{"id":"fcaf5386-725f-42e9-ae54-39123242328f","arxiv_id":"2412.13883","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Using the center of mass as the reference center reduces the anisotropic weak lensing mass bias from miscentering to below one percent in simulated galaxy clusters.","lead":"This paper tests whether galaxy cluster masses inferred from weak lensing are less biased when the cluster center is defined as the center of mass instead of the bottom of the gravitational potential. Using hydrodynamical simulations, the authors find the non-isotropic miscentering bias drops from about 6 percent to within 1 percent.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Distance-resampling in Sec. 2.3 confounds tau; sub-percent CoM claim needs rotation-only test.","rationale":"I read the paper as a concise simulation study with a clear goal: to measure the anisotropic component of weak-lensing mass bias and argue that using the center of mass as reference reduces it below 1%. The strongest claim is supported by Table 1, but the estimator tau depends on how the isotropic baseline is constructed. The reader identified the randomization procedure in Sec. 2.3 as the weakest assumption, focusing on directional correlations that the scalar-distance randomization erases. My analysis agrees that the randomization is the load-bearing step, but I identify a more fundamental and technical flaw: redrawing distances from the pooled distribution breaks the pairing between each halo's offset distance and its own properties. This can bias tau even in the absence of any directional anisotropy, because the mass-bias response to miscentering depends on halo concentration, mass, and dynamical state. The proposed rotation-only test is a clean, computationally cheap way to settle whether the reported sub-percent result is real or an artifact of the resampling scheme. I therefore maintain the reader's CONDITIONAL verdict, adding the specific condition that the rotation-only test be performed. The paper is otherwise internally consistent, and I see no reason to escalate to rejection without this test. My agreement is partial because I emphasize the distance-resampling confound rather than erased directional correlations, though both concern the same Section 2.3 construction.","tokens_in":11382,"tokens_out":13421,"duration_ms":125629,"concrete_test":"Run a rotation-only isotropization: for each of the 825 projected targets, keep the measured absolute miscentering |Δ| fixed (e.g., |SZE−G| or |SZE−CoM|) and assign a uniformly random angle to generate a rotated center. Repeat the NFW mass fitting and compute tau comparing these rotated centers to the actual SZE centers, exactly as in Sec. 3. Under per-halo isotropy, this rotation-only tau is zero in expectation. If it reproduces the reported ~6% (G) and ~0 (CoM), the conclusion is robust. If it shifts by more than about 1%, the distance-resampling artifact is significant and the central claim is not supported. This requires no new simulations, only rerunning the existing pipeline with modified centers.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The pivotal methodological step is the construction of the isotropic baseline in Sec. 2.3. For each halo, the authors draw a new absolute miscentering (with replacement) from the global empirical distance distribution and couple it with a uniform random angle. This does more than randomize directions: it also breaks the association between the offset distance and the individual halo's properties (mass, concentration, dynamical state). The weak-lensing mass bias b is a nonlinear function of offset distance and the halo's density profile. Even if every halo's miscentering were perfectly isotropic (uniform angle, independent of distance), the mean of b over the actual sample would not equal the mean over the randomized sample whenever the halo-specific distance distribution or the mass-bias response correlates with the same halo properties. The estimator tau, defined in Sec. 2.4 as the mean relative overcorrection relative to this randomized baseline, therefore does not isolate the anisotropic component; it contains a resampling artifact. This directly affects the central claim: the reported tau ≈ 6% for the gravitational center and tau ≈ 0 for the center of mass (Table 1 rows 4 and 6/8) could be biased by this artifact. The reader's concern about erased directional correlations (e.g., alignment with merger axes) is related, but the distance-resampling confound is more fundamental because it does not require any non-uniform angle distribution to produce a spurious signal. Without a test that preserves per-halo distances, the statement that the non-isotropic component is reduced to within one percent with the center of mass is not yet established.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper investigates whether the choice of the reference center used to define a cluster's 'true' mass affects the directional (anisotropic) part of the weak-lensing mass bias caused by miscentered observational centers. Using a z=0.67 snapshot of the Magneticum box2b-hr simulation, the authors select 275 massive halos and project them along three axes, identifying the gravitational center (most bound particle), the iterative center of mass, and the SZE peak in noiseless, 1-arcmin-smoothed Compton-y maps. They fit NFW weak-lensing masses centered on the actual SZE peaks and on randomized miscentering positions drawn from the empirical absolute offset distributions with random angles. The difference between the two is summarized by the overcorrection tau. They find tau of about 6% when the gravitational center is the reference and tau consistent with zero when the center of mass is the reference, concluding that using the center of mass removes the non-isotropic component of the mass bias to within about one percent.","tokens_in":11613,"tokens_out":18652,"duration_ms":177581,"significance":"If the conclusion survives the methodological concerns below, this is a valuable result for cluster cosmology: miscentering is a leading systematic in lensing mass calibration, and a simulation-side change of reference center would be inexpensive to implement. The paper is clearly written and has useful built-in robustness tests, including a constant concentration, no mass recalibration, and 0.5-arcmin SZE smoothing. The central quantitative claim, however, depends on the definition of the isotropic baseline; as argued in the major comments, the current randomization procedure conflates directional anisotropy with resampling effects, so the sub-percent claim is not yet established.","major_comments":[{"comment":"The randomization does not isolate the directional (anisotropic) component. Drawing a new absolute miscentering with replacement from the global empirical distance distribution and coupling it with a uniform angle changes more than the direction: it breaks the association between the offset magnitude and the individual halo's properties (mass, concentration, dynamical state, orientation). The weak-lensing mass bias b is a nonlinear function of the offset distance and of the halo's density profile, so the mean of b over the randomized sample is not the same as the mean over the same halos with their own offsets rotated by uniform angles, even if every halo's true miscentering were perfectly isotropic. Hence tau can be nonzero purely because of the resampling step. The correct null for the stated question is a rotation-only randomization that keeps each target's absolute offset fixed and randomizes only the angle. Without this control, the reported tau of about 6% for the gravitational-center reference and tau consistent with zero for the center-of-mass reference do not measure the non-isotropic component of the mass bias, and the sub-percent conclusion in the abstract is not established. Please rerun the analysis with rotation-only randomization, and if desired report the resampling-based baseline separately.","section":"Section 2.3 and 2.4, Table 1 rows 4, 6, 8"},{"comment":"The 825 targets are not independent. Each of the 275 halos is projected along three mutually orthogonal lines of sight, so the three shear images of the same halo share the same three-dimensional density field and are correlated. The quoted 1-sigma uncertainties on the fitted distribution parameters and on tau appear to treat all 825 targets as independent draws; if so, they are underestimated. This matters for the central statement that the residual is consistent with zero at the one-percent level. Please provide a halo-level bootstrap (resampling the 275 halos, not the 825 projections) or otherwise quantify the effective number of independent targets.","section":"Section 2.2 and Table 1"},{"comment":"The analysis uses a single snapshot at z=0.67 from one simulation box. The abstract's claim that the non-isotropic component 'can be reduced to within one percent' is therefore demonstrated only for this mass range and redshift. Since the proposed remedy is intended to inform survey calibration across a range of redshifts, either add at least one additional snapshot (or a redshift-binned sample) or explicitly restrict the abstract and conclusions to the tested regime.","section":"Section 2.2 and Section 5"}],"minor_comments":[{"comment":"The first sentence contains a typo ('UsingssnapshotoftheMagneticumsimulations') and gives the redshift as z=0.7 while the body uses z=0.67; please fix and harmonize.","section":"Section 5"},{"comment":"The table header appears to place the symbols mu_b, sigma_b and mu_log b, sigma_log b under the wrong distribution blocks. For example, row 0 lists -0.024 and 0.167 before 0.991 and 0.179, which is inconsistent with the text's definitions of the normal and log-normal fit parameters. Please use separate column groups labeled 'log-normal fit' and 'normal fit' and define the parameters explicitly in the caption.","section":"Table 1 and Section 2.4"},{"comment":"The schematic would be clearer if the caption stated explicitly that the randomized SZE centers are generated from the empirical absolute-distance distributions with independent uniform angles, not from the full two-dimensional offset distribution.","section":"Section 2.3 and Figure 3"},{"comment":"The statement that an isotropic broadening of the SZE miscentering distribution 'would not be expected to cause anisotropic effects' is plausible but is asserted rather than demonstrated; one sentence of justification or a reference would help.","section":"Section 4"}],"recommendation":"major_revision","confidential_remarks":"The randomization issue is the crux of the paper. If the authors can show with a rotation-only randomization that tau remains consistent with zero for the center-of-mass reference, I would be happy to accept a revised version. If not, the central claim as stated does not follow from the presented test."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this short simulation study has a genuinely new and practical suggestion — calibrate weak lensing mass bias with the center of mass as the reference instead of the bottom of the gravitational potential. The reported improvement (about 6% overcorrection with the gravitational center, consistent with zero for the CoM) is plausible and the paper is honestly written. But the randomization that defines the isotropic baseline in Sec. 2.3 mixes two effects, and the sub-percent claim is not yet established without an extra test.\n\nWhat is new and good: the CoM idea is new in this context, the paper is clear, and the authors are upfront about limitations — one snapshot from one Magneticum box, noiseless SZE images, no X-ray centroid check. The robustness checks (constant concentration, no mass recalibration, 0.5 arcmin smoothing) all support the qualitative result, and the consistency with Sommer+24 is reassuring. The paper does not oversell; the X-ray extension is explicitly labelled a speculation.\n\nThe soft spot is load-bearing. In Sec. 2.3, the isotropic baseline is built by resampling absolute offsets from the global distance distribution and pairing each with a uniform random angle. That does two things at once: it randomizes directions, but it also breaks the link between each halo's offset distance and its own mass, concentration, and dynamical state. The weak lensing mass bias is a nonlinear function of both the offset and the halo's density profile, so the mean bias over the randomized sample does not have to equal the mean over the actual sample even if every offset angle were perfectly uniform. Tau therefore does not purely measure anisotropy; it also includes a resampling artifact that could be part of the 6% versus 0% gap. The authors' own limitations paragraph dismisses the noiseless-SZE issue quickly, but this confound is not mentioned there.\n\nThe fix is cheap: rerun the comparison rotating each halo's actual offset vector by a uniform random angle, preserving per-halo distances. That test is absent, and without it the claim that the non-isotropic component drops below one percent with the CoM is not fully supported. The rest of the evidence base is also narrow, and no code or data are released — \"reasonable request\" only — so independent checks are harder than they need to be.\n\nProportion: this is a solid, useful paper with one untested methodological wrinkle under its headline number; it is not a takedown. Cluster cosmologists calibrating mass-observable relations with SZE or X-ray centers should read it, and I would cite it with a cautionary note. I would bring it to a reading group; the confound is instructive.\n\nRecommendation: send it to a serious referee. The idea is new and the execution is mostly sound, but the editor should ask for the rotation-only test (and ideally another simulation box) before the sub-percent result enters the calibration literature.","headline":"Practical, new idea — use the center of mass as the reference for weak lensing mass calibration — but the sub-percent claim rests on a randomization test that confounds distance resampling with direction, so it needs a rotation-only check before it enters the calibration literature.","tokens_in":12192,"tokens_out":6710,"would_cite":true,"duration_ms":58602,"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":"Switching galaxy-cluster mass reference from the potential minimum to the center of mass reduces the directional weak lensing mass bias to about one percent.","keywords":["weak gravitational lensing","galaxy clusters","miscentering","mass bias","Sunyaev-Zeldovich effect","center of mass","hydrodynamical simulations","cluster mass calibration"],"falsifier":"A direct check is to measure the distribution of angles between the vector $\\Theta_{\\mathrm{SZE}}-\\Theta_{\\mathrm{CoM}}$ (or $\\Theta_{\\mathrm{SZE}}-\\Theta_{\\mathrm{G}}$) and the projected elongation axis of the cluster or the surrounding large-scale structure. If that angular distribution is not uniform, the isotropic-randomization null model erases real directional information and the quoted sub-percent $\\tau$ is not the full anisotropic bias. A sharper version: repeat the mass fits with miscentering offsets artificially aligned along the cluster major axis and compare $\\tau$ with the minor-axis-aligned case; any difference means the center-of-mass reference does not remove the directional dependence.","tokens_in":11152,"feed_emoji":"🔭","tokens_out":11686,"duration_ms":92048,"temperature":0.7,"pith_summary":"Weak lensing estimates of galaxy cluster masses are biased when the adopted cluster center is slightly off, and when centers come from X-ray or Sunyaev-Zeldovich observations the bias carries a directional component that standard isotropic miscentering corrections miss. The paper uses hydrodynamical cosmological simulations to test a simple remedy: redefine the reference center of a cluster as its center of mass instead of the bottom of the gravitational potential. With the center-of-mass reference, the mass overcorrection produced by applying an isotropic miscentering correction drops from about 6 percent to a value consistent with zero at the one-percent level. This matters because miscentering is currently a dominant systematic in cluster mass calibration, and the proposed change keeps the usual isotropic correction machinery intact while removing most of the directional bias.","feed_headline":"Center of mass reference cuts weak-lensing mass bias to ~1 percent","feed_subtitle":"Switching the reference center from the potential minimum to the center of mass removes a ~6 percent overcorrection.","key_machinery":"The argument is carried by a randomization scheme that separates directional from scalar miscentering. For each halo, the absolute miscentering distance $|\\Theta_{\\mathrm{SZE}}-\\Theta_{\\mathrm{G}}|$ (or the corresponding distance relative to $\\Theta_{\\mathrm{CoM}}$) is drawn and paired with a uniformly random angle $\\varphi\\in[0,2\\pi)$, creating an isotropic comparison distribution that has the same radial miscentering statistics but no preferred direction. The overcorrection $\\tau = \\langle (M_{\\mathrm{biased}}-M_{\\mathrm{unbiased}})/M_{\\mathrm{unbiased}}\\rangle$ then measures the mass bias introduced purely by the directional structure of the real miscentering. The second ingredient is the iterative center-of-mass definition, obtained by mass-weighting particles inside $r_{500}$ and recomputing $M_{500}$ and $r_{500}$ at each step until convergence, which reduces the absolute SZE miscentering by about half. The fact that $\\tau$ is consistent with zero when the reference center is $\\Theta_{\\mathrm{CoM}}$ is what carries the conclusion.","core_discovery":"The paper claims that the non-isotropic component of the weak lensing mass bias can be reduced to within one percent of the cluster mass by choosing the center of mass $\\Theta_{\\mathrm{CoM}}$ as the reference center rather than the gravitational potential minimum $\\Theta_{\\mathrm{G}}$. For the 275 most massive halos of a hydrodynamical cosmological simulation snapshot at $z=0.67$, projected along three orthogonal axes, the peak of a one-arcminute-convolved Sunyaev-Zeldovich (SZE) image serves as the observed center proxy. Comparing mass fits centered on the actual SZE position with fits centered on randomized isotropic miscentering, the overcorrection $\\tau$ is about 6 percent when $\\Theta_{\\mathrm{G}}$ is the reference, but becomes $\\tau=-0.2\\pm0.8$ percent (log-normal model) or $\\tau=-0.3\\pm0.9$ percent (Gaussian model) when $\\Theta_{\\mathrm{CoM}}$ is the reference. Because $\\tau$ is consistent with zero in the center-of-mass frame, the paper concludes that standard isotropic miscentering distributions remain adequate provided the halo center is defined by the center of mass, and that the halo mass function need not be redefined.","pith_inferences":["The paper does not recompute mass-observable relations, but a practical consequence is that survey calibrations built on simulations with gravitational-center references carry a roughly 6 percent directional correction that would disappear if the calibrations were re-derived with center-of-mass references; this could shift inferred cluster counts and cosmological parameters.","A stronger test of the null model would preserve the angle between the miscentering vector and the projected shear field when randomizing; if the directional bias is produced by correlated structure rather than by the scalar distance distribution, the sub-percent conclusion may be specific to the angle-averaged comparison.","The same center-of-mass prescription likely applies to optical and X-ray centroid proxies: X-ray centroids trace gas density and earlier work found X-ray and SZE overcorrections similar, so the bias reduction should transfer, and the prediction is testable with the same simulation pipeline.","A falsifiable prediction follows for real surveys: using SZE-selected clusters, masses calibrated with center-of-mass references should show no dependence of the weak lensing mass bias on the position angle of the SZE centroid relative to cluster elongation, whereas gravitational-center calibrations should show such a dependence."],"forward_implications":["Adopting the center of mass as the reference center reduces the residual anisotropic miscentering systematic in weak lensing mass calibration to about one percent, comparable to the statistical goals of current and planned cluster surveys.","Standard isotropic empirical miscentering distributions can still be used for simulation-based corrections; only the reference center used to measure them needs to change.","The halo mass function does not need to be recomputed for the new center definition, since $M_{500,\\mathrm{CoM}}$ differs from $M_{500,\\mathrm{G}}$ by only about 5 percent on average.","Because SZE peaks and X-ray centroids showed similar overcorrections in earlier work, and X-ray emission traces gas density, the same center-of-mass reference is expected to reduce biases for X-ray-determined centers, though this is not demonstrated here.","The result is insensitive to the SZE smoothing scale between 0.5 and 1 arcminute and to the concentration-mass relation, so the sub-percent conclusion does not depend on these modeling choices."],"supporting_citations":[{"why":"Establishes that SZE- and X-ray-derived centers produce several-percent anisotropic weak lensing mass biases, and justifies using noiseless SZE images via its noise-broadening test.","marker":"S+24"},{"why":"The Magneticum Pathfinder simulation suite from which the cluster sample is drawn.","marker":"Hirschmann et al. 2014"},{"why":"Describes the Magneticum simulations including the galaxy formation physics used for the redshift slice and halos.","marker":"Dolag et al. 2016"},{"why":"Provides the NFW density profile used to model clusters and fit weak lensing masses.","marker":"Navarro et al. 1997"},{"why":"Supplies the concentration-mass relation used in the mass fitting.","marker":"Diemer & Kravtsov 2015"},{"why":"Supplies the corrected concentration-mass relation parameters adopted in the fits.","marker":"Diemer & Joyce 2019"},{"why":"Provides the analytic surface density and shear expressions used to predict reduced tangential shear.","marker":"Bartelmann 1996"},{"why":"Provides the ClusterFind and SIMCUT tools used to identify halos and to recompute masses and radii while iterating the center of mass.","marker":"Ragagnin et al. 2017"}],"fun_headline_variants":["Center of mass reference reduces weak-lensing bias to ~1%","Swap reference center to fix weak-lensing mass bias","Miscentering bias in weak lensing tamed by center of mass","Weak-lensing mass bias drops to 1% with different center","Center of mass cuts directional miscentering bias to 1%"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that pairing each observed miscentering distance with a uniformly random angle creates a fair isotropic baseline; if real SZE miscentering directions are correlated with the surrounding structure or the shear field, the randomization destroys that correlation and the measured $\\tau$ underestimates the true directional bias.","fun_headline_variants_meta":{"raw":{"variants":["Center of mass reference reduces weak-lensing bias to ~1%","Swap reference center to fix weak-lensing mass bias","Miscentering bias in weak lensing tamed by center of mass","Weak-lensing mass bias drops to 1% with different center","Center of mass cuts directional miscentering bias to 1%"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000625,"raw_usage":{"total_tokens":2896,"prompt_tokens":952,"completion_tokens":1944,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":568,"completion_tokens_details":{"reasoning_tokens":1855}},"tokens_in":568,"tokens_out":1944,"duration_ms":12945,"temperature":1.0,"reasoning_tokens":1855,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T12:41:08.118155+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct check is to measure the distribution of angles between the vector $\\Theta_{\\mathrm{SZE}}-\\Theta_{\\mathrm{CoM}}$ (or $\\Theta_{\\mathrm{SZE}}-\\Theta_{\\mathrm{G}}$) and the projected elongation axis of the cluster or the surrounding large-scale structure. If that angular distribution is not uniform, the isotropic-randomization null model erases real directional information and the quoted sub-percent $\\tau$ is not the full anisotropic bias. A sharper version: repeat the mass fits with miscentering offsets artificially aligned along the cluster major axis and compare $\\tau$ with the minor-axis-aligned case; any difference means the center-of-mass reference does not remove the directional dependence.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the analytic surface density and shear expressions used to predict reduced tangential shear."}],"review_version":1}