction and the sports section is illustrative only.","tokens_in":27850,"tokens_out":1130,"would_cite":true,"duration_ms":13729,"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 shows that the complete finite-time statistics of the maximum and minimum of a biased random walk's position and entropy production are encoded in a single generating function, whose relaxation spectrum is the Marchenko-Pastur…","keywords":["biased random walk","extreme value statistics","entropy production","first-passage times","Marchenko-Pastur distribution","random matrix theory","molecular motors","sports scoring"],"falsifier":"Perform a high-precision simulation of a 1D biased random walk at, say, $A=1$, $\\nu=1$, and compare the empirical distributions of $-S_{\\min}(t)$ and $S_{\\max}(t)-S(t)$ at $t=0.1, 1, 10$ against the paper's integral formula and the symmetry relation. Any statistically significant violation of the mirror symmetry, or a relaxation spectrum estimated from the time series that fails to match the Marchenko-Pastur density with parameter $\\delta=e^{-A}$, would falsify the central claim.","tokens_in":26576,"feed_emoji":"📊","tokens_out":12678,"duration_ms":123879,"temperature":0.7,"pith_summary":"The paper establishes that for a continuous-time biased random walk, the complete finite-time statistics of the maximum and minimum of displacement and entropy production are exactly computable from one generating function. Every moment and every extremal probability is fixed by two parameters, the bias $A$ and the jump rate $\\nu$, with all time dependence carried by a relaxation spectrum that is exactly the Marchenko-Pastur law of random-matrix theory. The same formulas yield a symmetry between the distributions of minima and maxima of entropy production and a bound on how much entropy production can exceed its running average. A sympathetic reader cares because these quantities answer practical questions—how far a molecular motor backs up against its bias, or how long the weaker team can lead—and the paper gives closed-form answers rather than asymptotics alone.","feed_headline":"Extremes of biased random walks follow the Marchenko-Pastur law","feed_subtitle":"Exact formulas give every moment of the minimum and maximum of position and entropy, at any finite time.","key_machinery":"The central object is the generating function $G_{\\min}(z;t)$ of the minimum distribution, expressed as an integral against the Marchenko-Pastur law—the limiting eigenvalue density of Wishart-Laguerre random matrices—on the relaxation-time interval $[\\tau_0,\\tau_\\infty]$, with $\\tau_0=(\\sqrt{k_+}+\\sqrt{k_-})^{-2}$ and $\\tau_\\infty=(\\sqrt{k_+}-\\sqrt{k_-})^{-2}$. The machinery is the Laplace transform of the first-passage-time density, $\\hat P_{\\rm fpt}(s;x)=e^{Ax/2}e^{-|x|\\cosh^{-1}(s/2\\nu+\\cosh(A/2))}$, whose algebraic form turns the convolution over repeated first passages into the Marchenko-Pastur integral. The same first-passage duality $\\tilde P_{\\rm fpt}(T;x)=P_{\\rm fpt}(T;-x)$ produces the max-min symmetry, and the effective-parameter formulas $\\nu_{\\rm eff}=\\sum_\\alpha\\nu_\\alpha$, $A_{\\rm eff}=2\\cosh^{-1}\\left(\\sum_\\alpha\\frac{\\nu_\\alpha}{\\nu_{\\rm eff}}\\cosh\\frac{A_\\alpha}{2}\\right)$ extend the 1D result to the 2D molecular-motor model.","core_discovery":"The paper's central claim is that for a 1D continuous-time biased random walk with rates $k_\\pm=\\nu e^{\\pm A/2}$, the generating function of the finite-time minimum is exactly $G_{\\min}(z;t)=1+\\frac{1-z}{e^A-1}\\int_{\\tau_0}^{\\tau_\\infty}\\frac{1-e^{-t/\\tau}}{1+f(z)\\,\\tau}\\rho(\\tau/\\bar\\tau)\\,\\frac{d\\tau}{\\bar\\tau}$, with $f(z)=k_+(z-1)+k_-(z^{-1}-1)$ and $\\rho$ the Marchenko-Pastur density. All moments and probabilities of the minimum, and by a mirror symmetry likewise of the maximum, follow by differentiating this expression. The paper proves $P(S(0)-S_{\\min}(t)=s)=P(S_{\\max}(t)-S(t)=s)$ and derives the supremal bound $\\langle S_{\\max}(t)\\rangle-\\langle S(t)\\rangle\\le 1$, generalizing the infimum law of stochastic thermodynamics. It further shows that the finite-time average minimum relaxes through the Marchenko-Pastur spectrum, so the extreme-value statistics of the walk can be estimated from the eigenvalues of a single Wishart or $\\beta$-Laguerre random matrix.","pith_inferences":["The same Marchenko-Pastur structure should reappear for any Markov process whose Laplace-transformed first-passage density is an algebraic function of one master quantity, which would make extreme-value statistics of a broader class of currents estimable from random-matrix spectra.","The mirror symmetry could be tested experimentally as an extremal fluctuation theorem: recording only minima in single-molecule trajectories would predict the distribution of maxima above the running value, and deviations would flag non-Markovian effects.","The sports analysis is testable with play-by-play score data: if real scoring rates are time-dependent, the empirical extrema should deviate systematically from the Poisson fixed-rate prediction, and the deviation would quantify the minimal model's limits.","Near stall, the predicted divergence of average extreme spatial steps could be probed in single-motor assays, since the divergence timescale depends on $A_{\\rm eff}$ in a specific way and load-dependent measurements would discriminate the effective 1D mapping from alternative multi-state descriptions."],"forward_implications":["Exact finite-time probabilities and moments for the extrema of any 1D biased random walk can be computed directly from the generating function, so no simulation is needed for the idealized model.","The distribution of the maximum of entropy production above its running value is the mirror image of the distribution of the minimum below its initial value at every time; measuring one gives the other.","A single sufficiently large Wishart or $\\beta$-Laguerre random matrix with rectangularity $e^{-A}$ estimates the full time dependence of the average minimum to within a few percent in the tested cases.","For the two-dimensional motor model, the effective parameters of Eqs. (17)-(18) reproduce the simulated distributions of entropy extrema, and the average extreme displacements grow as the external force approaches the stall force while entropy extremes remain bounded.","For sports scores, the comeback window for the weaker team is bounded by $\\tau_0\\le t\\le \\tau_\\infty$; the data suggest that in the 2018-19 Women's Champions League this window closes at roughly 60 minutes, while for Men's it extends past the 90-minute match."],"supporting_citations":[{"why":"Supplies the infimum law for entropy production and the martingale methods at stopping times that the paper generalizes to finite-time extrema and to the supremum bound.","marker":"[31]"},{"why":"Provides the first-passage-time formalism used to write the exact absorption probability and first-passage density for the biased walk.","marker":"[54]"},{"why":"Gives the Marchenko-Pastur limiting eigenvalue density that the paper identifies as the relaxation spectrum of the extrema generating function.","marker":"[59]"},{"why":"Establishes the convergence of Wishart-Laguerre eigenvalue distributions to the Marchenko-Pastur law, justifying the random-matrix estimates.","marker":"[62]"},{"why":"Defines the beta-Laguerre matrix ensembles used to realize the Marchenko-Pastur spectrum for arbitrary bias and to build the numerical estimators.","marker":"[63]"}],"fun_headline_variants":["Exact finite-time extremes for biased walks: Marchenko-Pastur","Marchenko-Pastur spectrum sets extremes of random walks","Symmetry of max and min extremes in stochastic transport","Random-matrix eigenvalues predict walk extrema exactly"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The motor and sports conclusions depend on compressing a genuinely multidimensional, time-varying process into a one-dimensional biased walk with fixed rates; the compression is exact for the 1D model, but for the molecular motor it is an unproven approximation checked at only one parameter set.","fun_headline_variants_meta":{"raw":{"variants":["Exact finite-time extremes for biased walks: Marchenko-Pastur","Marchenko-Pastur spectrum sets extremes of random walks","Symmetry of max and min extremes in stochastic transport","Random-matrix eigenvalues predict walk extrema exactly"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000243,"raw_usage":{"total_tokens":1539,"prompt_tokens":965,"completion_tokens":574,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":581,"completion_tokens_details":{"reasoning_tokens":507}},"tokens_in":581,"tokens_out":574,"duration_ms":6109,"temperature":1.0,"reasoning_tokens":507,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:13:06.952203+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Perform a high-precision simulation of a 1D biased random walk at, say, $A=1$, $\\nu=1$, and compare the empirical distributions of $-S_{\\min}(t)$ and $S_{\\max}(t)-S(t)$ at $t=0.1, 1, 10$ against the paper's integral formula and the symmetry relation. Any statistically significant violation of the mirror symmetry, or a relaxation spectrum estimated from the time series that fails to match the Marchenko-Pastur density with parameter $\\delta=e^{-A}$, would falsify the central claim.","supporting_citations":[{"cited_title":"Kinetic characterization of heat bath and the energetics of thermal ratchet models,","cited_arxiv_id":null,"evidence_quote":"Supplies the infimum law for entropy production and the martingale methods at stopping times that the paper generalizes to finite-time extrema and to the supremum bound."},{"cited_title":"Universal statistics of longest lasting records of random walks and l´ evy flights,","cited_arxiv_id":null,"evidence_quote":"Provides the first-passage-time formalism used to write the exact absorption probability and first-passage density for the biased walk."},{"cited_title":"1 (Elsevier, 1992)","cited_arxiv_id":null,"evidence_quote":"Gives the Marchenko-Pastur limiting eigenvalue density that the paper identifies as the relaxation spectrum of the extrema generating function."},{"cited_title":"Generic properties of stochastic entropy pro- duction,","cited_arxiv_id":null,"evidence_quote":"Establishes the convergence of Wishart-Laguerre eigenvalue distributions to the Marchenko-Pastur law, justifying the random-matrix estimates."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the beta-Laguerre matrix ensembles used to realize the Marchenko-Pastur spectrum for arbitrary bias and to build the numerical estimators."}],"review_version":1}