{"id":"8029ea71-cfc4-44af-be09-7eec7863f184","arxiv_id":"2412.05308","paper_version":2,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Weighted sums of the mixed volumes V(K[j],-K[n-j]) are bounded by Vol(K), sharp for simplices, giving an averaged Godbersen inequality and extending Rogers-Shephard.","lead":"This paper proves new upper bounds on weighted sums of the mixed volumes of a convex body and its reflection, giving partial progress toward a 1938 conjecture of Godbersen. One consequence is that Godbersen's bound holds 'on average' for every convex body, and a related inequality extends the classical Rogers-Shephard bound.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Equality statement in Theorems 2 and 5 fails at λ=0 and λ=1; the LHS equals Vol(K) for every K at those endpoints, so non-simplex bodies also attain equality. The main inequalities survive if λ is restricted to (0,1).","rationale":"The core inequality in Theorem 2 is proved correctly: the volume computation for C, the application of the Rogers–Shephard section-projection inequality, and the expansion into mixed volumes are coherent once one tracks the λ↔1−λ symmetry in the indexing. The reader's flagged concern about the imported equality case of Lemma 9 is a reasonable caution, but it is a cited result and not the place where the argument actually breaks. The concrete internal flaw is the endpoint degeneracy: at λ=0 and λ=1 the statement 'equality if and only if K is a simplex' is simply false for every non-simplex body. This is a genuine correctness issue in the theorem as stated, though it is localized and easily repaired by restricting λ to (0,1). Because the fix is small and does not affect the main inequalities or Corollary 3, the appropriate verdict is conditional acceptance rather than rejection.","tokens_in":12235,"tokens_out":29636,"duration_ms":264304,"concrete_test":"Evaluate Eq. (5) at λ=0 for a Euclidean ball B in R^3: the left-hand side is Vol(B), so equality holds although B is not a simplex, contradicting the stated 'if and only if'. Then verify that after changing the quantifier to λ∈(0,1), the proof of Lemma 8 uses only interior λ, so the corrected equality statement holds and Corollary 3 remains valid.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Theorems 2 and 5 are stated for all λ∈[0,1] with equality if and only if K is a simplex. This is false at the endpoints. In Theorem 2, setting λ=0 leaves only the j=0 term in (5), giving (1−0)^n V(K[0],−K[n]) = Vol(−K)=Vol(K), so every convex body attains equality; the same happens at λ=1 via the j=n term. The same degeneration occurs in Theorem 5: at λ=0 only the m=n term survives, and its value is exactly Vol(K). The source of the problem is visible in the proof of Lemma 8: the argument divides by Vol(λ(1−λ)K), which vanishes at λ=0,1, so the equality analysis is degenerate there. The inequalities themselves are not in question, and Corollary 3 is unaffected since endpoints have measure zero, but the stated equality characterization of the central theorem requires the quantifier λ∈(0,1), with endpoints handled separately or excluded.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript studies Godbersen's conjecture via weighted sums of the mixed volumes V_j = V(K[j], -K[n-j]). Its main results are Theorem 2, a binomial-type bound sum_{j=0}^n lambda^j (1-lambda)^{n-j} V_j <= Vol(K), and Theorem 5, a more intricate double-sum bound, both with equality claimed exactly for simplices. Corollary 3 integrates Theorem 2 to obtain a uniform average form of Godbersen's conjecture. Section 4 shows how Theorems 2 and 5 would follow from the unbalanced difference body conjecture (Conjecture 6) and proves Conjecture 6 for n=4,5 by reducing to the known j=1 bound and the Rogers-Shephard inequality. Section 5 gives a new proof of an inequality from [2]. The arguments are non-circular and use Rogers-Shephard section-projection inequalities, a covariogram/Brunn-Minkowski argument, and the local Steiner formula.","tokens_in":12477,"tokens_out":26679,"duration_ms":230850,"significance":"Assuming the equality statements are corrected, the paper gives genuine progress on a longstanding conjecture: it proves a Godbersen-type bound on a uniform average for every body, generalizes Rogers-Shephard in a different direction, and settles the unbalanced difference body conjecture in dimensions 4 and 5. The proofs are self-contained and do not assume the conjectures they address; Theorem 7 is reduced to classical inequalities. The main weakness is the endpoint degeneracy in the equality statements, which is local and fixable rather than fatal to the inequalities themselves.","major_comments":[{"comment":"The equality characterization in Theorem 2 is false at lambda=0 and lambda=1. Setting lambda=0 in (5) leaves only the j=0 term, which equals Vol(-K)=Vol(K), so every convex body attains equality; the lambda=1 endpoint is analogous. Lemma 8 inherits the same problem: at lambda=0 the body C is a pyramid over K with volume Vol(K)/(n+1) for every K. In the proof this degeneracy is visible because the argument divides by Vol(lambda(1-lambda)K), which vanishes at the endpoints, so the homothety-based equality analysis is not valid there. The inequality (5) is unaffected, but the equality statement should be restricted to lambda in (0,1), with the endpoints noted as trivial equality cases or excluded.","section":"Section 2, Theorem 2 and Lemma 8"},{"comment":"The equality claim in Theorem 5 is false at lambda=0. At lambda=0 the double sum in (6) reduces to the m=n term, whose value is exactly Vol(K), so equality holds for every K. The proof also degenerates at both endpoints: B=lambda(K-K) is {0} at lambda=0 and K'=(1-lambda)K is {0} at lambda=1, so the local Steiner formula (9), which requires 0 in int(B), is not directly applicable. The correct statement is that equality characterizes simplices for lambda in (0,1); the case lambda=1 then follows from the Rogers-Shephard inequality, and lambda=0 is trivial.","section":"Section 3, Theorem 5"}],"minor_comments":[{"comment":"In the displayed formula for T_{theta,y}, the last set should be theta*lambda*K intersected with ((1-theta)(1-lambda)K - y), not with (1-theta)(1-lambda)K; the subsequent sentence appears to use the corrected version.","section":"Section 2, equality case of Lemma 8"},{"comment":"In the n=5 case the sentence that equality follows from the equality case of Rogers-Shephard alone is slightly compressed; when alpha-beta > 0 it also uses equality in V_1 <= 5, whose equality case gives the same simplex conclusion. Expanding this sentence would avoid an apparent gap.","section":"Section 4, Theorem 7"},{"comment":"The phrase \"with equality if and only if K is a simplex\" in the abstract and in the statements of Theorems 2 and 5 is overbroad; it should be qualified to lambda in (0,1) after the endpoint issue is resolved.","section":"Abstract and Introduction"}],"recommendation":"major_revision","confidential_remarks":"The endpoint issue is the only substantive obstruction I see; it is easily repaired by restricting the equality assertions to lambda in (0,1) and treating lambda=0 and lambda=1 separately. I would encourage a revised version rather than a rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The headline is simple: this is a good note, worth refereeing, with one small bug in an equality claim. The main results are new, correct inequalities in the direction of Godbersen's conjecture, and they deserve to be in the literature.\n\nWhat's actually new: Theorem 2 improves the coefficientwise bound from [2] to a full binomial-weighted sum, giving the uniform-average Corollary 3. That average is genuinely different from the Rogers-Shephard average, and the Markov/median consequences are a nice touch. Theorem 5 generalizes Rogers-Shephard through an unbalanced covariogram; the proof is clean. Theorem 7 settles the unbalanced difference body conjecture in dimensions 4 and 5, and Section 5 gives a pleasant geometric proof of a known dual inequality. I checked the algebra in the proofs of Theorems 2, 5, and 7, and it is correct. The reliance on the equality case of the Rogers-Shephard section-projection inequality from [2] is explicit and legitimate.\n\nThe soft spot: the equality statements in Theorems 2 and 5, and also in Lemma 8, are not true at λ=0 and λ=1. At those endpoints the left-hand side collapses to Vol(K) for every body, so equality is attained for all K, not only simplices. The cause is visible in the proof of Lemma 8: it divides by vol(λ(1-λ)K), which vanishes at the endpoints. The fix is trivial: state the theorems for λ∈(0,1), or explicitly say the equality characterization is only for the interior. The inequalities survive as written, and Corollary 3 is unaffected because endpoints are measure zero. A referee should require this correction, but it does not threaten the central content.\n\nOne more caveat in proportion: the equality cases for the interior also lean on an imported lemma ([9, Lemma 4]) and on the known equality case of Rogers-Shephard. These are the right tools, but the paper's own contribution to the equality analysis is mostly assembling them.\n\nWho is this for: anyone working in mixed volumes, convex geometry, or Rogers-Shephard type inequalities. It is a short note, clearly written, with no circularity and no overclaiming beyond the endpoint issue. I would bring it to our reading group and cite it.\n\nBest.","headline":"Solid short note: new binomial-weighted and uniform-average mixed-volume inequalities that would follow from Godbersen's conjecture, with a small but real equality-case bug at λ=0 and λ=1.","tokens_in":13027,"tokens_out":6944,"would_cite":true,"duration_ms":58279,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["52A39","52A40"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that Godbersen's conjecture holds on average for every convex body, with equality only for simplices.","keywords":["convex geometry","mixed volumes","Godbersen's conjecture","Rogers-Shephard inequality","simplex extremality","unbalanced difference body","covariogram","Brunn-Minkowski inequality"],"falsifier":"Search for a convex body $K$ that is not a simplex and a value $\\lambda\\in(0,1)$ for which every section $\\lambda(1-\\lambda)K\\cap(\\lambda(1-\\lambda)K+y)$ is homothetic to $\\lambda(1-\\lambda)K$ as $y$ varies; if such a body exists, the equality part of Lemma 8 and hence Theorem 2 would fail. Equivalently, numerically evaluate the sum in Theorem 2 for a non-simplex such as a cube or an ellipsoid and check whether equality holds for any $\\lambda$; any equality would falsify the theorem.","tokens_in":12053,"feed_emoji":"📐","tokens_out":6789,"duration_ms":56413,"temperature":0.7,"pith_summary":"This paper establishes unconditional bounds on weighted averages of Godbersen's mixed volumes $V(K[j],-K[n-j])$, quantities that measure how a convex body and its reflection interact. The main theorem, Theorem 2, shows that for every convex body $K\\subset\\mathbb{R}^n$ and every $\\lambda\\in[0,1]$, the sum $\\sum_{j=0}^n \\lambda^j(1-\\lambda)^{n-j}V(K[j],-K[n-j])$ is at most $\\operatorname{Vol}(K)$, with equality only for simplices. Averaging this over $\\lambda$ gives Corollary 3: $\\frac{1}{n+1}\\sum_{j=0}^n \\binom{n}{j}^{-1}V_j\\le \\operatorname{Vol}(K)$, a uniform version of Godbersen's conjecture \"on average\" that does not require any single $V_j$ to obey the conjecture. A second theorem generalizes the Rogers-Shephard inequality by bounding another weighted sum of mixed volumes, recovering the difference-body bound at one end and the known $j=1,n-1$ cases at the other. If correct, the results do not settle Godbersen's conjecture, but they show its extremal behavior—simplex maximality—survives in integrated form for every convex body.","feed_headline":"Godbersen's conjecture holds on average for every convex body","feed_subtitle":"New theorems prove binomial-weighted sums of the mixed volumes stay below the simplex value, equality only for simplices.","key_machinery":"Throughout, the paper works with Rogers-Shephard-type constructions: the $(n+1)$-dimensional body $C=\\operatorname{conv}(\\{0\\}\\times(1-\\lambda)K\\cup\\{1\\}\\times(-\\lambda K))$ in Lemma 8, whose volume is computed by slicing and expanding in mixed volumes, and a $(2n+1)$-dimensional body $T$ built from two scaled copies of $K$. The engine is Lemma 9, Rogers and Shephard's section-projection inequality, which bounds the product of the volume of a section and the volume of an orthogonal projection of a convex body. Fubini slicing of $C$ converts the bound on $\\operatorname{Vol}(C)$ into exactly the binomial-weighted sum of Theorem 2. Theorem 5 uses an \"unbalanced covariogram\" $f(x)=\\operatorname{Vol}((\\lambda K+x)\\cap K)/\\operatorname{Vol}(\\lambda K)$ and a local Steiner formula to produce a second weighted sum. Equality analysis is carried by the homothety statement in Lemma 9: equality forces all relevant sections to be homothetic, which pins down the simplex.","core_discovery":"The central claim is that several families of linear combinations of the mixed volumes $V_j=V(K[j],-K[n-j])$ are maximized, for fixed volume, by the $n$-simplex, even though the individual inequalities $V_j\\le \\binom{n}{j}\\operatorname{Vol}(K)$ (Godbersen's conjecture) remain open. Theorem 2 proves the binomial-weighted sum $\\sum_{j=0}^n \\lambda^j(1-\\lambda)^{n-j}V_j\\le\\operatorname{Vol}(K)$ for all $\\lambda\\in[0,1]$, with equality if and only if $K$ is a simplex; Theorem 5 proves an analogous bound for a second weighted sum involving $V(K[n-j],-K[j])$, again sharp exactly for simplices. Theorems 2 and 5 would both be immediate consequences of Godbersen's conjecture, and the paper shows conversely that both follow from the unbalanced difference body conjecture stated as Conjecture 6, which it proves in dimensions up to five. The method identifies the simplex as the unique maximizer in these averaged senses, so the extremal content of Godbersen's conjecture is correct even though the pointwise form remains unresolved.","pith_inferences":["A direct consequence the authors leave implicit is that differentiating Theorem 2 with respect to $\\lambda$ yields polynomial inequalities in $\\lambda$ whose coefficients constrain the differences $V_j-\\binom{n}{j}$; these may be easier to test than the individual conjecture.","The reduction in Section 4 suggests a program for Godbersen's conjecture: if the unbalanced difference-body ratio $\\operatorname{Vol}(D_\\lambda K)/\\operatorname{Vol}(K)$ can be shown to be maximized by the simplex in all dimensions, then both Theorem 2 and Theorem 5 follow by integration; the low-dimensional verification for $n\\le 5$ supports but does not prove this.","One could test Conjecture 6 numerically in dimension six or seven for random polytopes; since Theorem 2 and Theorem 5 already hold unconditionally, such tests would probe only the stronger unbalanced-difference-body statement, not the paper's main results."],"forward_implications":["Corollary 3 gives, for any convex body of volume one, $\\frac{1}{n+1}\\sum_{j=0}^n \\binom{n}{j}^{-1}V_j\\le 1$, so Godbersen's conjecture holds on average over $j$ for every body.","The median of the normalized quantities $\\binom{n}{j}^{-1}V_j$ is less than two, and Corollary 4 gives a Markov-type statement: for at least $k$ of the $n-1$ interior indices, $V_j\\le \\frac{n-1}{n-k}\\binom{n}{j}$ when $\\operatorname{Vol}(K)=1$.","Theorems 2 and 5 recover the known boundary cases $j=1$ and $j=n-1$ of Godbersen's conjecture, namely $V(K,-K[n-1])\\le n\\operatorname{Vol}(K)$.","Taking $\\lambda=1$ in Theorem 5 recovers the classical Rogers-Shephard inequality for the difference body, while Theorem 2 supplies a new proof of the earlier pointwise bound from reference [2].","Conjecture 6, which asks whether the unbalanced difference body $D_\\lambda K=(1-\\lambda)K-\\lambda K$ is volume-maximized by simplices, is proved in dimensions $n\\le 5$, so simplex maximality of $D_\\lambda K$ is verified in low dimensions."],"supporting_citations":[{"why":"Supplies the convex-body construction $T$ and the section-projection inequality that drive the proof of Theorem 2 via Lemma 8.","marker":"[8]"},{"why":"Provides the difference-body inequality whose proof is generalized and the simplex characterization used for equality cases.","marker":"[9]"},{"why":"Gives the pointwise bound (4) that Theorem 2 improves, and states the full equality condition for Lemma 9 used in the equality analysis.","marker":"[2]"},{"why":"Contains the local Steiner formula, equation (9), that turns the covariogram estimate into the mixed-volume sum of Theorem 5.","marker":"[13]"},{"why":"Supplies the standard mixed-volume properties, the exposed-point fact, and convex-body background used throughout.","marker":"[12]"},{"why":"States Godbersen's conjecture, the target that the paper verifies on average.","marker":"[5]"},{"why":"Establishes the boundary cases $j=1,n-1$ of Godbersen's conjecture that Theorem 2 recovers.","marker":"[11]"}],"fun_headline_variants":["Godbersen’s conjecture true on average for all convex bodies","Weighted mixed volumes maximized by simplices","Averaged Godbersen inequalities: simplex is extremal","Godbersen’s conjecture proved on average","Simplex maximizes weighted sums of mixed volumes"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that Rogers and Shephard's equality characterization in the section-projection inequality is complete: if equality holds in Lemma 9, then all relevant sections are homothetic; the inequalities themselves do not depend on this, but the assertions that equality occurs only for simplices do.","fun_headline_variants_meta":{"raw":{"variants":["Godbersen’s conjecture true on average for all convex bodies","Weighted mixed volumes maximized by simplices","Averaged Godbersen inequalities: simplex is extremal","Godbersen’s conjecture proved on average","Simplex maximizes weighted sums of mixed volumes"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000872,"raw_usage":{"total_tokens":3799,"prompt_tokens":993,"completion_tokens":2806,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":609,"completion_tokens_details":{"reasoning_tokens":2729}},"tokens_in":609,"tokens_out":2806,"duration_ms":17666,"temperature":1.0,"reasoning_tokens":2729,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T11:05:18.459310+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Search for a convex body $K$ that is not a simplex and a value $\\lambda\\in(0,1)$ for which every section $\\lambda(1-\\lambda)K\\cap(\\lambda(1-\\lambda)K+y)$ is homothetic to $\\lambda(1-\\lambda)K$ as $y$ varies; if such a body exists, the equality part of Lemma 8 and hence Theorem 2 would fail. Equivalently, numerically evaluate the sum in Theorem 2 for a non-simplex such as a cube or an ellipsoid and check whether equality holds for any $\\lambda$; any equality would falsify the theorem.","supporting_citations":[{"cited_title":"Convex bodies asso ciated with a given convex body","cited_arxiv_id":null,"evidence_quote":"Supplies the convex-body construction $T$ and the section-projection inequality that drive the proof of Theorem 2 via Lemma 8."},{"cited_title":"The difference body of a convex body","cited_arxiv_id":null,"evidence_quote":"Provides the difference-body inequality whose proof is generalized and the simplex characterization used for equality cases."},{"cited_title":"Florentin, and Ya ron Ostrover","cited_arxiv_id":null,"evidence_quote":"Gives the pointwise bound (4) that Theorem 2 improves, and states the full equality condition for Lemma 9 used in the equality analysis."},{"cited_title":"Stochastic and integral geometry , volume 1","cited_arxiv_id":null,"evidence_quote":"Contains the local Steiner formula, equation (9), that turns the covariogram estimate into the mixed-volume sum of Theorem 5."},{"cited_title":"Convex Bodies: the Brunn-Minkowski Theory , volume 151 of Encyclopedia of Mathematics and its Applications","cited_arxiv_id":null,"evidence_quote":"Supplies the standard mixed-volume properties, the exposed-point fact, and convex-body background used throughout."},{"cited_title":"Der Satz vom Vektorbereich in Raumen beliebiger Dimension","cited_arxiv_id":null,"evidence_quote":"States Godbersen's conjecture, the target that the paper verifies on average."},{"cited_title":"Stability for some extremal properties of the s implex","cited_arxiv_id":null,"evidence_quote":"Establishes the boundary cases $j=1,n-1$ of Godbersen's conjecture that Theorem 2 recovers."}],"review_version":1}