{"id":"79f5e639-f281-41ae-99a8-49b9a921ed45","arxiv_id":"2501.12347","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A domain in R^3 whose Yamabe quotient is close to the maximal ball value is close to a ball: diffeomorphic, nearly round, and Gromov-Hausdorff close after scaling.","lead":"This paper proves a stability version of Escobar's rigidity theorem: in three dimensions, a bounded domain whose Yamabe quotient is nearly that of a ball must actually be nearly a ball in several precise geometric senses. The result gives quantitative teeth to Escobar's suggestion that the Yamabe quotient measures how far a domain is from being round.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Stability proof hinges on unproved Lemma 6 of [33] asserting every regular level set Σ_t={w=t} of w=-log u is connected when ∂Ω is connected; this premise is used in Proposition 22 and the deficit estimate Proposition 24, so a gap here would invalidate the main theorems.","rationale":"The reader's weakest_assumption identifies exactly the unproved connectedness of regular level sets, and the reader's verdict is already CONDITIONAL. My read does not change that verdict: the concern is real and load-bearing, but it is likely repairable by supplying the short minimum-principle proof. The separate gap in the case split of Theorem 1 (negation of Case 1 only yields d(x,∂Ω) ≤ ε/2, while Case 2 assumes ε/4) is also present but is a minor, fixable issue. The central unresolved issue remains the missing justification for the topological premise that underlies the stability machinery.","tokens_in":23099,"tokens_out":33718,"duration_ms":325214,"concrete_test":"Re-derive Lemma 6 of [33] via the minimum principle: for s ∈ (0,1), extend u by 1 on Ω and set E_s = {u > s}; show the complement has no bounded components (otherwise a harmonic function would be constant on a nonempty open set), hence the boundary of E_s, which is Σ_s, is connected. If this derivation cannot be completed, numerically compute the capacitary potential for a dumbbell or solid-torus domain and check whether every regular level set is connected; a disconnected regular level set would falsify the premise and break Proposition 24.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 3.1 states: 'The assumption that ∂Ω is connected ensures that Σ_t is connected for all regular values t (see [33, Lemma 6]).' No proof is given; it is cited to the authors' own preprint. This connectedness is load-bearing: Proposition 22 bounds U'(t) using the Gauss-Bonnet formula and the Willmore energy estimate for closed connected surfaces, and the derivation of the deficit term ∫∥˚A∥² (Proposition 24) -- the engine for Propositions 28/29 and all main theorems -- depends on it. If Σ_t had k components, Gauss-Bonnet would give ∫K ≤ 4πk rather than ≤ 4π, and the constant 8π in Proposition 22 would not follow, so the quantitative stability estimate would fail. The lemma is likely true: the strong minimum principle applied to u in a bounded component of {u ≤ s} forces u ≡ s, impossible; hence {u ≤ s} has no bounded components and Σ_s is connected. But the paper does not provide this proof, instead citing an unpublished preprint. A revision should include the argument or a direct, accessible reference.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies quantitative stability for Escobar's rigidity theorem in R3: for smooth bounded domains with connected boundary, the Yamabe quotient Q(Ω) is maximized by balls, and equality holds only for balls. The authors prove that if Q(B)−Q(Ω) is sufficiently small then Ω is Hausdorff-close to a ball (Theorem 1), GH-close to the unit ball in its induced length metric (Theorem 2), and diffeomorphic to a ball (Theorem 3), with the quantitative deficit δ=O(ε^9) recorded in Remark 5. The proof uses the capacitary potential of the exterior domain, a test function built from it, Miao's monotonicity formulas, the De Lellis–Müller almost-umbilic estimate, and medial-axis/flow arguments to convert a nearly round level set into containment of Ω in a slightly larger ball. A final section draws qualitative consequences for Gehring's coefficient of quasi-conformality in the presence of spikes, ridges, and hairs.","tokens_in":23281,"tokens_out":16424,"duration_ms":155522,"significance":"If correct, these results constitute a substantial advance: they give the first quantitative stability for the Euclidean Yamabe quotient of domains, confirming Escobar's heuristic that Q(Ω) measures distance from a ball, and they connect this invariant with quasiconformal geometry. The proof strategy is well suited to the problem: the test function is constructed from the capacitary potential without fitted parameters, the deficit is expressed through an explicit ∫|Å|² term, and the dependence on the De Lellis–Müller constant is stated honestly. The medial-axis arguments are original in this context and give a concrete route from a nearly umbilical level set to global containment. However, the proof of the main theorem contains a repairable gap in the case split, and a load-bearing connectivity lemma is cited to an unpublished preprint rather than proved; these issues must be fixed before the results can be considered established.","major_comments":[{"comment":"The case split in the proof of Theorem 1 is not exhaustive. Case 1 assumes there is x∈Ω with |x|≥1 and B(x,ε/2)⊂Ω, while Case 2 assumes that every x∈Ω with |x|≥1 satisfies d(x,∂Ω)≤ε/4. A point with ε/4<d(x,∂Ω)<ε/2 satisfies neither condition, and such points are not a priori excluded by the standing assumptions, including Proposition 35. The subsequent flow argument in Case 2 only needs the bound R≤ε/2: since R(C(ε,y))−R(y)≤ε/2, the average of ∥∇∥² over [0,ε] is at most 1/2, giving a time with ∥∇∥²≤1/2. Thus the gap is repairable by replacing ε/4 by ε/2 in Case 2, but as written the contradiction is not established for all configurations. This same gap propagates to Theorem 3, whose proof states that 'Case 1' gives R(x)≤ε/2 for x∈Ω\\B(0,1); that statement belongs to the repaired Case 2.","section":"Section 4, proof of Theorem 1"},{"comment":"The paper asserts without proof that connectedness of ∂Ω implies connectedness of every regular level set Σ_t={w=t}, citing [33, Lemma 6], an unpublished preprint by the same authors. This fact is load-bearing: Proposition 22 applies Gauss–Bonnet and the Willmore bound ∫H²≥16π for closed connected surfaces; if Σ_t had k components, the bounds would become 4πk and 16πk and the constant 8π in Proposition 22 would fail. The lemma is true by a standard strong-minimum-principle argument applied to the capacitary potential u in a bounded component of {u≤t}, but the manuscript should include that proof or cite a published, accessible reference instead of relying on an unpublished preprint.","section":"Section 3.1, first paragraph and Proposition 22"}],"minor_comments":[{"comment":"The interval for s is misstated: for t∈[−log(1−η/2), −log(1−η)], one has s=e^{−t}∈[1−η, 1−η/2], not s∈[η/2, η]; the displayed interval contradicts the statement '1−η<s<1' in the proposition.","section":"Section 3.3, Proposition 28"},{"comment":"The phrase 'By Case 1 of the proof of Theorem 1' is incorrect; the needed fact, that R(x)≤ε/2 for every x∈Ω\\B(0,1), comes from Case 2 (after the repair described above).","section":"Section 4, proof of Theorem 3"},{"comment":"The comparison with the coefficient of quasi-conformality is formulated for nonsmooth features such as spikes and ridges, whereas Q is only defined for smooth domains; the parenthetical smoothing argument should be made quantitative, since the smoothing scale affects the stated lower bound δ(1/10,θ(α)).","section":"Section 5"},{"comment":"There are several typographical slips, including 'quasi-confromality' in Section 1.1 and 'Haussdorff' in Definition 36; these should be corrected in the final version.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The main results are plausible and the proof strategy is promising. The two issues identified—the non-exhaustive case split in Theorem 1 and the reliance on an unpublished preprint for a load-bearing connectivity lemma—are both repairable, so I am not recommending rejection. However, the connectivity lemma should be proved directly or cited to a published source, and the proof of Theorem 1 should be rewritten with an exhaustive dichotomy. The misstated interval in Proposition 28 and the wrong cross-reference in Theorem 3 also indicate that a careful proofreading pass is needed."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The headline result is real: for a smooth bounded domain Ω ⊂ R³ with connected boundary, having Q(Ω) within δ of Q(B) forces the ball containment B(x,r) ⊂ Ω ⊂ B(x,r(1+ε)), Gromov–Hausdorff closeness to the unit ball under the induced length metric, and diffeomorphism to a ball. Theorems 1–3 are new quantitative stability statements; Escobar's rigidity inequality was known, but nobody had a quantitative version. The Sobolev-quotient analogues and the qualitative comparison with Gehring–Väisälä's quasi-conformality coefficient are useful additions that speak to Escobar's own question.\n\nThe method is the best part. The capacitary potential plus Miao's monotonicity formulas give a compact re-derivation of Escobar's inequality, and the deficit estimate in Proposition 24 — turning monotonicity slack into an L² bound on trace-free curvature of level sets — is the right engine. The De Lellis–Müller almost-umbilic step and the capacity-to-proximity lemma (Proposition 27) are clean. The final act, using Lieutier's medial-axis flow to find a smaller certificate ball when Ω protrudes past B(0,1+ε), is clever and is the most original part of the paper.\n\nSoft spots, in proportion. The serious one: connectedness of every regular level set of the capacitary potential is load-bearing — it is what lets Proposition 22 use Gauss–Bonnet and the Willmore bound for closed connected surfaces, and the 8π constant feeds the deficit estimate. It is cited to the authors' own unpublished preprint [33, Lemma 6], with no proof. The lemma is true; a bounded component of {u ≤ s} would force u ≡ s by the maximum principle, and connectedness of ∂Ω forces {u > s} to be connected. But a load-bearing lemma needs its half-page proof in the paper, or an accessible reference, not a self-citation to an arXiv preprint.\n\nThe second issue: the case split in Theorem 1 is not exhaustive as written. Case 2 assumes d(x,∂Ω) ≤ ε/4, while the negation of Case 1 gives only the bound ε/2. This is cosmetic — the flow argument in Case 2 runs verbatim with ε/2, using R(y) > 0 for the strict contradiction — but as printed it is a logical hole. Smaller still: the equality case in Proposition 20 is asserted in the introduction's sketch without proof (harmless, since the stability theorems do not need it), and the proof of Theorem 3 is compressed, especially the flow-homotopy claim for starting points in B(0,3/4).\n\nWho this is for: geometric analysts working on Yamabe invariants, stability questions, or level-set methods. It deserves a serious referee; the core arguments are sound and both flagged issues are repairable. I would send it out, with a request that the revision prove the connectedness lemma and fix the case split.","headline":"New quantitative stability theorems for Escobar's Yamabe inequality in R³, built with capacitary level sets and a medial-axis flow; the core holds up and the two main gaps are repairable.","tokens_in":23864,"tokens_out":30013,"would_cite":true,"duration_ms":267391,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["53C21","53C24"],"pacs":[],"model":"deepseek-v4-flash","headline":"If the Yamabe quotient of a domain in R3 is nearly maximal, the domain is nearly a ball.","keywords":["Yamabe quotient","conformal invariant","stability of rigidity","capacitary potential","almost-umbilic estimate","medial axis","Gromov-Hausdorff distance","quasi-conformality coefficient"],"falsifier":"Find one smooth bounded domain with connected boundary whose Yamabe quotient is closer than δ to Q(B) but which contains a long thin hair or neck, so that no ball satisfies B(x,r) ⊂ Ω ⊂ B(x,r(1+ε)) for a fixed small ε; the theorem predicts this cannot happen. A more targeted check is to compute the capacitary potential of a solid torus and inspect its regular level sets: a disconnected regular level set would invalidate the level-set connectedness assumption without changing the hypothesis on ∂Ω.","tokens_in":22833,"feed_emoji":"🔵","tokens_out":10821,"duration_ms":106166,"temperature":0.7,"pith_summary":"The paper proves a quantitative stability theorem for the sharp conformal inequality Q(Ω) ≤ Q(B): if a smooth, bounded domain Ω in R3 with connected boundary has Yamabe quotient within δ of the ball's Yamabe quotient, then Ω is close to a ball in a precise sense. Concretely, for every ε>0 there is δ>0 such that Q(B)-Q(Ω)<δ implies B(x,r) ⊂ Ω ⊂ B(x,r(1+ε)), and after scaling Ω is Gromov-Hausdorff ε-close to the unit ball with its induced length metric; for δ small enough Ω is diffeomorphic to a ball. The proof works through level sets of the capacitary potential of the exterior domain, converting the deficit Q(B)-Q(Ω) into a weighted integral of the trace-free second fundamental form of those level sets, then applying an almost-umbilic estimate and a medial-axis flow argument. The same results hold for the boundary Sobolev quotient, and the paper shows that geometric features known to make the quasi-conformal coefficient large—spikes, ridges, and hairs—force a definite Yamabe deficit as well.","feed_headline":"Near-maximal Yamabe quotient forces near-roundness","feed_subtitle":"Tiny conformal deficit forces a ball sandwich, length-metric closeness, and diffeomorphism to a ball.","key_machinery":"The engine is the capacitary potential u of the exterior domain: Δu = 0 outside Ω, u = 1 on ∂Ω, u → 0 at infinity, and its logarithm w = −log u, whose level sets Σ_t = {w = t} foliate the exterior. The paper computes two level-set quantities, W(t) = ∫_{Σ_t} |∇w|^2 and U(t) = ∫_{Σ_t} H|∇w|, and uses monotonicity formulas from [34] to get ODE-type inequalities W′(t) = 2W(t) − U(t) and U′(t) ≤ 8π − U(t) + W(t) − ⋯, with equality for balls. A carefully chosen test function f = s(w)|∇w|^{1/2} with s(t) = $e^{{t/2}}$(1 + $e^{{2t}}$)^{-1/2} is then substituted into the Yamabe quotient of the exterior domain; the monotonicity formulas show its quotient is at most Q*(B), and tracking the losses yields a weighted bound ∫ b(τ) ∫_{Σ_τ} ∥Å∥^2 da dτ ≤ Q*(B) − Q*(Ω). A small deficit therefore produces a level set with small total trace-free second fundamental form; the almost-umbilic estimate [15] turns that into closeness to a round sphere, and a capacity comparison (Proposition 27) propagates roundness to ∂Ω. For the interior-domain results, conformal inversion and the medial-axis flow [29] are used to select a smallest ball B(x,r) ⊂ Ω whose boundary touches ∂Ω in two separated directions; the structural proposition applies to every such ball, and the flow pushes out a contradiction unless Ω already lies in B(x, r(1+ε)).","core_discovery":"The paper's central claim is a stability theorem: the sharp inequality Q(Ω) ≤ Q(B)—where Q is the Yamabe quotient of a smooth bounded Euclidean domain and B is the unit ball, with equality exactly for balls—is quantitatively stable in R3. If Q(B)-Q(Ω) is smaller than δ, then Theorem 1 produces a ball B(x,r) with B(x,r) ⊂ Ω ⊂ B(x,r(1+ε)); Theorem 2 says that after translation and scaling the domain, equipped with its induced length metric, is Gromov-Hausdorff ε-close to the unit ball; Theorem 3 says a sufficiently small deficit forces Ω to be diffeomorphic to a ball. The authors compute the dependence δ = O($ε^{9}$) with a constant that is not explicit, and they transfer all three theorems to the Sobolev quotient Q(Ω, ∂Ω). The final section establishes a qualitative comparison with the coefficient of quasi-conformality K: spikes, ridges, and hairs, which are known to force K away from 1 by definite amounts [22], also force Q(B)-Q(Ω) to be bounded away from 0 by definite amounts.","pith_inferences":["Editorial inference: the rate δ = O(ε^9), with a constant inherited from the almost-umbilic estimate [15], is probably not sharp; testing whether the exponent 9 can be lowered is a concrete open problem.","Editorial inference: all machinery is three-dimensional because the Gauss-Bonnet and Willmore controls act on closed surfaces in R3, so the theorem should not be expected to generalize verbatim to domains in R4 or higher.","Editorial inference: the spike/ridge/hair comparison suggests there may be a direct quantitative inequality between Q(B) − Q(Ω) and K(Ω) − 1 near the round ball; the paper does not state such an inequality, but its structural proposition is a natural tool for trying to prove one."],"forward_implications":["If Q(Ω) is within δ of Q(B), then for the same ε there is a two-sided ball inclusion with radii ratio 1+ε, so the deficit controls Hausdorff distance from a round ball in a quantitative way.","With the induced length metric, a nearly maximal quotient forces the domain to be Gromov-Hausdorff close to the unit ball, so geodesic distances inside Ω are nearly Euclidean at the scale of the domain.","A uniform deficit threshold guarantees Ω is diffeomorphic to a ball, so the stability regime is topologically trivial and the boundary is a smoothly embedded sphere.","The same stability holds for the Sobolev quotient Q(Ω, ∂Ω), so the result is not an artifact of the particular boundary term in the Yamabe functional.","Spikes, ridges, and hairs each force a definite gap Q(B) − Q(Ω) > δ > 0, matching the known quantitative lower bounds for the coefficient of quasi-conformality."],"supporting_citations":[{"why":"Establishes the sharp inequality Q(Ω) ≤ Q(B) with equality only for balls, the rigidity result being stabilized here.","marker":"[20]"},{"why":"Supplies the level-set monotonicity formulas for W(t) and U(t) used to bound the test-function quotient and to control ∥Å∥².","marker":"[34]"},{"why":"Provides the almost-umbilic estimate that turns small total trace-free second fundamental form of a level surface into closeness to a round sphere.","marker":"[15]"},{"why":"Gives the medial-axis flow used to prove that a smallest ball in A(Ω) forces the two-sided inclusion in Theorem 1.","marker":"[29]"},{"why":"Supplies the quantitative lower bounds for the coefficient of quasi-conformality of spikes, ridges, and hairs that the final comparison section mirrors for Q.","marker":"[22]"},{"why":"Justifies absolute continuity of U(t) through Green's-function monotonicity, allowing the integration by parts that produces the deficit estimate.","marker":"[4]"},{"why":"Supplies the cited lemma that connectedness of ∂Ω implies connectedness of regular level sets of w, on which the Gauss-Bonnet and Willmore step rests.","marker":"[33]"}],"fun_headline_variants":["Almost-maximal Yamabe quotient forces near-round domains","Yamabe equality almost forces ball sandwich","Near equality in Yamabe gives ball and diffeomorphism","Small Yamabe deficit implies domain is nearly a ball","Yamabe stability: near-roundness from near-max quotient"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument depends on the claim, cited to the authors' own prior preprint [33, Lemma 6] and not proved here, that when ∂Ω is connected every regular level set of w = −log u in the exterior domain is connected; Proposition 22's Gauss-Bonnet and Willmore estimates need closed connected level surfaces.","fun_headline_variants_meta":{"raw":{"variants":["Almost-maximal Yamabe quotient forces near-round domains","Yamabe equality almost forces ball sandwich","Near equality in Yamabe gives ball and diffeomorphism","Small Yamabe deficit implies domain is nearly a ball","Yamabe stability: near-roundness from near-max quotient"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000214,"raw_usage":{"total_tokens":1431,"prompt_tokens":954,"completion_tokens":477,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":570,"completion_tokens_details":{"reasoning_tokens":400}},"tokens_in":570,"tokens_out":477,"duration_ms":5349,"temperature":1.0,"reasoning_tokens":400,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T17:15:33.025276+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find one smooth bounded domain with connected boundary whose Yamabe quotient is closer than δ to Q(B) but which contains a long thin hair or neck, so that no ball satisfies B(x,r) ⊂ Ω ⊂ B(x,r(1+ε)) for a fixed small ε; the theorem predicts this cannot happen. A more targeted check is to compute the capacitary potential of a solid torus and inspect its regular level sets: a disconnected regular level set would invalidate the level-set connectedness assumption without changing the hypothesis on ∂Ω.","supporting_citations":[{"cited_title":"The yamabe problem on manifolds with boundary.Journal of Differential Geometry , 35(1):21–84, 1992","cited_arxiv_id":null,"evidence_quote":"Establishes the sharp inequality Q(Ω) ≤ Q(B) with equality only for balls, the rigidity result being stabilized here."},{"cited_title":"Mass, capacitary functions, and the mass-to-capacity ratio","cited_arxiv_id":null,"evidence_quote":"Supplies the level-set monotonicity formulas for W(t) and U(t) used to bound the test-function quotient and to control ∥Å∥²."},{"cited_title":"Optimal rigidity estimates for nearly umbilical surfaces.Journal of Differential Geometry, 69(1):075–110, 2005","cited_arxiv_id":null,"evidence_quote":"Provides the almost-umbilic estimate that turns small total trace-free second fundamental form of a level surface into closeness to a round sphere."},{"cited_title":"Any open bounded subset of Rn has the same homotopy type as its medial axis","cited_arxiv_id":null,"evidence_quote":"Gives the medial-axis flow used to prove that a smallest ball in A(Ω) forces the two-sided inclusion in Theorem 1."},{"cited_title":"The coefficients of quasiconformality of domains in space.Acta Mathematica, 114(1):1– 70, 1965","cited_arxiv_id":null,"evidence_quote":"Supplies the quantitative lower bounds for the coefficient of quasi-conformality of spikes, ridges, and hairs that the final comparison section mirrors for Q."},{"cited_title":"A green’s function proof of the positive mass theorem","cited_arxiv_id":null,"evidence_quote":"Justifies absolute continuity of U(t) through Green's-function monotonicity, allowing the integration by parts that produces the deficit estimate."}],"review_version":1}