{"id":"3cbbff93-0a73-48df-abf1-e35f38a48c4b","arxiv_id":"2606.01907","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Group-voids identified from galaxy group catalogs reproduce halo-void size functions and density profiles in CSST mocks above ~40% spectroscopic completeness, with incompleteness at 30% captured by an effective redshift-error parameter.","lead":"This paper tests whether cosmic voids found in galaxy group catalogs can stand in for voids found in dark-matter halos, using mock data for the planned CSST space telescope. It finds they match when about 40% of galaxies have accurate redshifts, and even at 30% if simulations add a calibrated redshift-error term.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Post-hoc Gaussian σ0 fit to the same VSF data it claims to reproduce leaves the '30% completeness is acceptable' conclusion circular and untested; no predictive validation is shown.","rationale":"The central claim of the paper is that group-voids can replace halo-voids in simulations, with quantitative thresholds of 40% and 30% spectroscopic completeness. The ideal-case (100% spec-z) comparison is the strongest part: zspec group-void VSFs and density profiles are consistent with halo-voids, which is a credible demonstration of the core method. The load-bearing weakness lies in the incomplete-spectroscopy extension. The paper's only support for the 30% threshold is the σ0 model, but the model is calibrated post hoc on the same data it is used to validate. This makes the 'reliably describe' claim circular and leaves it without predictive power. The reader's weakest assumption identified exactly this issue, and I agree: the equivalence between spectroscopic incompleteness plus photo-z errors and a single Gaussian halo redshift error is not established by the present analysis. The concrete test of generating multiple photo-z realizations would directly expose whether the σ0 model is predictive or merely a curve-fitting device. Because the ideal-case result remains valid and the overclaim is fixable with a proper calibration, the CONDITIONAL verdict is appropriate; no change is needed. I see no basis for REJECT, and the paper's core methodological proposal is sound enough to warrant further development.","tokens_in":15589,"tokens_out":7849,"duration_ms":80411,"concrete_test":"Generate ~50 independent realizations of the conservative galaxy catalog by re-sampling the photometric redshift error for galaxies without spectroscopic redshifts, using the same σz=(0.01+0.015z_spec)(1+z_spec) as in §2.1. Run the group finder and VIDE on each realization, and compute the group-void VSF in each redshift bin. Compare the resulting distribution of VSFs to the halo-void VSF with the claimed σ0 values. If the σ0-perturbed halo VSF falls outside the 2σ band of the realization distribution in any bin, or if a likelihood fit yields σ0 values that vary between realizations by more than the bin-to-bin variation, the single-parameter Gaussian model is not a faithful description.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 3.1 introduces a free parameter σ0 to model spectroscopic incompleteness: halos are assigned redshifts drawn from a Gaussian with σz=σ0(1+z), and the resulting halo-void VSF is compared with the group-void (zCSST) VSF. The authors then report that σ0=0.005, 0.01, 0.01, and 0.02 'match' in the four redshift bins above z=0.2. This is a post-hoc selection: σ0 is tuned to the very same data used to claim agreement. No goodness-of-fit statistic, uncertainty on σ0, or validation on an independent redshift bin, completeness level, or the void density profile is provided. The text itself concedes 'We intend to investigate this strategy in future work,' confirming the effective-error model is not yet established. Consequently, the abstract's statement that at 30% completeness group-voids can be 'reliably described' via halo-voids with a redshift error term overstates the evidence: the data demonstrate only that some Gaussian smoothing can make the large-radius VSF curves overlap in the chosen bins, not that σ0 is a predictive effective parameter that would make the 30% threshold robust.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes using voids identified in galaxy group catalogs (group-voids) as cosmological probes, arguing that groups are directly linked to dark matter halos and therefore that group-void statistics can be modeled from halo catalogs in simulations, bypassing expensive galaxy-formation modeling. Using the CSST MGRS mock, the authors construct group catalogs in an ideal case (100% spectroscopic redshifts, z_spec) and a conservative case (~30% spectroscopic completeness, z_CSST), identify voids with the VIDE/ZOBOV algorithm, and measure the void size function (VSF) and stacked void density profile in five redshift bins from z=0 to 1. They compare these statistics with halo-voids and further attempt to model the conservative-case VSF by adding a Gaussian redshift error sigma0(1+z) to the halo redshifts, claiming that sigma0=0.005, 0.01, and 0.02 in different redshift bins reproduces the group-void VSF. The central assertion is that group-voids with ≥40% spectroscopic completeness faithfully reproduce halo-void statistics, and that even at 30% completeness the effect can be captured by a single effective redshift-error parameter.","tokens_in":15918,"tokens_out":3910,"duration_ms":39566,"significance":"If fully validated, the proposed group-void approach could significantly simplify void-based cosmological analyses by replacing galaxy-void emulators requiring galform models or HOD/abundance matching with halo-void emulators. The paper uses realistic CSST mock catalogs with a sophisticated slitless-spectroscopy simulation and a well-established group finder, and the ideal-case comparisons (z_spec group-voids vs halo-voids) are a useful sanity check of the pipeline. The density-profile analysis with BCG centers is also a potentially valuable practical improvement. However, the headline claim about 30% completeness rests on a post-hoc, unvalidated effective parameter, and the 40% claim is stronger than the plotted 2-sigma and density-profile evidence supports. With additional formal fitting and out-of-sample validation, the idea could be made compelling.","major_comments":[{"comment":"The effective redshift error sigma0 is chosen per redshift bin (0.005, 0.01, 0.02) by visually matching the halo-void VSF to the group-void z_CSST VSF. No goodness-of-fit statistic, uncertainty on sigma0, or independent test is provided. Because sigma0 is tuned to the exact data it is then claimed to reproduce, the statement in the Abstract and Conclusion that group-voids at 30% completeness can be 'reliably described' via a redshift error term is circular and unsupported. The text itself (Section 3.1) concedes 'We intend to investigate this strategy in future work,' which confirms the model is not yet established. This is load-bearing: the 30% tolerance is a key result. A formal likelihood fit with jackknife covariances and an out-of-sample prediction (e.g., predicting the density profile, or the VSF at an intermediate sigma0/completeness not used in the fit) is needed.","section":"Section 3.1 / Figure 3"},{"comment":"The conclusion that 'group-voids can reliably represent halo-voids in redshift bins with spectroscopic completeness of 40% or higher' is overstated. In the 0.2<z≤0.4 bin (the 40%-completeness bin), Figure 2 shows agreement only for voids with R_v ≳ 40 h−1Mpc and at the 2-sigma level, not a full 1-sigma match. Moreover, Figure 6 shows that the density profiles from group-voids with z_CSST deviate significantly at z>0.2 unless BCG centers are used; the improvement from BCG is demonstrated only for z≤0.6. Thus the blanket '40% or higher' claim is not uniformly supported by the plotted statistics. The conclusion should be qualified to specify radius range, center choice, and statistic.","section":"Section 4 / Section 3.2 / Figure 6"},{"comment":"The group-void versus halo-void comparison is partly self-referential by construction: the mock galaxies are placed into subhalos via SHAM, and the group finder assigns halo masses by abundance matching to an assumed HMF. The near-agreement in the ideal case is therefore an expected consistency check of the group finder and SHAM, not an independent empirical validation that group-voids generally trace halo-voids in real surveys. This does not invalidate the practical modeling approach, but it means the claims of 'faithfully reproduce' should be framed as a pipeline validation on mocks rather than an independent discovery, and the paper should discuss how group-finder uncertainties or mass errors in real data might break this agreement.","section":"Section 2 / Overall interpretation"}],"minor_comments":[{"comment":"The title and several headings contain 'V oids' with an apparent space; this is likely a LaTeX/compilation artifact and should be corrected.","section":"Title and text"},{"comment":"The caption does not list the exact sigma0 values for each panel in the printed text, relying on the reader to infer from the line styles; please state explicitly (e.g., '0.2<z≤0.4: sigma0=0.005') in the caption.","section":"Figure 3"},{"comment":"The relative deviation δ is mentioned in the text and figure subpanels but never defined by an equation. Please define δ explicitly (e.g., δ = (VSF_group − VSF_halo)/VSF_halo) and state how the jackknife errors are propagated into the subpanels.","section":"Section 3.1"},{"comment":"The number of sigma0 values tested is only three (0.005, 0.01, 0.02), yet the paper states these 'match' the group-void VSF. It would be clearer to describe these as discrete trial values rather than a fit, unless a fitting procedure is actually used.","section":"Section 3.1"}],"recommendation":"major_revision","confidential_remarks":"The paper is a pipeline-validation study in a series for the CSST survey. The central practical idea is promising, but the current draft presents a post-hoc effective-parameter match as a discovery. The authors should be encouraged to include a proper likelihood fit and, ideally, a prediction for an independent statistic or redshift bin. The overlap between the SHAM-based mock and the halo-comparison also needs explicit acknowledgment; otherwise, the paper may overstate the degree of empirical support. I do not see fundamental errors in the ideal-case analysis; the issue is the strength of the claims relative to the evidence."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThis is a useful calibration paper for CSST void cosmology, but the headline that 30% completeness is acceptable rests on a hand-tuned σ0 that isn't yet predictive. The cleanest result is the ideal-case validation: with full spectroscopic redshifts, group-void VSFs and stacked density profiles agree with halo-voids within 1σ across z=0–1. That is worth having, and the BCG-centered group-void density profiles at z≤0.6 also show a real improvement over luminosity-weighted centers.\n\nThe 40% completeness claim is plausible but loosely demonstrated; it's inferred from the first two redshift bins where the spec-z fraction is ~40% and only large voids are checked. The bigger issue is the σ0 model in Section 3.1. The authors assign halos a Gaussian redshift error σz=σ0(1+z) and choose σ0=0.005, 0.01, 0.01, 0.02 per bin to visually match the zCSST group-void VSF. There is no goodness-of-fit statistic, no uncertainty on σ0, and no validation on an independent redshift bin or on the density profile. The text even says the strategy will be investigated in future work, so the abstract's statement that group-voids at 30% completeness can be 'reliably described' via halo-voids overstates what is shown. What is shown is that some Gaussian smoothing can make the large-radius VSF overlap in the chosen bins — a descriptive fit, not a predictive model.\n\nA related concern: the comparison is partly circular, since the group catalog is built from the same halos via abundance matching. That doesn't invalidate the ideal-case agreement, but it means the σ0 values are fitting noise as much as signal. The paper also doesn't cite earlier group-catalog void work, which matters if such exists; the novelty here is mostly the CSST-specific calibration, not the group-void concept.\n\nThe paper is aimed at cosmology practitioners building emulators for void statistics, especially for CSST; for that audience it's a useful starting point despite the σ0 weakness. Still, the work is worth refereeing. The mock is realistic, the void finder is standard, and the ideal-case validation plus BCG-center test are solid. A revision that adds a proper fit for σ0 with error bars, validates it on a held-out redshift bin or a different completeness level, and tones down the abstract would make this a genuinely useful pipeline paper. I'd send it to a serious referee.","headline":"Useful CSST group-void calibration, but the 30%-completeness headline relies on a hand-tuned σ0 that is descriptive, not yet predictive.","tokens_in":16431,"tokens_out":7367,"would_cite":true,"duration_ms":58158,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Voids found in galaxy groups can stand in for dark-matter halo voids in cosmological analyses, even when only 40% of galaxies have precise redshifts.","keywords":["cosmic voids","galaxy groups","void size function","void density profile","large-scale structure","mock galaxy survey","spectroscopic completeness","cosmological emulators"],"falsifier":"Run the same group-finder and void-finder on mocks with spectroscopic completeness varying continuously between 20% and 50% while holding everything else fixed, and check whether the best-fit Gaussian redshift error on halos reproduces the group-void size function at every level and in every redshift bin. A break in that mapping—e.g., a sigma0 that cannot be found or that changes abruptly—would disprove the one-parameter equivalence claim.","tokens_in":15483,"feed_emoji":"🕳️","tokens_out":4675,"duration_ms":39487,"temperature":0.7,"pith_summary":"This paper establishes that cosmic voids identified from galaxy group catalogs (group-voids) reproduce the void size function and void density profiles obtained from dark-matter halos (halo-voids) in mock surveys. The match holds across redshift bins out to z=1 when at least 40% of galaxies have accurate spectroscopic redshifts. Even at about 30% completeness, the size function can be recovered if halo redshifts are given an effective Gaussian error whose amplitude is tuned per redshift bin. The practical payoff: cosmological analyses could model group-voids directly from halo catalogs in simulations, skipping the expensive galaxy-formation step that galaxy-based void studies require.","feed_headline":"Group voids match halo voids at 40% completeness","feed_subtitle":"Cosmic-void surveys could skip galaxy-formation modeling and use dark-matter halos directly in emulators.","key_machinery":"The argument rests on four pieces: (1) a halo-based group finder that assigns galaxies to dark matter halos, removing the finger-of-god distortion and stabilizing tracer positions; (2) a watershed-based void finder that identifies underdense zones without assuming a shape and returns volume-weighted centers and effective radii; (3) use of the brightest central galaxy (BCG) as the group center, which improves void-center accuracy when spectroscopic redshifts are sparse; and (4) a parameterization of incompleteness as an effective Gaussian redshift error applied to simulation halos, which the authors show reproduces the group-void size function in each redshift bin.","core_discovery":"The paper's central claim is that group-voids—underdensities traced by galaxy groups rather than individual galaxies—are statistically equivalent to halo-voids for the void size function and stacked void density profile, provided the group catalog has sufficient redshift completeness. In the ideal case of full spectroscopic redshifts, the two statistics agree within 1σ at z<0.8 and for large voids at higher redshift. In a conservative case with partial spectroscopic coverage, agreement holds where spectroscopic completeness is at least 40%; below that, a single free parameter describing a Gaussian redshift error applied to halos (with values 0.005, 0.01, and 0.02 in successive redshift bins)","pith_inferences":["The single-redshift-error mapping may also absorb other residual systematics (e.g., group-finder incompleteness, photometric outliers), so the inferred sigma0 values should not be interpreted literally as redshift errors; this can be tested by varying completeness in controlled mocks.","If the equivalence holds in real data, void statistics from group catalogs could be combined with galaxy-void statistics to cross-check systematics and tighten dark-energy constraints.","The approach might generalize to other tracer types, such as galaxy clusters or emission-line galaxies, wherever a group/halo correspondence can be established.","A direct test on a real survey with both photometric and spectroscopic redshifts (like the one the mock emulates) would confirm whether the 30% floor holds outside the mock."],"forward_implications":["Group-void size functions and density profiles can be predicted directly from dark-matter-only simulations, making emulator-based cosmological inference feasible without galaxy-formation prescriptions.","Spectroscopic incompleteness in a survey can be captured by one effective redshift-error parameter per redshift bin, greatly simplifying nuisance-parameter treatment.","The method extends the usable redshift range and void-size range for void cosmology with partial spectroscopic surveys, because group tracers are less sensitive to individual galaxy redshift errors.","Using the brightest central galaxy as the void center recovers accurate low-density regions out to z≈0.6 even when only a third of galaxies have precise redshifts.","The 40% completeness threshold provides a target for survey design: surveys that meet it can use group-voids without extra modeling."],"fun_headline_variants":["Group voids rival halo voids at 40% redshift completeness","Void stats from groups match halos at 40% completeness","CSST group-voids reproduce halo-void statistics at ≥40% completeness","Group-voids as faithful proxies for halo-voids past 40% completeness","Void astronomy: groups stand in for halos at 40% completeness"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The load-bearing premise is that the combined effect of incomplete spectroscopy and photometric redshift errors on the group-void size function is exactly equivalent to assigning halos a Gaussian redshift error with a single amplitude per redshift bin; if this equivalence fails at other completeness levels or redshifts, the 30% acceptance claim does not follow.","fun_headline_variants_meta":{"raw":{"variants":["Group voids rival halo voids at 40% redshift completeness","Void stats from groups match halos at 40% completeness","CSST group-voids reproduce halo-void statistics at ≥40% completeness","Group-voids as faithful proxies for halo-voids past 40% completeness","Void astronomy: groups stand in for halos at 40% completeness"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000906,"raw_usage":{"total_tokens":3789,"prompt_tokens":855,"completion_tokens":2934,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":599,"completion_tokens_details":{"reasoning_tokens":2846}},"tokens_in":599,"tokens_out":2934,"duration_ms":18338,"temperature":1.0,"reasoning_tokens":2846,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T12:32:55.607446+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the same group-finder and void-finder on mocks with spectroscopic completeness varying continuously between 20% and 50% while holding everything else fixed, and check whether the best-fit Gaussian redshift error on halos reproduces the group-void size function at every level and in every redshift bin. A break in that mapping—e.g., a sigma0 that cannot be found or that changes abruptly—would disprove the one-parameter equivalence claim.","supporting_citations":[],"review_version":2}