{"id":"e613d219-7f84-45a4-a481-616c0887e738","arxiv_id":"2505.10657","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A Vicsek-like flocking model with index-ordered time delays exhibits a new phase deep in the ordered state: two counter-propagating density bands with opposite transverse velocities.","lead":"This paper simulates a flocking model where particles react to neighbors with tiny time delays, and finds that at low noise the flock breaks into two bands moving with opposite sideways motion. The result suggests that very small asymmetries in how quickly agents respond can create entirely new collective states.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The patterned-phase claim lacks thermodynamic-limit evidence: Binder-cumulant sign change appears only at the largest sizes, no finite-size scaling or wavelength-vs-L analysis is shown, and the proposed index-field mechanism is never directly measured.","rationale":"Good-faith reading: the paper does not claim an analytical derivation; it is a computational discovery paper. The strongest support is the visual phase, the Binder sign change at large N, and the permutation susceptibility. These are real and creditable. However, the load-bearing point for the headline claim is not the mechanism but the status of the 'second transition' itself. Without a finite-size scaling analysis that locates a transition in the thermodynamic limit, the positive Binder cumulant at the two largest sizes could be a pre-asymptotic effect; the negative-to-positive sign change with N is exactly what one would see if a finite-size crossover occurs at a length scale larger than the smaller boxes. The absence of any wavelength-vs-L measurement is especially serious because the pattern is presented as two bands spanning the box; a finite-wavelength spatial pattern should have k_max independent of L, while phase separation or a single-domain wall would give k_max ~ 1/L. The paper's own statements—that the index field is 'prohibitively difficult to measure', that no analytical mapping was found, and that there is 'no clear mechanism' for spatial index structure—flag the mechanistic part as unresolved. I agree with the reader's conditional verdict; the concern here is a sharper formulation of the same missing evidence, focusing on the phase's thermodynamic status rather than the index-field explanation. If the proposed structure-factor and finite-size-scaling tests confirm a constant k_max and a common crossing, the concern is resolved.","tokens_in":9058,"tokens_out":12628,"duration_ms":139827,"concrete_test":"Run at fixed η=0.02 (and η=0.03) for N = 2^15, 2^16, 2^17, 2^18 with at least 10 independent seeds each. Compute the structure factor S(k) of the density and of v⊥ along the flocking direction and extract the peak position k_max; also compute Binder cumulant with bootstrap error bars. If k_max trends as 1/L (or the number of bands remains fixed at two with band width proportional to L), the pattern is a finite-size or macroscopic-demixing effect. If k_max approaches a finite constant and the Binder cumulants for different L cross at a common η_c within errors, the transition claim is supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is a second thermodynamic transition into a spatially patterned phase. The evidence for this is Fig. 3: Binder cumulants of the tilt angle become positive only for N ≥ 2^16, while all smaller systems give negative values; no error bars, no crossing analysis, and no finite-size scaling to L→∞ are presented. Given the well-known finite-size sensitivity of Vicsek-like models (cited by the authors), this does not by itself establish a bulk transition. A second, more specific problem is that the order parameter used is a Binder cumulant of a scalar tilt distribution, while the actual pattern is a spatial modulation of the transverse velocity; the manuscript never reports the dependence of the modulation wavelength on L. In Fig. 1 and Fig. 4 the pattern appears as two bands / one sinusoidal mode across the box, so if k_max ~ 1/L the 'patterned phase' is a macroscopic two-domain coexistence rather than a finite-wavelength spatially patterned phase. The paper's own limitations—'prohibitively difficult to measure' index field, 'not able to find an analytical mapping', and 'no clear mechanism' by which the negative damping would produce spatial index structure—mean the mechanistic explanation is also unverified. The permutation test only shows the phase requires fixed index ordering, not that a persistent spatial index-order field exists.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces a Vicsek-like model with an index-ordered update rule (Eq. 4) that breaks reaction-time symmetry within a single timestep, mimicking a distribution of reaction times. The authors report that, in addition to the usual disorder-to-order transition, the model exhibits a second transition deep inside the polar flocking phase to a spatially patterned state: two high-density bands propagate together in the global flocking direction with opposite transverse velocities. The transition is characterized with a Binder cumulant of angles relative to the flocking direction (Fig. 3), number fluctuations (Fig. 2), sinusoidal longitudinal profiles of transverse velocity and of the non-reciprocal torque field Ψ (Fig. 4), and a susceptibility to random permutations of particle indices (Supplemental Fig. 5). The paper argues that the patterned phase is stabilized by a slow-relaxing spatial organization of the particle-index field.","tokens_in":9312,"tokens_out":5072,"duration_ms":52145,"significance":"If the second transition survives the thermodynamic limit, this is a genuinely interesting result: it shows that a minimal, seemingly innocuous temporal asymmetry in agent updates can produce a new collective phase beyond the standard Vicsek phenomenology, with potential implications for hydrodynamic descriptions of time-delayed active matter. The work uses large-scale GPU simulations (up to N=2^17, and N=8e5 for the metric model) and includes an independent permutation-based control that ties the patterned phase to index ordering. These are real strengths. However, the central claim of a thermodynamic transition currently rests on finite-size evidence that is not yet quantitatively established, and the proposed index-field mechanism is explicitly acknowledged in the text to be difficult to measure and not analytically mapped. The paper is therefore a promising candidate for publication after substantial revision, but the evidence as presented is not yet conclusive.","major_comments":[{"comment":"The central claim of a second thermodynamic transition is not yet established. The Binder cumulant U_L changes sign only for N >= 2^16, no error bars are shown, and there is no finite-size scaling or crossing analysis that would extrapolate the behavior to L -> infinity. Given the authors' own statement that Vicsek-like models are notoriously sensitive to finite-size effects, a sign change at the largest sizes could be a finite-size crossover rather than a bulk phase transition. Please provide U_L vs. noise for several sizes with statistical uncertainties, and either a finite-size collapse or an explicit discussion of how the transition point and order-parameter distribution behave as L grows.","section":"Fig. 3 and 'Order parameters for the spatially patterned flocking phase'"},{"comment":"The order parameter used in Fig. 3 is a Binder cumulant of scalar relative angles, which does not directly probe the spatial modulation that defines the patterned phase. The manuscript never reports the wavelength or dominant Fourier mode of the transverse-velocity pattern as a function of system size L. In Fig. 1 and Fig. 4 the pattern appears as a single sinusoidal mode across the box; if the dominant mode scales as k_max ~ 1/L, the state is better described as macroscopic two-domain coexistence (a bulk phase separation into two oppositely moving bands) rather than a finite-wavelength spatially patterned phase. Please report the L-dependence of the pattern wavelength and, if possible, the number of bands in larger systems.","section":"Figs. 1 and 4"},{"comment":"The mechanistic explanation in the abstract and main text is stronger than the evidence. The text states that the relative index field is 'prohibitively difficult to measure at levels beyond the noise' and that there is 'no clear mechanism' by which negative damping would create spatial structure in the index field. The permutation susceptibility in Supplemental Fig. 5 demonstrates only that the patterned phase requires fixed index labels; it does not directly establish the existence of a persistent, slowly relaxing spatial index-order field. The claim that stability is 'directly tied to a subtle spatial organization' of an index-order field therefore needs either a direct measurement of a coarse-grained index gradient (or a proxy that tracks index-order persistence over time), or a more cautious statement that the mechanism is a plausible but unverified hypothesis.","section":"Discussion of the index field and Supplemental Fig. 5"},{"comment":"The paper does not report statistical uncertainties for any of the key order parameters: the Binder cumulant U_L, the sinusoidal amplitude A in the Fig. 2 inset, and the profiles in Fig. 4. Because the sign of U_L at large N is the primary quantitative evidence for the transition, error bars or block-averaged standard errors are necessary to determine whether the sign change is robust and to support any subsequent finite-size analysis. Please add error bars and specify how many independent runs or blocks were used.","section":"Figs. 2 and 3"}],"minor_comments":[{"comment":"Equation (2) appears to be missing the current angle theta_i(t) on the right-hand side; the orientation update should read theta_i(t + dt) = theta_i(t) + dt * tau_i + eta sqrt(dt) * zeta_i(t), or the definitions should be clarified.","section":"Eq. (2)"},{"comment":"The caption reads 'N = 217 particles'; this should be 'N = 2^17 particles' (or the equivalent superscript notation) to match the text.","section":"Fig. 1 caption"},{"comment":"The caption of Fig. 5 does not define whether the plotted change in the global order parameter is the absolute value, the signed difference, or an ensemble-averaged magnitude, nor over what time window it is measured. Please clarify.","section":"Fig. 5 caption"},{"comment":"The definition 'theta_i = arcsin(n_i . v_perp)' gives only the transverse component of the unit director, not a signed angle about v_parallel; using atan2 of the transverse and parallel components would be more transparent and would make the bimodal distributions easier to interpret.","section":"Fig. 2 top panel"},{"comment":"Reference [43] is listed as '[url to be inserted by publisher]'; in the current preprint this should point to the actual Supplemental Material or be replaced by a stable identifier.","section":"References"},{"comment":"The text should state explicitly that the update in Eq. (4) is performed sequentially in increasing index order, so that the theta_j(t + dt) term for i > j refers to neighbors updated earlier in the same timestep; this is clear from context but would benefit from being stated directly.","section":"Eq. (4)"}],"recommendation":"major_revision","confidential_remarks":"The manuscript fits the scope of the journal and the central observation is likely of interest even if the thermodynamic interpretation is not yet conclusive. The main risk is overclaiming the transition and the index-field mechanism; both are addressable with additional simulation analysis and more cautious wording. I see no reason to suspect the simulations themselves are flawed, but the quantitative evidence needs strengthening before publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe headline: this paper reports a genuinely new computational result. The index-ordered update rule, where low-index neighbors contribute time-local torques and high-index neighbors time-delayed torques within a single timestep, is a clever minimal construction. The resulting phase—two high-density bands that move together in the global flocking direction but with opposite transverse velocities—is not in the existing time-delay Vicsek literature. That is a real phenomenon worth taking seriously.\n\nWhat the paper does well: it probes the phase from several directions. The Binder cumulant turning positive only for N ≥ 2^16 indicates a bimodal distribution that emerges with system size. The index-permutation susceptibility is a smart control: shuffling indices destroys the pattern, showing that the phase genuinely depends on index ordering. The measured non-reciprocal torque field Ψ has a sinusoidal profile out of phase with the transverse velocity, consistent with the proposed mechanism. The authors also show a metric version of the model exhibits similar behavior, which argues against the pattern being a topological-neighbor artifact. They are honest about what they cannot do: measure the index field directly, or find an analytical mapping.\n\nThe soft spots are real and non-trivial. The thermodynamic-limit evidence is thin. There are no error bars on the Binder cumulant or other order parameters, no finite-size scaling extrapolation, and no analysis of the modulation wavelength as a function of L. The patterns shown in Figs. 1 and 4 are a single sinusoidal mode across the box, so the wavelength scales with L. That means the 'patterned phase' could be a long-wavelength two-domain coexistence rather than a spontaneously selected finite-wavelength modulation. If the wavelength keeps growing with system size, the thermodynamic phase may be homogeneous or something else. The stress-test note makes this point correctly, and the paper does not address it. Also, the index-field mechanism is never directly measured; the permutation test demonstrates necessity of index ordering but not the existence of a slow-relaxing spatial index field.\n\nOn balance, the central observation is plausible and novel, but the paper reads as a compelling first report, not a closed case. For a letter, that can be enough if the revisions add error bars and a wavelength-vs-L check. The citation pattern to the time-delay Vicsek literature looks appropriate; the novelty claim against Refs. 35 and 38 holds up.\n\nMy recommendation: send it to peer review. A serious referee can push for the missing finite-size analysis. Even if the phase turns out to be a finite-size effect, identifying this class of behavior in time-asymmetric flocking is a useful contribution. I would bring it to a reading group and cite it.","headline":"A genuinely new computational observation of a second, spatially modulated phase deep in the polar flocking state, driven by an index-ordered time-asymmetric update; the main weakness is the lack of thermodynamic-limit evidence.","tokens_in":9870,"tokens_out":2506,"would_cite":true,"duration_ms":26293,"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":"Time-delayed alignment splits an ordered flock into counter-moving bands.","keywords":["flocking","Vicsek model","time delay","reaction-time symmetry","non-reciprocal interactions","active matter","banded phase","Binder cumulant"],"falsifier":"Simulate the model with the same parameters but replace the full index-ordered update with a two-species rule (half the particles react instantly, half with a one-timestep delay) and measure the transverse-velocity order parameter; if no bimodal banded phase appears in that simpler rule, the slow-relaxing index-order mechanism—not mere time delay—is what the phase depends on. Alternatively, directly measure the autocorrelation time of a local index-gradient observable; if it decays on a timescale much shorter than the lifetime of the bands, the proposed mechanism fails.","tokens_in":8837,"feed_emoji":"🐦","tokens_out":6512,"duration_ms":60502,"temperature":0.7,"pith_summary":"This paper introduces a minimal flocking model in which equal-time alignment is replaced by an index-ordered update rule: lower-index particles react to higher-index neighbors with a one-timestep delay. The model keeps the standard disorder-to-order transition of Vicsek-like flocks, but deep inside the ordered phase it finds a second transition to a spatially patterned state—two high-density bands travel together along the global flocking direction while moving in opposite transverse directions. The authors argue that this phase is stabilized by a slowly relaxing spatial organization of the particle index field, and they support that claim by showing the pattern is destroyed when indices are randomly permuted and that the non-reciprocal torque field has a sinusoidal longitudinal profile. The wider message is that even a tiny asymmetry in reaction time, occurring entirely within one numerical timestep, can change the collective state of active matter.","feed_headline":"Time-delayed alignment splits flocks into counter-moving bands","feed_subtitle":"A minimal flocking model keeps its usual order transition, then adds a second one deep inside the ordered phase.","key_machinery":"The central object is the index-ordered update rule (Eq. 4): within each timestep, particle i receives time-local information from lower-index neighbors and one-timestep-delayed information from higher-index neighbors. This creates a non-reciprocal, reaction-time-asymmetric interaction. The load-bearing quantity is the spatial organization of the particle index field—the relative index differences between neighboring particles—which the paper argues organizes along the flocking direction and relaxes slowly, although it cannot be measured directly beyond noise. Its effect is captured by Ψ, the difference between the torques actually applied and the torques a time-local Hamiltonian would produce; the sinusoidal longitudinal profile of Ψ supplies the persistent torque that keeps the two bands from aligning into one homogeneous flock.","core_discovery":"In a Vicsek-like flocking model with torques derived from a ferromagnetic Hamiltonian and Voronoi neighbor lists, the authors replace the simultaneous update rule with an index-ordered rule in which particle i aligns to lower-index neighbors using their updated orientations and to higher-index neighbors using their old orientations. This breaks reaction-time symmetry within a single timestep. Below the usual flocking transition, at low noise, the system develops a new phase in which the flock remains globally polarized but separates into two dense bands with opposite transverse velocities. The Binder cumulant of angles relative to the flocking direction becomes positive for large systems (N ≳ $2^{16}$), indicating a genuinely bimodal distribution rather than long tails, and giant number fluctuations persist. The authors show the phase is unstable to random permutation of particle indices and that the longitudinal profile of Ψ—the difference between the index-ordered torques and the time-local Hamiltonian torques—is sinusoidal and out of phase with the transverse-velocity profile. They conclude that a slowly relaxing spatial organization of the index-order field sustains the counter-propagating bands.","pith_inferences":["A direct testable extension: a two-species version of the model (one species delayed by one timestep, the other reacting instantly) should reproduce the patterned phase if reaction-time asymmetry is the operative ingredient; the paper flags this as future work.","The Ψ measurement implies that time-delayed alignment is equivalent, at a coarse-grained level, to time-local non-reciprocal torques; if so, other models with explicit non-reciprocal alignment should exhibit analogous banded phases, and an analytic mapping might be derivable via Markovian embedding.","Because the index-order field is slow-relaxing yet unmeasurable, the patterned phase could represent a genuinely new universality class whose hydrodynamic description couples to an index-gradient field rather than standard polar-flock fields; a continuum derivation would be the next step.","The finite-size requirement (N ≳ 2^16) suggests the phase may be hard to observe in small experimental groups; detecting it would require large flocks with measurable reaction-time distributions."],"forward_implications":["The standard disorder-to-order transition survives the reaction-time asymmetry, so the patterned phase is an additional ordering transition nested inside the polar flocking phase rather than a replacement for it.","The pattern requires large systems: Binder cumulants show bimodality only for N ≳ 2^16, so smaller simulations would miss the phase entirely.","Because random permutation of indices destroys the bands, any model that labels particles symmetrically cannot produce this phase; the index ordering is a genuine degree of freedom.","A metric version of the model shows similar but spatially more complex banded phases, suggesting the mechanism is not an artifact of Voronoi neighbor lists.","The sinusoidal, out-of-phase relationship between Ψ and the transverse velocity provides a coarse-grained signature that could be measured in other time-delayed flocking models."],"supporting_citations":[{"why":"Defines the baseline model of self-propelled particles whose order-disorder transition this paper retains and extends.","marker":"[3]"},{"why":"Establishes the hydrodynamic theory of polar flocks that the new patterned phase would need to be incorporated into.","marker":"[9]"},{"why":"Provides empirical evidence of hierarchical reaction delays in pigeon flocks that motivates the index-ordered delay construction.","marker":"[27]"},{"why":"Supplies the GPU-accelerated simulation code used for the large-scale Voronoi-neighbor simulations.","marker":"[42]"},{"why":"Furnishes the finite-size scaling and number-fluctuation expectations used to place the patterned phase relative to the known flocking universality class.","marker":"[45]"},{"why":"Shows how time-delayed systems can be represented through Markovian embedding, supporting the paper's link between time delays and non-reciprocal torques.","marker":"[46]"},{"why":"Prior work by the authors on banded phases in topological flocks, providing context for banding regimes in this model family.","marker":"[8]"}],"fun_headline_variants":["Time delays split flocks into opposing bands","Delayed alignment drives flocks into banded phases","Delay-induced symmetry breaking creates banded flocks","Slow reactions make flocks form counter-moving bands"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The patterned phase survives only if the spatial ordering of particle indices persists over long times; the paper states this index-order field is prohibitively difficult to measure beyond noise, so its stability is inferred indirectly from the destruction of the pattern under index permutation and from the sinusoidal torque field Ψ.","fun_headline_variants_meta":{"raw":{"variants":["Time delays split flocks into opposing bands","Delayed alignment drives flocks into banded phases","Delay-induced symmetry breaking creates banded flocks","Slow reactions make flocks form counter-moving bands"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000727,"raw_usage":{"total_tokens":3256,"prompt_tokens":941,"completion_tokens":2315,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":557,"completion_tokens_details":{"reasoning_tokens":2260}},"tokens_in":557,"tokens_out":2315,"duration_ms":15712,"temperature":1.0,"reasoning_tokens":2260,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T21:06:09.723112+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Simulate the model with the same parameters but replace the full index-ordered update with a two-species rule (half the particles react instantly, half with a one-timestep delay) and measure the transverse-velocity order parameter; if no bimodal banded phase appears in that simpler rule, the slow-relaxing index-order mechanism—not mere time delay—is what the phase depends on. Alternatively, directly measure the autocorrelation time of a local index-gradient observable; if it decays on a timescale much shorter than the lifetime of the bands, the proposed mechanism fails.","supporting_citations":[{"cited_title":"Toner and Y","cited_arxiv_id":null,"evidence_quote":"Establishes the hydrodynamic theory of polar flocks that the new patterned phase would need to be incorporated into."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides empirical evidence of hierarchical reaction delays in pigeon flocks that motivates the index-ordered delay construction."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the GPU-accelerated simulation code used for the large-scale Voronoi-neighbor simulations."},{"cited_title":"Chat´ e, F","cited_arxiv_id":null,"evidence_quote":"Furnishes the finite-size scaling and number-fluctuation expectations used to place the patterned phase relative to the known flocking universality class."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows how time-delayed systems can be represented through Markovian embedding, supporting the paper's link between time delays and non-reciprocal torques."}],"review_version":1}