{
  "packet_id": "full_manuscript",
  "scope_summary": "Main conclusion: Theorem 1 \u2014 the product K=T_10\u00d7T_10\u2282R^20 has R_20(K)/c_20 = 121\u00b7C(20,10)/(21\u00b72^20) = 22355476/22020096 > 1, so K violates the proposed simplex upper bound |\u03a0K| \u2264 c_d|K|^{d-1} (Brannen's conjectured maximum) in dimension 20; Corollary 7 adds exponential excess \u03bb^n for all large n via products of simplices. All six proof-bearing claims (Lemma 2 facet/zonotope formula, Proposition 3 product identity, Lemma 4 affine covariance, Proposition 5 simplex value, Theorem 1, Corollary 7) plus the unnumbered general product-test formula were audited, including independent recomputation of every numeric constant and low-dimensional consistency checks (d=2 via Rogers\u2013Shephard, d=3 via c_3=18).",
  "audited_steps": [
    {
      "anchor": "Lemma 2 (lem:facets), eqs (eq:facet),(eq:zonotope)",
      "result": "checked",
      "evidence": "Cauchy projection formula for polytopes: for unit u, vol_{d-1}(proj_{u^\u22a5}P)=(1/2)\u03a3_F s_F|\u27e8\u03bd_F,u\u27e9| since a.e. point of the projection is covered by exactly two facet projections (entry/exit), each facet projecting with Jacobian |\u27e8\u03bd_F,u\u27e9|; exceptional sets (boundary of projection, projections of faces of dim \u2264 d\u22122, facets parallel to u) lie in finitely many affine subspaces of dim \u2264 d\u22122 or in the boundary of a convex set, hence null. h_{[-v/2,v/2]}(u)=|\u27e8v,u\u27e9|/2 and support functions add under Minkowski sum, so (eq:zonotope) follows. Standard; proof is sketch-level but correct."
    },
    {
      "anchor": "Proposition 3 (prop:product), eqs (eq:product-body),(eq:multiplicative)",
      "result": "checked",
      "evidence": "Exposed faces of A\u00d7B under (u,v) are products of exposed faces; a face of dimension r+s\u22121 must be F\u00d7B or A\u00d7G. Areas: vol_{r+s-1}(F\u00d7B)=s_F|B| (orthogonal factors), normal (\u03bd_F,0). Lemma 2 gives h_{\u03a0(A\u00d7B)}(u,v)=|B|h_{\u03a0A}(u)+|A|h_{\u03a0B}(v)=h_{(|B|\u03a0A)\u00d7(|A|\u03a0B)}(u,v). Volume: |B|^r|\u03a0A|\u00b7|A|^s|\u03a0B|/(|A||B|)^{r+s-1}=(|\u03a0A|/|A|^{r-1})(|\u03a0B|/|B|^{s-1}). Correct."
    },
    {
      "anchor": "Lemma 4 (lem:affine)",
      "result": "checked",
      "evidence": "For polytopes: L^{-t}\u03bd_F is outward for LF since \u27e8L^{-t}\u03bd,Lx\u2212Ly\u27e9=\u27e8\u03bd,x\u2212y\u27e9\u22640; prism argument: image prism LF+[0,L\u03bd_F] has volume |det L|s_F and height \u27e8L\u03bd_F,L^{-t}\u03bd_F/\u2016L^{-t}\u03bd_F\u2016\u27e9=1/\u2016L^{-t}\u03bd_F\u2016, so s_{LF}=|det L|\u2016L^{-t}\u03bd_F\u2016s_F and area-normal vectors map by M=|det L|L^{-t}. Zonotope maps by M, |det M|=|det L|^{d}|det L|^{-1}=|det L|^{d-1}, matching |LP|^{d-1}. Polytope approximation: aB\u2282P_m gives \u03b5_mB\u2282(\u03b5_m/a)P_m so P\u2282(1+\u03b5_m/a)P_m; 0\u2264F\u2212f_m\u2264(\u03bb_m^{d-1}\u22121)\u03ba_{d-1}R^{d-1} by monotonicity; limit passage valid. Only the polytope case is actually used downstream."
    },
    {
      "anchor": "Proposition 5 (prop:simplex)",
      "result": "checked",
      "evidence": "|T_d|=1/d!. Coordinate facets: area 1/(d\u22121)!, normal \u2212e_i. Slanted facet: projection to e_d^\u22a5 is T_{d-1} (volume 1/(d\u22121)!) with Jacobian |\u27e8w/\u221ad,e_d\u27e9|=1/\u221ad, so area \u221ad/(d\u22121)! and area-normal w/(d\u22121)!. \u03a0T_d is a translate of ([0,1]^d+[0,w])/(d\u22121)!. |Q+[0,w]|=|Q|+\u2016w\u2016vol_{d-1}(proj_{w^\u22a5}Q)=1+h_{\u03a0Q}(w), h_{\u03a0Q}(w)=(1/2)\u00b72d\u00b71=d, so |Q+[0,w]|=d+1 (d=2: hexagon area 3; d=3: 1+\u221a3\u00b7\u221a3=4). |\u03a0T_d|=(d+1)/((d\u22121)!)^d; R_d=(d+1)(d!)^{d-1}/((d\u22121)!)^d=(d+1)d^d/d!. Consistency: c_2=6=Rogers\u2013Shephard equality value |T\u2212T|/|T| for triangles; c_3=18 matches cited 3-dimensional bound."
    },
    {
      "anchor": "Theorem 1 (thm:counterexample), proof in Section 4",
      "result": "checked",
      "evidence": "R_20(K)=c_10^2 by Prop 3 (r=s=10\u22652) and Prop 5. c_10^2/c_20=(121\u00b710^20/(10!)^2)\u00b7(20!/(21\u00b720^20))=121\u00b7C(20,10)/(21\u00b72^20). C(20,10)=184756; 121\u00b7184756=22355476; 21\u00b72^20=22020096; difference 335380>0. Verified. Also c_10=17187500/567 verified (10^10/10!=4\u00b75^8/567)."
    },
    {
      "anchor": "Unnumbered product-test formula after Theorem 1 proof",
      "result": "checked",
      "evidence": "c_rc_s/c_{r+s}=(r+1)(s+1)r^rs^s(r+s)!/(r!s!(r+s+1)(r+s)^{r+s})=((r+1)(s+1)/(r+s+1))\u00b7C(r+s,r)\u00b7r^rs^s/(r+s)^{r+s}. Matches. Spot check r=s=9: 100\u00b748620/(19\u00b72^18)<1, so dimension-18 symmetric product does not violate; r=s=10 is the first symmetric violation, consistent with the Theorem."
    },
    {
      "anchor": "Corollary 7 (cor:exponential-excess)",
      "result": "checked",
      "evidence": "n^n/n!<e^n (one term of the series), e<5/2+(1/6)\u03a33^{-j}=5/2+1/4=11/4=\u03b2 (tail ratios 1/(j+1)\u22641/4<1/3). c_n<(n+1)\u03b2^n. A=c_10/\u03b2^10=17187500\u00b74^10/(567\u00b711^10)=18022400000000/14706519748767=1638400000000/1336956340797\u22481.2255>1 (both reduced by common factor 11; verified 1336956340797\u00b711=14706519748767). With n=10k+d_r: R_n(K_n)/(c_n\u03bb^n)>c_{d_r}\u03b2^{-d_r}\u03bb^{-d_r}(A/\u03bb^{10})^k/(10k+d_r+1)\u2192\u221e for \u03bb<A^{1/10}. Representation n=10k+d_r, k\u22651, unique for n\u226520. Prop 3 iterates since all factors are full-dimensional polytopes of dimension \u226510."
    }
  ],
  "objections": [
    {
      "anchor": "Lemma 2 proof, paragraph 'Apart from a set of (d\u22121)-dimensional measure zero...'",
      "classification": "unverified_step",
      "claim": "The multiplicity-two argument is stated at sketch level: it is not explicitly argued that every point of the interior of proj_{u^\u22a5}P lies on a line meeting int P (so that the line meets P in a nondegenerate interval), nor that the endpoints then lie in facet relative interiors after discarding lower-dimensional faces.",
      "evidence": "The proof says 'Every remaining line meets P in a nondegenerate interval with its two endpoints in facet interiors' without justification.",
      "derivation": "Repairable: proj(int P) is open, convex, dense in proj P, so int(proj P)=proj(int P); hence each line over an interior point meets int P in a nondegenerate open interval whose endpoints are boundary points; after discarding projections of faces of dim \u2264 d\u22122 (null in u^\u22a5) the endpoints lie in facet relative interiors, and such a facet cannot be parallel to u.",
      "hypotheses_checked": [
        "P full-dimensional polytope",
        "u unit vector",
        "exceptional sets contained in finitely many affine subspaces of dim \u2264 d\u22122 or in \u2202(proj P)"
      ],
      "possible_repair": "Insert the two sentences above, or cite Cauchy's projection formula (e.g. Schneider, Convex Bodies, \u00a75.3) directly.",
      "dependency_impact": "None on correctness: the formula is classical and the stated factor 1/2 is correct; all downstream constants are consistent with independent checks (c_2=6, c_3=18)."
    },
    {
      "anchor": "Introduction, attribution 'Brannen conjectured in 1996 that simplices maximize R_d among all convex bodies [Brannen]' and citations [ChenEtAl],[FengEtAl]",
      "classification": "unverified_step",
      "claim": "The historical framing (exact statement of Brannen's conjecture, and the 2026 preprints asserting d=3 maximality and d\u22659 counterexamples) is external and not audited here.",
      "evidence": "These statements are not used in any proof; the mathematical content of Theorem 1 is self-contained.",
      "derivation": "Not applicable; the theorem's inequality R_20(K)>c_20 holds independently of the attribution.",
      "hypotheses_checked": [
        "Theorem 1 depends only on Lemma 2, Prop 3, Prop 5"
      ],
      "possible_repair": "None needed for the mathematics; attribution should be checked against Brannen (1996).",
      "dependency_impact": "No effect on the validity of Theorem 1 or Corollary 7."
    }
  ],
  "unverified_obligations": [
    {
      "anchor": "[Brannen] Mathematika 43 (1996)",
      "obligation": "Confirm that Brannen's conjecture is the simplex maximum for |\u03a0K|/|K|^{d-1} in all dimensions, as stated in the Introduction."
    },
    {
      "anchor": "[ChenEtAl], [FengEtAl] 2026 preprints",
      "obligation": "External results (R_3\u226418 with tetrahedra extremal; counterexamples for d\u22659) are cited for context only and were not audited."
    },
    {
      "anchor": "Lemma 4, general convex body case",
      "obligation": "The polytope-approximation passage is not used by Theorem 1 or Corollary 7 (both involve only polytopes); it was checked but is not load-bearing."
    }
  ],
  "assessment": "no_demonstrated_error_found"
}