{"id":"00b56b80-ee67-4554-8955-a8cc6ea73de0","arxiv_id":"1908.07118","paper_version":3,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Extremal functions of real convex polytopes equal the maximum of extremal functions of supporting simplices and strips, yielding a self-contained proof of the existence of extremal ellipses for all real convex bodies.","lead":"This paper gives a method to compute the extremal function of a real convex polytope by decomposing it into simpler sets, simplices and strips. It uses this to reprove, without Lempert theory, that every real convex body has inscribed extremal ellipses on which the extremal function is harmonic.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 13.1's approximation step proves V_K(f(ζ_z))=log|ζ_z| only at one point, but condition (ii) requires equality for all |ζ|>1; the missing propagation (via uniform convergence or maximum principle) is the load-bearing gap.","rationale":"Focused on the two central claims. Theorem 5.4's proof is sound: the strip lemma 5.3 is justified by continuity of V_{B_1} at the origin, and the product/diagonal argument in the proof of Theorem 5.4 is correct. Proposition 10.6 has a small unproved construction, but an explicit choice of the off-diagonal coefficient m=√((1−A²)(1−B²))/(AB) makes the matrix [[1/A²,m],[m,1/B²]] have smallest eigenvalue 1, so that gap is fillable. The genuinely load-bearing soft spot is the final approximation step in Theorem 13.1: the equality V_K(f(ζ))=log|ζ| is shown at one parameter value only, yet it is asserted for all |ζ|>1. This is not a defect in the underlying mathematics; it is a missing standard argument, either via Dini's theorem applied to the monotone sequence V_{K_j} or via the maximum principle on the subharmonic function u. Because the gap is localized and fillable, the appropriate verdict remains CONDITIONAL rather than REJECT or ACCEPT. I agree with the reader that the Section 13 approximation is the weakest premise; my check isolates the precise unproven implication.","tokens_in":30430,"tokens_out":19997,"duration_ms":188997,"concrete_test":"Run the following analytical check on the limiting ellipse E_C obtained in Theorem 13.1: (1) prove E⊂K by the stated limit argument, so V_K=0 on E; (2) verify that u(ζ)=V_K(f(ζ))−log|ζ| is subharmonic on {|ζ|>1}, since V_K is psh and f is holomorphic, while log|ζ| is harmonic; (3) note u≥0 because V_K≥V_{K_j} and V_{K_j}(f_j(ζ))=log|ζ|. If u has boundary value 0 on the unit circle, the maximum principle forces u≡0, giving V_K(f(ζ))=log|ζ| for every |ζ|>1. If this check goes through, the gap in Theorem 13.1 is closed; if not, the theorem is unproved in its current form.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In the proof of Theorem 13.1, after the normal-families limit f_j→f and ζ_j→ζ_z, the text says: 'By the pointwise convergence V_{K_j}↗V_K (Proposition 5.1(3)) we have V_K(f(ζ_z))=... Hence (i) and (ii) hold.' This establishes (ii) only at the single parameter value ζ_z. For an arbitrary ζ with |ζ|>1, the available facts are V_{K_j}≤V_K and V_{K_j}(f_j(ζ))=log|ζ|, which give only V_K(f(ζ))≥log|ζ|; they do not yield equality. Equality requires either locally uniform convergence V_{K_j}→V_K (which follows from Dini's theorem since V_{K_j} and V_K are continuous and the convergence is monotone, but this is not stated) or a maximum principle argument on E_C showing that u(ζ)=V_K(f(ζ))−log|ζ| is subharmonic on the exterior, has boundary value 0 on |ζ|=1, and is nonnegative, forcing u≡0. The paper also asserts without detail that the approximating polytopes K_j can be chosen with the generic normal-independence condition and that the coefficients a_j,c_j are bounded; these assertions are true but are precisely the unstated ingredients on which the convergence argument rests. Since condition (ii) is used globally in all applications of Theorem 13.1, the proof as written is incomplete at this point.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops an explicit method for computing the Siciak–Zaharjuta extremal function of a compact convex polytope in R^d in terms of a finite family of supporting simplices and strips, and uses this to reprove the existence of extremal ellipses for general real convex bodies. The main theorems are Theorem 5.4, asserting V_K(z)=max{V_S(z): S∈S(K)} for polytopes, and Theorem 13.1, asserting that for every z outside a compact convex body K there is a complexified ellipse through z on which V_K equals log|ζ| and whose real trace is inscribed in K. The proof route goes through geometric lemmas on simplices and strips, explicit extremal functions for balls and simplices, Hooke and Newton ellipses, the Robin exponential map for polytopes, and a final approximation argument from polytopes to general convex bodies.","tokens_in":30698,"tokens_out":15361,"duration_ms":152981,"significance":"If the main results are correct, the paper provides a self-contained and largely elementary route to structural facts about extremal functions for real convex bodies that were previously obtained using Lempert theory. The explicit barycentric-coordinate formulas and the constructive polytope algorithm in Section 7 are valuable, as is the new derivation of Lundin and Baran formulas through the square map. The Robin exponential map for polytopes and the approximation scheme for general convex bodies are potentially useful tools. However, the proof of the central polytope formula contains a false convex-geometric step, and the approximation step in Theorem 13.1 has a gap; both need to be addressed before the main claims are established.","major_comments":[{"comment":"The step 'The set Π is a product of convex sets, hence convex, and therefore Π⊂π_Δ^{-1}(K_Δ)' is false: orthogonal projection of a convex set onto a subspace need not be contained in the set. The paper's own Section 7 example gives a concrete counterexample. Let S_3={x_1≥0, x_2≥0, 3−x_1−3x_2≥0} and S_4={x_1≥0, x_2≥0, 3−3x_1−x_2≥0}; these are the two supporting triangles in S(K) for the quadrilateral of Section 7. The points x=(2,0.2)∈S_3 and y=(0.2,2)∈S_4 satisfy π_Δ(x,y)=((1.1,1.1),(1.1,1.1)), and (1.1,1.1) lies in neither S_3 nor S_4, so π_Δ(x,y)∉K_Δ even though (x,y)∈Π. Thus the claimed inclusion Π⊂π_Δ^{-1}(K_Δ) fails in the exact class of examples considered. Since the reverse inequality V_K≤max_j V_{S_j} is the core of Theorem 5.4 and is used later in Lemma 8.2 and in the Robin exponential map construction, this is a load-bearing gap in the proof.","section":"Section 5, proof of Theorem 5.4"},{"comment":"The normal-families argument proves V_K(f(ζ_z))=log|ζ_z| only at the single parameter value ζ_z, whereas condition (ii) requires equality for every ζ with |ζ|>1. The sentence 'Hence (i) and (ii) hold' is not justified by pointwise convergence alone. The gap is repairable: since the sets K_j and K are regular by Lemma 5.2, the functions V_{K_j} and V_K are continuous, and Proposition 5.1(3) gives monotone pointwise convergence V_{K_j}↗V_K; Dini's theorem on compact sets then gives locally uniform convergence, and combined with f_j→f locally uniformly this yields V_K(f(ζ))=log|ζ| for all |ζ|>1. As written, however, the proof is incomplete at a point that is essential for all applications of Theorem 13.1.","section":"Section 13, proof of Theorem 13.1"}],"minor_comments":[{"comment":"The statement says 'polynomial mapping of degree d' and gives the formula V_{P^{-1}(K)}(z)=1/d V_K(P(z)), but the correct factor is the reciprocal of the degree of the polynomial mapping, not the dimension d. As stated, the proposition is false for a linear map when d>1. The later use in Corollary 6.4 correctly uses the factor 1/2 for the degree-2 square map, so this is likely a typographical error, but it should be corrected.","section":"Section 5, Proposition 5.1(2)"},{"comment":"The labels l_3 and l_4 for the two non-axis supporting lines are interchanged between Remark 3.8 and Section 7: Remark 3.8 has l_3(x)=3−3x_1−x_2 and l_4(x)=3−x_1−3x_2, while Section 7 defines l_3(z)=3−z_1−3z_2 and l_4(z)=3−3z_1−z_2. Please align the notation to avoid confusion in checking the example.","section":"Section 7"},{"comment":"The existence of a decreasing sequence of polytopes K_j satisfying the linear independence condition of Theorem 3.9 and converging to K is asserted without proof. This is a standard genericity/refinement argument, but since the convergence step of Theorem 13.1 depends on it, a brief justification would improve the exposition.","section":"Section 13"}],"recommendation":"major_revision","confidential_remarks":"The false inclusion in the proof of Theorem 5.4 is the most serious issue; the author should be asked to supply a correct proof of that theorem before the paper can be accepted. The Theorem 13.1 gap is more easily repairable, but it should also be addressed explicitly. The paper's reliance on earlier formulas from [4] and [5] should be checked for novelty disclosure once the technical issues are resolved."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Ma'u's paper is a genuinely useful piece of elementary pluripotential geometry. The main new result is Theorem 5.4: for a compact convex polytope, the Siciak-Zaharjuta extremal function is the maximum of the extremal functions of finitely many supporting simplices and strips. That decomposition is real and new as far as I can tell, and the convex-geometric construction behind it is worked out in satisfying detail. The second half builds ellipses on which V_K is harmonic, first for balls and simplices, then for polytopes via the Robin exponential map, and finally for general convex bodies by approximation. The upshot is a new proof of the Burns-Levenberg-Ma'u existence theorem that avoids Lempert theory entirely. That is a legitimate contribution: the paper makes the existence of extremal ellipses more accessible and gives an explicit computational recipe for polytopes.\n\nThe soft spots are concentrated in the last step. In Theorem 13.1, after taking the normal-families limit f_j→f and ζ_j→ζ_z, the text claims (i) and (ii) hold, but the displayed computation only establishes V_K(f(ζ_z)) = log|ζ_z| at that single parameter value. Condition (ii) requires equality for every |ζ|>1. The missing argument is standard: monotone convergence of continuous extremal functions gives local uniform convergence (Dini), and then uniform convergence on compact annuli transfers the equalities from the approximating ellipses to the limit. The paper even notes uniform convergence on compact annular regions, but does not connect the dots. A referee should ask for that one paragraph. The other asserted ingredient, that the approximating polytopes can be chosen with the generic linear-independence condition and with a_j, c_j bounded, is also true but under-documented; it needs a sentence or two.\n\nNone of this shakes the central argument. Section 5.4 is the real result, and it is proved cleanly. The reliance on earlier papers, including the author's own [4] and [6], is for standard formulas (barycentric coordinates, Robin exponential terminology) rather than for the main novelty, so I do not see a circularity problem. The paper is honest about what it reproves and what is new.\n\nThis is a paper I would send to a competent referee. It deserves a careful reading and likely a minor revision. I would cite it if I worked on extremal functions for real convex bodies, and I would bring it to our reading group.","headline":"Solid, mostly elementary paper with a genuinely new polytope decomposition and a minor but easily patched gap in the final approximation argument.","tokens_in":31274,"tokens_out":3039,"would_cite":true,"duration_ms":28678,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["32U05","32U15","52A20"],"pacs":[],"model":"deepseek-v4-flash","headline":"Every point outside a real convex body lies on an extremal complex ellipse.","keywords":["Siciak-Zaharjuta extremal function","pluripotential theory","real convex body","convex polytope","extremal ellipse","Robin exponential map","barycentric coordinates"],"falsifier":"For a concrete test, take the unit square in $\\mathbb R^2$, choose an exterior point such as $z=(2+i,3)$, solve the equations $z=a+c\\zeta+c/\\zeta$ numerically for $a\\in\\mathbb R^2$, $c\\in\\mathbb C^2$, and $|\\zeta|>1$, and compare $\\log|\\zeta|$ with a high-degree polynomial-envelope approximation of $V_K(a+c\\zeta+c/\\zeta)$. A nonzero difference for every such solution would falsify the theorem.","tokens_in":30175,"feed_emoji":"📐","tokens_out":8750,"duration_ms":92779,"temperature":0.7,"pith_summary":"The paper establishes an explicit decomposition for the Siciak-Zaharjuta extremal function of a real convex polytope: it is the maximum of the extremal functions of a finite list of supporting simplices and strips. This reduces polytope computations to barycentric-coordinate formulas on simplices together with a dimension-reduction formula for strips. Using this polytope result, the author gives a new proof that for every real convex body and every exterior point there is a complex ellipse through that point on which the extremal function equals $\\log|\\zeta|$, with the real ellipse inscribed in the body. The significance is a self-contained route to the ellipse foliation that avoids the earlier geodesic-based construction.","feed_headline":"Every point outside a convex body lies on an extremal ellipse","feed_subtitle":"New proof computes polytope extremal functions from simplices and strips, settling the ellipse foliation.","key_machinery":"The load-bearing identity is $V_K(z)=\\max\\{V_S(z):S\\in\\mathcal S(K)\\}$ (Theorem 5.4), where $\\mathcal S(K)$ is the finite collection of supporting simplices and strips whose intersection is $K$. For a simplex, the extremal function is $V_S(z)=\\log h(|\\lambda_0(z)|+\\cdots+|\\lambda_d(z)|)$, with $\\lambda_j$ the barycentric coordinates and $h(\\eta)=\\eta+\\sqrt{\\eta^2-1}$ the inverse Joukowski map. Strips are handled by the reduction $V_{K\\times \\mathbb R^{d-j}}(z',z'')=V_K(z')$, proved by a ball-limit argument. The ellipse side is carried by the Robin exponential map $R_K(c\\zeta)=a(c)+c\\zeta+c/\\zeta$, where $a(c)$ is the center of the unique inscribed ellipse of shape $c$; for the ball this is a generalized Joukowski map, and for the simplex it is obtained from the ball case by the square map.","core_discovery":"The central claim is that the extremal function $V_K$ of a compact convex polytope is fully determined by supporting simplices and strips: $V_K(z)=\\max\\{V_S(z): S\\in\\mathcal S(K)\\}$, where the sets in $\\mathcal S(K)$ are finitely many simplices and strips whose common intersection is $K$. For a simplex, $V_S$ has the closed form $\\log h(|\\lambda_0(z)|+\\cdots+|\\lambda_d(z)|)$ in terms of barycentric coordinates and the inverse Joukowski function; for a strip, $V_S$ reduces to the extremal function of a lower-dimensional cross-section. From this the paper constructs, first for the ball, then for a simplex, then for a polytope, a foliation of $\\mathbb C^d\\setminus K$ by complexified ellipses on which $V_K=\\log|\\zeta|$, and obtains the general convex-body case by decreasing polytopal approximation.","pith_inferences":["Editorial inference: the construction is algorithmic in nature: given the half-space description of a polytope, the list of supporting simplices and strips is produced by rank checks on normal vectors, so $V_K$ and the equilibrium measure could be evaluated symbolically for any polytope.","Editorial inference: the strip factorisation suggests that when face normals are linearly dependent, the computation genuinely lowers dimension; the polytope problem projects to a lower-dimensional simplex problem, which may explain why strips require no separate analytic machinery beyond the simplex formula.","Editorial inference: the identity $\\log|\\zeta|=V_K(a+c\\zeta+c/\\zeta)$ is concrete enough to be a numerical check of the theory for arbitrary bodies: choose a body, choose an exterior point, solve for the ellipse data, and compare the two sides by polynomial approximation."],"forward_implications":["For every compact convex polytope, the extremal function is explicitly computable as a maximum of barycentric-coordinate formulas over a finite list of supporting simplices and strips.","For polytopes satisfying the face-normal independence condition, the Robin exponential map is a homeomorphism from $\\mathbb C^d\\setminus K_\\rho$ to $\\mathbb C^d\\setminus K$ whose leaves are complexified ellipses on which $V_K=\\log|\\zeta|$.","For any compact convex body, each exterior point lies on a complex ellipse whose real part is an inscribed real ellipse and on which the extremal function is exactly $\\log|\\zeta|$.","The paper notes that known results on the regularity of $V_K$ and on the complex equilibrium measure $(dd^c V_K)^d$ follow from this self-contained theorem."],"supporting_citations":[{"why":"Supplies the closed-form extremal function of a real simplex via barycentric coordinates and the inverse Joukowski function.","marker":"[4]"},{"why":"Supplies the explicit extremal function of the real unit ball, which is the starting point for the Hooke-ellipse foliation and for the ball-limit argument in the strip formula.","marker":"[14]"},{"why":"States the theorem on extremal ellipses that this paper reproves, providing the result being given a new self-contained derivation.","marker":"[5]"},{"why":"Supplies the geodesic-based framework used by the earlier proof of existence of extremal ellipses, which this paper bypasses.","marker":"[10]"},{"why":"Supplies further geodesic and holomorphic-retract results used in the earlier indirect route to extremal ellipses.","marker":"[11]"},{"why":"Supplies the symmetry and Monge-Ampere transformation results used by the earlier existence proof.","marker":"[12]"},{"why":"Provides the standard extremal-function properties, including continuity, monotone limits for decreasing compact sets, and the product formula, used throughout the polytope and approximation arguments.","marker":"[9]"}],"fun_headline_variants":["Ellipse foliation proven for every convex body","New proof: extremal functions from simplices and strips","Simplices and strips determine extremal functions","How simplices and strips yield extremal ellipses","Extremal ellipses: new proof for all convex bodies"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof assumes that the extremal functions of the approximating polytopes converge, point by point, to the extremal function of the convex body they shrink to, and that the corresponding inscribed ellipses have a locally uniform limit; the entire convex-body theorem rests on that convergence.","fun_headline_variants_meta":{"raw":{"variants":["Ellipse foliation proven for every convex body","New proof: extremal functions from simplices and strips","Simplices and strips determine extremal functions","How simplices and strips yield extremal ellipses","Extremal ellipses: new proof for all convex bodies"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000949,"raw_usage":{"total_tokens":3968,"prompt_tokens":782,"completion_tokens":3186,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":398,"completion_tokens_details":{"reasoning_tokens":3110}},"tokens_in":398,"tokens_out":3186,"duration_ms":21496,"temperature":1.0,"reasoning_tokens":3110,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:24:10.540028+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For a concrete test, take the unit square in $\\mathbb R^2$, choose an exterior point such as $z=(2+i,3)$, solve the equations $z=a+c\\zeta+c/\\zeta$ numerically for $a\\in\\mathbb R^2$, $c\\in\\mathbb C^2$, and $|\\zeta|>1$, and compare $\\log|\\zeta|$ with a high-degree polynomial-envelope approximation of $V_K(a+c\\zeta+c/\\zeta)$. A nonzero difference for every such solution would falsify the theorem.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the closed-form extremal function of a real simplex via barycentric coordinates and the inverse Joukowski function."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the explicit extremal function of the real unit ball, which is the starting point for the Hooke-ellipse foliation and for the ball-limit argument in the strip formula."},{"cited_title":"Burns, N","cited_arxiv_id":null,"evidence_quote":"States the theorem on extremal ellipses that this paper reproves, providing the result being given a new self-contained derivation."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the geodesic-based framework used by the earlier proof of existence of extremal ellipses, which this paper bypasses."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies further geodesic and holomorphic-retract results used in the earlier indirect route to extremal ellipses."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the symmetry and Monge-Ampere transformation results used by the earlier existence proof."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the standard extremal-function properties, including continuity, monotone limits for decreasing compact sets, and the product formula, used throughout the polytope and approximation arguments."}],"review_version":1}