{"id":"dbd3fd37-a0a4-4a90-9e9c-0916e403576c","arxiv_id":"2505.09200","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Ball-bodies admit a near-unique duality that controls their curvature, diameter, and symmetrization behavior in Euclidean space.","lead":"This paper develops the geometry of ball-bodies, sets formed by intersecting translates of unit balls, and their duality operation. It proves new structural results about their boundaries, curvature, symmetrization, and connections to bodies of constant width and the Kneser-Poulsen conjecture.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Section 5.4's counterexample may be invalid: the stated vertical fiber lies outside the lens (the two ball intervals are disjoint), so the curvature computation does not apply to a Steiner symmetral.","rationale":"The reader's conditional verdict is appropriate. The main engine of the paper—Proposition 1.14, Theorem 1.19, Corollaries 1.22 and 1.23, the isometry theorem, and the curvature duality—appears logically sound; I found no circularity or missing steps there. The 3D counterexample in Section 5.4, however, has a more fundamental problem than the decimal approximations the reader flagged: the chosen vertical line misses the lens, so the computed h does not describe the Steiner symmetral. This does not disprove the claimed failure of S_n preservation (a corrected choice of parameters may well work), but it means the claim is not established as printed. Since the reader's verdict already conditions on this section, I leave it unchanged. The remaining contributions—Santalo-type inequalities, Caratheodory theorems, boundary structure, and curvature relations—are substantial and appear correct.","tokens_in":52311,"tokens_out":36473,"duration_ms":327693,"concrete_test":"Recompute Section 5.4 with exact arithmetic: evaluate the two vertical intervals at the stated coordinates (e.g., using the implicit z0=-0.4) and check whether max(lower1,lower2) <= min(upper1,upper2); if not, the fiber is empty and the example is void. Then, choose an interior point of the lens (e.g., reduce |z0| so that |2z0| <= s1+s2) and verify the claimed kappa_h<1 using exact rational or interval arithmetic on the unrounded expressions, with the denominator sqrt(1+|grad h|^2)(1+h_x^2). If the inequality fails or cannot be reproduced, the counterexample should be withdrawn or revised.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 5.4's numerical counterexample does not compute a Steiner symmetral fiber. With the implicit z0=-0.4 (the only value matching the printed values of fu, gu, -fd, -gd), the vertical intervals of the two unit balls at (x,y)=(0.4142,0.7268) are approximately [-0.934,0.134] for B(c0) and [0.209,0.591] for B(-c0). These intervals are disjoint, so the lens L has empty intersection with this vertical line; the true symmetral fiber has length 0. The quantity h=(fu+gd)/2 is then negative, not a half-length of any fiber. Moreover, the displayed numerical values do not match the curvature formula: read literally, the printed ψ(s,t)=sqrt(1+t^2+s^2(1+t^2)) gives roughly 4.47 for the averaged gradient, not 6.313, and the cited ψ-values for the individual gradients are not reproducible with the standard denominator sqrt(1+s^2+t^2)(1+s^2). The conclusion that Steiner symmetrization fails in dimension 3 is therefore not established as written; it may be repairable, but the example needs a corrected, exact check.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a systematic theory of ball-bodies, i.e., intersections of translates of the Euclidean unit ball, together with the associated c-duality. It proves a number of structural results: the identity K - K^c = B(0,1) (Proposition 1.14), linearity of the c-duality with respect to Minkowski averages (Theorem 1.19), Santaló-type inequalities for volume and mixed volumes, an order-isomorphism characterization of the class (Theorem 2.1), continuity of the c-duality with respect to Hausdorff distance, a curvature duality relating principal radii (Theorem 4.27), Carathéodory-type theorems for c-hulls, and results on Steiner and Minkowski symmetrization, including a claimed 3D counterexample to closure under Steiner symmetrization. The paper also discusses connections to the Kneser-Poulsen conjecture, bodies of constant width, and optimal transport.","tokens_in":52601,"tokens_out":12542,"duration_ms":104397,"significance":"If the main results hold, this paper is a valuable unified reference for ball-bodies. The identity K - K^c = B(0,1) is clean and powerful; it yields several inequalities simply and is used throughout. Theorem 1.19 (linearity of c-duality under Minkowski averages) is surprising, and Theorem 4.27 generalizes a known curvature duality for constant-width bodies. Many proofs are self-contained, with clear citations for classical tools. However, the main advertised new phenomenon in Section 5, namely that Steiner symmetrization preserves S_n only for n≤2, is currently not established because the numerical counterexample in dimension 3 is flawed.","major_comments":[{"comment":"The proposed counterexample does not compute a Steiner symmetral. For the parameters as used in the displayed values (z0=-0.4), the vertical intervals of the two unit balls at (x,y)=(0.4142,0.7268) are approximately [-0.934,0.134] for B(c0) and [0.209,0.591] for B(-c0). These intervals are disjoint, so the lens L has empty intersection with this vertical line and the Steiner symmetral S_{e3}(L) has an empty fiber there. The interval [(x,y,-gd),(x,y,fu)] used in the text has -gd≈0.209 > fu≈0.134, so it is not a fiber of L. Therefore the curvature computation is not applied to a valid fiber, and the claim that Su(L) leaves S_3 is not established. Because this is the only evidence for the claims that Steiner symmetrization preserves S_n only for n≤2 and that Vol(L_t) can be non-convex for n≥3 (Remark 5.7), these claims must either be proved with a correct example or removed.","section":"Section 5.4"},{"comment":"The numerical verification is internally inconsistent. With the printed definition ψ(s,t)=sqrt(1+t^2+s^2(1+t^2)), ψ(0.9996,2.9972) evaluates to about 4.47, not 6.313. The quoted individual values (4.251 and 7.658) also do not follow from the standard curvature denominator sqrt(1+||∇φ||^2)(1+φ_x^2); with that denominator, the individual values are approximately 5.03 and 7.80. The authors should provide an exact algebraic inequality or an interval-arithmetic certificate for the curvature comparison, and should first ensure that the fiber being considered is actually a fiber of L.","section":"Section 5.4"}],"minor_comments":[{"comment":"The parameter line 'z−0 = 0.4' is ambiguous and inconsistent with the functions fu, gu, -fd, -gd displayed directly below it, which correspond to z0=-0.4. Please state the sign of z0 unambiguously.","section":"Section 5.4"},{"comment":"There are several typographical errors, including 'geneal settings' in Section 1, 'Miknowski average' in Corollary 1.21, 'saisfy' in Lemma 6.7, and 'Deonte' in the proof of Lemma 5.4. A careful proofread is needed.","section":"Throughout"},{"comment":"The full isometry classification is attributed to unpublished preprints [5,6]; since these are not available to the reader, the remark should either be stated as conditional or accompanied by a proof outline.","section":"Remark 2.5"},{"comment":"The example is described verbally and with a figure but without enough coordinate detail to verify the claimed smoothness; a short explicit formula for the body would improve reproducibility.","section":"Example 4.14"}],"recommendation":"major_revision","confidential_remarks":"The manuscript contains many correct and useful results, but the Section 5.4 counterexample is central to the advertised Steiner symmetrization claim. If the authors can replace it with a correct exact example (or with a proof that the fiber issue can be remedied), the paper would be acceptable; otherwise the revision may be substantial. The editor may also wish to ask the authors to clarify the status of preprints [5,6] cited in Remark 2.5."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is genuinely useful: it collects and sharpens the theory of ball-bodies, and most of the main structural claims hold up on reading. The identity K−Kc = B(0,1) is proved cleanly and drives the Santaló-type inequalities, the duality under Minkowski averaging, and the curvature duality. The Carathéodory-type theorems and the characterization of order-preserving and order-reversing maps on S_n are real contributions, and the proofs are mostly self-contained and cite the classical tools fairly. I see no circularity burden; the old results are clearly separated from the new ones, and the self-citations for the isometry classification are not load-bearing. The in-/out-radius comparison and the 1-lens volume convexity are nice additions.\n\nThe soft spot is Section 5.4. The stress-test is right: at the printed parameter values, the vertical interval of B(c0,1) is roughly [−0.934, 0.134] and the interval of B(−c0,1) is roughly [0.209, 0.591]. These are disjoint, so the lens has empty intersection with that vertical line. The true Steiner symmetral fiber has length 0, and h = (fu+gd)/2 is negative, not a half-fiber. The displayed curvature numbers also do not reproduce from the stated ψ formula: for the averaged gradient I get about 4.47, not 6.313, and the individual ψ-values are not consistent with the standard graph-curvature denominator. As written, the claim that Steiner symmetrization fails in dimension 3 is not established. It may well be repairable—pick a different fiber or do an exact algebraic check—but the paper should not ship this counterexample in its current form. Note also that the notation “z−0 = 0.4” is misleading; the printed numbers only work with z0 = −0.4, and even then the fiber is empty. This is a secondary claim, not the engine of the main theory. The planar Steiner result is separate and appears fine.\n\nCitation behavior is honest: classical theorems are credited, and the new results are clearly marked. The paper deserves a serious referee. The referee should ask for a corrected, exact verification of the 3D counterexample, or for the claim to be softened, before publication. The rest of the paper is in good shape.","headline":"A rich and mostly sound study of ball-bodies whose central structural theorems hold up, but the Section 5.4 counterexample to Steiner symmetrization in 3D is numerically broken as written and needs an exact fix before the claim is cited.","tokens_in":53146,"tokens_out":5453,"would_cite":true,"duration_ms":49078,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["52A20","52A40","52A39"],"pacs":[],"model":"deepseek-v4-flash","headline":"For every non-degenerate ball-body $K$, the identity $K-K^c=B(0,1)$ holds, and the paper shows that this one relation drives the class's inequalities, symmetries, and boundary structure.","keywords":["ball-bodies","c-duality","summands of Euclidean ball","Santaló inequality","constant width bodies","Steiner symmetrization","Kneser-Poulsen conjecture","spindle convexity"],"falsifier":"Using exact algebra or interval arithmetic at the stated fiber point $(x,y)=(0.4142,0.7268)$ of the lens $B(c_0,1)\\cap B(-c_0,1)$ with $c_0=(-0.2807,0.2457,0.4)$, verify whether the sectional curvature of the Steiner symmetral in direction $e_1$ is strictly below 1; if the two displayed values 6.313 and 5.9545 do not preserve their strict order under rigorous computation, the counterexample collapses.","tokens_in":52137,"feed_emoji":"📐","tokens_out":8188,"duration_ms":78957,"temperature":0.7,"pith_summary":"Ball-bodies—sets that can be written as intersections of translates of the Euclidean unit ball—are shown to form a self-contained analogue of convex bodies, with a duality that behaves like a geometric complement. The paper's central result is the identity $K-K^c=B(0,1)$ for every non-degenerate ball-body, which pins down the c-dual support function as $h_{K^c}(u)=1-h_K(-u)$. From this single relation the authors derive volume and mixed-volume inequalities of Santaló type, closure of the class under Minkowski averaging, an isometry theorem for the duality, a curvature duality at paired boundary points, and a characterization of bodies of constant width as fixed points. Along the way they prove structural analogues of Carathéodory and Krein–Milman theorems for the c-hull and show that Steiner symmetrization preserves the class in the plane but not in higher dimensions. If correct, this gives a unified setting for several open problems in convex geometry, including isoperimetric questions and the Kneser–Poulsen conjecture.","feed_headline":"One identity governs ball-bodies: K − K^c = B(0,1)","feed_subtitle":"Intersections of unit balls carry a duality that yields Santaló-type inequalities, isometries, and curvature relations.","key_machinery":"The c-dual of a set is $K^c=\\bigcap_{x\\in K}B(x,1)$, and on ball-bodies it is an order-reversing involution. The engine of the paper is the support-function identity $h_{K^c}(u)=1-h_K(-u)$, which is the analytic form of $K-K^c=B(0,1)$. This identity transfers ball-body problems into additive and mixed-volume form: it turns the duality into a linear operation under Minkowski averages, gives the Santaló-type inequalities via Brunn–Minkowski and Urysohn, and yields the isometry property by mapping support functions to their complements. It also connects boundary normals: for $x\\in\\partial K$ and $u\\in N_K(x)$, the point $x-u$ lies in $\\partial K^c$ with normal $-u$, which is the mechanism behind the curvature relation $r_i+s_{n-i}=1$.","core_discovery":"The core discovery is the duality identity of Proposition 1.14: for any $K\\in \\mathcal{S}_n\\setminus\\{\\emptyset,\\mathbb{R}^n\\}$, $K-K^c=B(0,1)$; equivalently, $h_{K^c}(u)=1-h_K(-u)$ for every unit vector $u$. The paper shows that this identity is equivalent to $K$ being a summand of the Euclidean ball, and that it forces $K+K^c$ to have constant width 2. From it the authors obtain a Santaló-type inequality $\\operatorname{Vol}(K)^{1/n}+\\operatorname{Vol}(K^c)^{1/n}\\le \\operatorname{Vol}(B(0,1))^{1/n}$, with equality only for balls, and a mixed-volume version. It also makes the c-duality linear under Minkowski averaging, an isometry of the class under Hausdorff distance, and the only order-reversing involution up to rigid motions. At smooth dual boundary points, the principal radii of curvature of $K$ and $K^c$ add to 1, a generalization of a classical constant-width fact.","pith_inferences":["The support-function identity suggests a full c-Mahler programme in all dimensions: minimize $\\operatorname{Vol}(K)\\operatorname{Vol}(K^c)$ over ball-bodies of fixed volume; the plane is solved here by lenses, and the natural conjectural extremizers in higher dimensions are 1-lenses and $(n-1)$-lenses.","Because c-duality is both an isometry and linear under averaging, ball-bodies form a natural testbed for optimal-transport costs that forbid pairs at distance at most 1; the necessary condition $\\mu(A)+\\nu(A^c)\\le 1$ from the transport literature becomes a genuinely geometric condition on this class.","The basin description $\\mathrm{Basin}(K)=\\{T\\in \\mathcal{S}_n: h_K-h_T \\text{ is even}\\}$ gives a recipe for constructing many constant-width bodies with prescribed odd support part; varying the even part may lead to new small-volume examples relevant to the Blaschke–Lebesgue problem."],"forward_implications":["Every ball-body outside the degenerate cases satisfies $K-K^c=B(0,1)$, so $K+K^c$ is a body of constant width 2; bodies of constant width 1 are exactly the fixed points $K=K^c$.","For any two ball-bodies $K,T$, the Minkowski average $(1-\\lambda)K+\\lambda T$ is again a ball-body and its c-dual is the same average of the duals; consequently Minkowski symmetrization and orthogonal projections preserve the class.","The Santaló-type inequality $\\operatorname{Vol}(K)^{1/n}+\\operatorname{Vol}(K^c)^{1/n}\\le \\operatorname{Vol}(B(0,1))^{1/n}$ holds, with equality only for balls; the same pattern holds for mixed volumes, yielding bounds for constant-width bodies.","The boundary structure of ball-bodies mirrors convexity: each boundary point lies in the c-hull of at most $n$ c-extremal points, each interior point in at most $n+1$, and $K$ is the c-hull of its c-extremal points.","Steiner symmetrization maps $\\mathcal{S}_2$ to itself but not $\\mathcal{S}_3$, so the class is not closed under every classical symmetrization in higher dimensions."],"supporting_citations":[{"why":"Proves that a convex body slides freely inside the unit ball exactly when it is a summand, identifying ball-bodies with ball-summands via Theorem 1.11.","marker":"[67]"},{"why":"Provides the order-reversing quasi-involution formalism that makes the c-dual an involution on the image class.","marker":"[14]"},{"why":"Supplies lens comparison theorems for volume versus in-radius and out-radius, which the paper extends to Santaló-type settings.","marker":"[21]"},{"why":"Proves the planar isoperimetric conjecture for ball-bodies, used here to derive the planar c-Mahler inequality.","marker":"[28]"},{"why":"Proves the reverse inradius inequality for ball-bodies in dimension three, a recent extremal result used as a benchmark.","marker":"[36]"},{"why":"Gives Schramm's volume bound for constant-width bodies and the illumination result that the paper discusses and partially reproves.","marker":"[69]"},{"why":"Beckman–Quarles theorem, used to show that any order-preserving bijection of the class is induced by a rigid motion.","marker":"[16]"}],"fun_headline_variants":["Ball-bodies obey K − K^c = B(0,1)","Ball-bodies: K + K^c has constant width 2","Santalo-type inequality for ball-bodies","Duality on ball-bodies: h_{K^c}(u)=1-h_K(-u)","For ball-bodies, curvature radii add to 1 at dual points"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper's negative result for Steiner symmetrization in $n\\ge 3$ depends on one computed lens example whose decisive inequality is checked by approximate decimals; if that comparison is not exact, the claim that $\\mathcal{S}_n$ is preserved only for $n\\le 2$ loses its proof.","fun_headline_variants_meta":{"raw":{"variants":["Ball-bodies obey K − K^c = B(0,1)","Ball-bodies: K + K^c has constant width 2","Santalo-type inequality for ball-bodies","Duality on ball-bodies: h_{K^c}(u)=1-h_K(-u)","For ball-bodies, curvature radii add to 1 at dual points"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001205,"raw_usage":{"total_tokens":4950,"prompt_tokens":919,"completion_tokens":4031,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":535,"completion_tokens_details":{"reasoning_tokens":3934}},"tokens_in":535,"tokens_out":4031,"duration_ms":29489,"temperature":1.0,"reasoning_tokens":3934,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T21:37:27.510329+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Using exact algebra or interval arithmetic at the stated fiber point $(x,y)=(0.4142,0.7268)$ of the lens $B(c_0,1)\\cap B(-c_0,1)$ with $c_0=(-0.2807,0.2457,0.4)$, verify whether the sectional curvature of the Steiner symmetral in direction $e_1$ is strictly below 1; if the two displayed values 6.313 and 5.9545 do not preserve their strict order under rigorous computation, the counterexample collapses.","supporting_citations":[{"cited_title":"Schneider","cited_arxiv_id":null,"evidence_quote":"Proves that a convex body slides freely inside the unit ball exactly when it is a summand, identifying ball-bodies with ball-summands via Theorem 1.11."},{"cited_title":"Artstein-Avidan, S","cited_arxiv_id":null,"evidence_quote":"Provides the order-reversing quasi-involution formalism that makes the c-dual an involution on the image class."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies lens comparison theorems for volume versus in-radius and out-radius, which the paper extends to Santaló-type settings."},{"cited_title":"Borisenko and K","cited_arxiv_id":null,"evidence_quote":"Proves the planar isoperimetric conjecture for ball-bodies, used here to derive the planar c-Mahler inequality."},{"cited_title":"Drach and K","cited_arxiv_id":null,"evidence_quote":"Proves the reverse inradius inequality for ball-bodies in dimension three, a recent extremal result used as a benchmark."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives Schramm's volume bound for constant-width bodies and the illumination result that the paper discusses and partially reproves."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Beckman–Quarles theorem, used to show that any order-preserving bijection of the class is induced by a rigid motion."}],"review_version":1}