{
  "packet_id": "full_manuscript",
  "scope_summary": "The main conclusion is that K=T_10\u00d7T_10 violates the displayed simplex upper bound in dimension 20, with R_20(K)/c_20=22,355,476/22,020,096>1. The secondary conclusion gives exponential excess over the simplex value in every sufficiently large dimension. All six formal proof-bearing claims were audited, including the arbitrary-convex-body approximation argument, the general product test, boundary-direction checks, and the residue-class construction. No demonstrated error was found in these arguments. The introductory assertions attributed to external publications were not audited externally and are not prerequisites of the supplied proofs.",
  "audited_steps": [
    {
      "anchor": "Lemma 2 (lem:facets), equations eq:facet and eq:zonotope, especially the projection-multiplicity paragraph",
      "result": "checked",
      "evidence": "For unit u, the projection Jacobian on a facet is |\u03bd_F\u00b7u|. The projected lower-dimensional faces and facets parallel to u have dimension at most d\u22122; the boundary of the full-dimensional projection also has measure zero. Every remaining fiber has exactly two transverse facet-interior endpoints. Consequently, \u03a3_F s_F|\u03bd_F\u00b7u|=2 vol_{d\u22121}(proj_{u\u22a5}P). Homogeneity handles nonunit u, including u=0. The centered segment sum has precisely half this sum as its support function, and the preceding separation argument justifies identification of the bodies. Boundary check: for P=[0,1]^2 and u=e_1, the parallel edges contribute zero and the other two contribute one each, yielding projection length one and confirming the factor 1/2."
    },
    {
      "anchor": "Proposition 3 (prop:product), equations eq:product-body and eq:multiplicative",
      "result": "checked",
      "evidence": "Full dimensionality ensures that a nonzero functional on either factor exposes a proper face. Thus a facet of A\u00d7B must be F\u00d7B or A\u00d7G; a functional with both blocks nonzero exposes a face of codimension at least two. Orthogonality gives the stated area-normal vectors (|B|s_F\u03bd_F,0) and (0,|A|s_G\u03bd_G). Checked Lemma 2 therefore gives h_{\u03a0(A\u00d7B)}(u,v)=|B|h_{\u03a0A}(u)+|A|h_{\u03a0B}(v), the support function of the asserted Cartesian product. Its volume is |B|^r|A|^s|\u03a0A||\u03a0B|. Dividing by (|A||B|)^{r+s\u22121} leaves exactly R_r(A)R_s(B). All volumes are positive under the stated hypotheses. As a scaling check, \u03a0[0,1]^d=[\u22121,1]^d and R_d([0,1]^d)=2^d, consistent with the product identity."
    },
    {
      "anchor": "Lemma 4 (lem:affine), the transformed-prism calculation and the approximation sandwich P_m\u2282P\u2282\u03bb_mP_m",
      "result": "checked",
      "evidence": "The transformed outward normal is L^{\u2212t}\u03bd/||L^{\u2212t}\u03bd||. The image prism has height 1/||L^{\u2212t}\u03bd||, because (L\u03bd)\u00b7(L^{\u2212t}\u03bd)=1. Hence its area-normal vector is |det L|L^{\u2212t}(s_F\u03bd_F), including orientation-reversing L. Lemma 2 gives covariance for polytopes. For M=|det L|L^{\u2212t}, |det M|=|det L|^{d\u22121}, exactly matching the denominator scaling. For general bodies, the fixed inner ball and finite-net construction give aB\u2282P_m\u2282P\u2282\u03bb_mP_m with \u03bb_m\u21921. Consequently f_m\u2264F\u2264\u03bb_m^{d\u22121}f_m, with f_m uniformly bounded by \u03ba_{d\u22121}R^{d\u22121}. The bounded increasing zonotopes \u03a0P_m therefore have a full-dimensional union closure with support F, establishing existence rather than assuming it. The volume sandwich has the correct exponent d(d\u22121). Applying L preserves the scalar sandwich, so the polytopal identity passes to the limit. Translation invariance follows from translation invariance of projection volumes."
    },
    {
      "anchor": "Proposition 5 (prop:simplex), the facet-vector calculation and the fiber computation for Q+[0,w]",
      "result": "checked",
      "evidence": "The coordinate facet vectors are \u2212e_i/(d\u22121)!, while the remaining facet has area \u221ad/(d\u22121)! and normal w/\u221ad, giving vector w/(d\u22121)!. Checked Lemma 2 thus makes \u03a0T_d a translate of (Q+[0,w])/(d\u22121)!. Adding [0,w] leaves the projection onto w\u22a5 unchanged and increases each nonempty parallel fiber's length by ||w||. Fubini gives volume 1+h_{\u03a0Q}(w); the cube facet formula gives h_{\u03a0Q}(w)=\u03a3_i|w_i|=d. Thus |\u03a0T_d|=(d+1)/((d\u22121)!)^d, and division by (1/d!)^{d\u22121} gives (d+1)d^d/d!. Checked Lemma 4 applies because every full-dimensional simplex is an invertible affine image of T_d. Local checks give |\u03a0T_2|=3, R_2(T_2)=6, and R_3(T_3)=18."
    },
    {
      "anchor": "Theorem 1 (thm:counterexample), its proof in Section 4, and the immediately following general product-test formula",
      "result": "checked",
      "evidence": "T_10\u00d7T_10 is a compact, full-dimensional convex polytope in the stipulated orthogonal product space. Checked Propositions 3 and 5 give R_20(K)=c_10^2. Dividing by c_20 yields 121\u00b7binom(20,10)/(21\u00b72^20). The integer values are binom(20,10)=184,756, numerator 22,355,476, denominator 22,020,096, and difference 335,380>0. This proves a strict violation without numerical approximation or external results. More generally, direct substitution gives c_rc_s/c_{r+s}=((r+1)(s+1)/(r+s+1))\u00b7binom(r+s,r)\u00b7r^r s^s/(r+s)^{r+s}, as displayed. Any universal upper constant must therefore be at least c_rc_s. As a low-dimensional check, r=s=2 gives ratio 27/40<1; the manuscript does not claim that every product exceeds the simplex."
    },
    {
      "anchor": "Corollary 6 (cor:exponential-excess), the definitions of d_r, \u03b2, A, and the final residue-class estimate",
      "result": "checked",
      "evidence": "For n\u226520, taking r=n mod 10 and k=floor(n/10)\u22121 gives the unique representation n=10k+(10+r), with k\u22651. All factors have dimension at least ten, so repeated application of checked Propositions 3 and 5 gives R_n(K_n)=c_10^k c_{d_r}; no dimension-one remainder is needed. The factorial tail ratios are at most 1/4<1/3, proving e<5/2+(1/6)/(1\u22121/3)=11/4=\u03b2. The n-th term of exp(n) gives c_n<(n+1)\u03b2^n. The exact values c_10=17,187,500/567 and A=1,638,400,000,000/1,336,956,340,797 are correct; the latter numerator exceeds its denominator by 301,443,659,203. Thus the required growth-rate prerequisite A>1 is independently verified, not inferred merely from Theorem 1. For any 1<\u03bb<A^(1/10), the displayed lower bound is a positive residue-dependent constant times (A/\u03bb^10)^k/(10k+d_r+1), which tends to infinity. There are only ten residue classes, so their thresholds yield one N valid for every n\u2265N."
    }
  ],
  "objections": [],
  "unverified_obligations": [
    {
      "anchor": "Introduction, the Brannen conjecture attribution and references [Brannen], [Petty], [Ludwig, Equation (1)], and [Saroglou, Section 1]",
      "obligation": "Confirm the historical attributions, the scope and normalization of Brannen's conjecture, and the cited discussion distinguishing the projection-body functional from polar projection inequalities and Petty's minimum problem. These external sources and proofs were not supplied or audited. The needed facet and affine-covariance identities were independently checked from the manuscript, so this obligation concerns the literature account rather than the internal counterexample proof."
    },
    {
      "anchor": "Introduction, paragraph beginning 'In dimension three'; [ChenEtAl, Theorems 1.1 and 1.2]",
      "obligation": "Inspect the external theorems and their proofs to verify the claimed solution of the three-dimensional minimum problem and the universal bound R_3(K)\u226418, including their hypotheses and normalization. Neither result is used in the six supplied proofs."
    },
    {
      "anchor": "Introduction, the assertions about [FengEtAl, Theorem 1.3 and Section 4]",
      "obligation": "Verify externally the counterexamples in every dimension d\u22659, the at-most-d+2-facets assertion, and the stated construction using perturbed cube projections and Minkowski's existence theorem. Their proofs were not provided and are not dependencies of the explicit product witness."
    },
    {
      "anchor": "Introduction, paragraph beginning 'The growth of upper constants'; [Henk]",
      "obligation": "Verify the stated lower bound 2^d(9/8)^{floor(d/3)} for the optimal upper constant in the centrally symmetric class, with the same normalization of R_d. The external proof was not audited; Corollary 6 instead supplies an independent argument for its own claim."
    }
  ],
  "assessment": "no_demonstrated_error_found"
}