{"id":"080a913d-9464-496d-bb87-9b2f36b1948d","arxiv_id":"1908.03802","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"A framework that is not rigid but is close to prestress-stable is proven to be confined to a small ball, with explicit radii computable by semidefinite programming.","lead":"This paper introduces almost-rigidity, a quantitative way to say that a flexible framework of fixed-length bars is nevertheless trapped in a small region of space. It proves computer-checkable conditions with explicit bounds on how far a framework can flex and how much edge lengths must change to escape.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition III is stated with the wrong threshold: (5.39) requires <3/8, but Strong(2/9) needs <1/3; a cubic energy with ratio 0.3 satisfies (5.39) yet violates the conclusion. Theorem III's proof chain relies on this proposition and needs correction.","rationale":"The reader's weakest_assumption focused on the restrictiveness of the smallness conditions D<1/2 and Dpss<1/2, and on the lack of code/data. That is a reasonable practical concern, but it is not the most load-bearing issue for the mathematical central claim. I found a concrete, localized error in the proof structure: Proposition III is stated with a condition that is weaker than the condition its proof actually uses. Concretely, the displayed bound in (5.39) is <3/8, whereas Strong(2/9), which Lemma 5.7 needs, corresponds to <1/3. A cubic counterexample with ratio 0.3 satisfies the displayed hypothesis but violates the conclusion. This does not necessarily falsify Theorem I or Theorem III, because the theorems' explicit hypotheses (e.g., (3.12)) pass through Corollary 5.12's stronger condition, which does imply the needed R<1/3. But the paper as written contains a false proposition that is cited in the proof chain, so the proof is not currently sound. The fix is small (change 3/8 to 1/3 in Proposition III, or state Strong(2/9) in derivative form consistently) and the main conclusions should survive; hence the reader's CONDITIONAL verdict remains appropriate. My disagreement with the reader is about where the true weak point lies: not primarily in the restrictiveness of the conditions, but in this internal proof inconsistency.","tokens_in":46834,"tokens_out":28406,"duration_ms":287386,"concrete_test":"Substitute the definitions a=H‴/6, b=H″/2, c=H′ into (5.35) and (5.39). Compute (max|a|/b)(max|c|/b) = (2/3)(max|H′|/H″)(max|H‴|/H″). For Strong(2/9) this product must be <2/9, hence the displayed product in (5.39) must be <1/3, not <3/8. To settle whether the printed condition is sufficient, test H(t)=t³+t²+0.3t: it satisfies R=0.3<3/8, but η1*=2|c|/b=0.6 and η3*=2b/(3|a|)=2/3<0.9=(3/2)η1*, so Proposition III's conclusion fails. If the authors correct the threshold to 1/3 and confirm that the proof of Corollary 5.12 and Proposition 7.4 still implies it, Theorem III stands; otherwise the proof is incomplete.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In Section 5.3, Proposition III states its hypothesis as Strong(2/9, 3/2η1*) but displays the bound (max |H′|/H″)(max |H‴|/H″) < 3/8. Under the paper's own definitions, a=H‴/6, b=H″/2, c=H′, so (max |a|/b)(max |c|/b) = (2/3)(max|H′|/H″)(max|H‴|/H″). Strong(C,r) is defined by (max|a|/b)(max|c|/b)2/9; then η1*=0.6 and η3*=2/3<0.9=(3/2)η1*, so the asserted energy-barrier radius does not exist. Since Corollary 5.12 invokes Proposition III and Theorem III is proved through Proposition 7.4/Corollary 5.12, the written proof chain for Theorem III contains a real gap. The intended theorem may still be correct—Corollary 5.12's stronger condition η1/η0(3η1/2)<8/9 does imply R<1/3—but the manuscript as written needs a corrected constant and a re-verification of the chain.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper introduces and proves a quantitative theory of \"almost-rigidity\" for bar frameworks. Given a framework (p,E), a subspace C complementary to the trivial motions, an almost-flex space V, an almost-stress ω whose stress matrix is positive definite on V, and derived quantities L, μ̄, η1, the authors state four theorems: (I) every continuous edge-length-preserving path in C_p starting at p remains within distance η1 of p; (II) any other configuration with the same edge lengths is either in the η1-ball or at distance at least η2>η1; (III) a deformation reaching distance η>(3/2)η1 forces a minimal squared-edge-length change e_min(η), hence any path to a distant equivalent configuration crosses a barrier e*_min; and (IV) under a stronger smallness condition there is a nearby prestress-stable framework. The proofs construct a spring energy H, prove general barrier propositions for functions that are almost critical (Section 5), specialize the derivative bounds to framework energies (Section 6), and then apply them (Section 7). The paper also reports SDP-based numerical tests on many examples and on 98,529 sphere clusters.","tokens_in":47288,"tokens_out":24243,"duration_ms":213401,"significance":"If the proof chain is repaired, this is a valuable contribution. It gives explicit, analytically defined radii and edge-length barriers that can be certified by linear algebra and semidefinite programming, without assuming that small singular values of the rigidity matrix are perturbed zeros. The framework is genuinely certificate-based: given V, W, ω satisfying (3.3), each theorem's conclusion follows from arithmetic verification of smallness conditions, and the examples and 98,529-cluster survey directly show how often the hypotheses hold and where they fail. The paper is also honest about limitations: the conditions are restrictive, one cluster fails in the survey, and Example 1 applies only for tiny perturbations. No parameter fitting or external ground truth is used in the estimates. However, the manuscript currently contains a load-bearing constant error in Proposition III and an overstatement in Theorem IV, so the central claims are defensible but need revision.","major_comments":[{"comment":"The hypothesis labeled Strong(2/9, (3/2)η1*) is stated with the wrong threshold. From the definitions in (5.9) and (5.35), a=H'''/6, b=H''/2, c=H', so (max |a|/b)(max |c|/b) = (2/3)(max |H'|/H'')(max |H'''|/H''). Strong(2/9) therefore requires the displayed product to be < (3/2)(2/9) = 1/3, not <3/8; 3/8 is the threshold for Strong(1/4). As written, Proposition III is false: for H(t)=t³+t²+0.24t one has η1*=0.48 and η3*=2/3<0.72=(3/2)η1*, while the displayed product equals 0.36<3/8 (using |H'(0)|/H''(0)=0.12 and max|H'''|/H''(0)=3), so the asserted positivity interval is empty. Because Theorem III is proved through Proposition 7.4 and Corollary 5.12, which invokes Proposition III, the proof chain for Theorem III has a real gap. The intended theorem is recoverable by correcting the constant to 1/3 and re-verifying the chain; note that Corollary 5.12's condition η1/η0(3η1/2)<8/9 is the correct Strong(2/9) analogue.","section":"Section 5.3, Eq. (5.39)"},{"comment":"Theorem IV as stated asserts existence of (ppss,E)∈C_p with the same edge lengths as p, but the proof only shows that the unconstrained energy H achieves an interior minimum ppss with ∇H(ppss)=0 and positive definite Hessian on C. A critical point of H is not forced to satisfy e(ppss)=e(p): for each edge energy h_i(x)=½κ(x-p_i²)²+ω_i x, the equation h_i(x)=h_i(p_i²) has a second root x=p_i²-2ω_i/κ, and the proof does not even show H(ppss)=H(p). Thus the \"same edge lengths as p\" clause is unsupported; it is also not used in the applications (Section 4.1.1 explicitly says \"we don't know that ppss has the desired edge lengths\"). The theorem should be restated with the weaker conclusion \"there is a nearby prestress-stable framework\" (matching the abstract and introduction), or an additional argument must be supplied. The phrase \"since the energy function is quadratic\" in the proof is also inaccurate: the energy in (6.2) is quartic in q.","section":"Section 3 (Theorem IV) and Section 7 proof"}],"minor_comments":[{"comment":"The sentence \"the the K3,4 framework\" contains a duplicated article.","section":"Section 4.2.4"},{"comment":"The phrase \"using the spectral data associated associated with\" contains a duplicated word.","section":"Section 9"},{"comment":"The phrase \"This is a slightly stronger than condition (5.14)\" is grammatically incomplete, and the \"constant of 2/8\" should be written as 1/4 to avoid confusion with the adjacent 2/9 constant.","section":"Section 5.2, after Eq. (5.22)"}],"recommendation":"major_revision","confidential_remarks":"The paper is a strong candidate after revision. Please ensure the authors correct the 3/8 versus 1/3 mismatch in Section 5.3 and either weaken Theorem IV or prove the same-edge-length claim. The rest of the proof architecture is coherent, and the numerical experiments are informative. I would also ask them to scan Section 5.3 for other places where H-derivative units and a,b,c units are interchanged."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear —,\n\nThis paper deserves serious refereeing. The core idea is new and useful: a quantitative, nonlocal variant of prestress stability, giving explicit radii and edge-change barriers via SDP-tested certificates. Theorems I–IV are real results, not restatements; the energy-barrier technique and the extension of the second-derivative test to almost-critical points are original and likely to be reused. The examples are well chosen, and the authors honestly report on a large dataset of 98,529 sphere clusters, including one cluster where the main condition fails.\n\nThe soft spot is a genuine error in Section 5.3. Proposition III announces the assumption Strong(2/9) but displays the bound (max|H'|/H'')(max|H'''|/H'') < 3/8. Under the paper's own definitions in (5.35), Strong(2/9) requires that product < 1/3; the displayed < 3/8 corresponds only to Strong(1/4). This is not a harmless typo: a cubic H(t)=t^3+t^2+0.3t satisfies the displayed bound yet has product 0.3 > 2/9, and the claimed energy-barrier radius does not exist. Since Theorem III is proved via Proposition 7.4 and Corollary 5.12, which invokes Proposition III, the written proof chain has a gap. I believe the intended theorem is recoverable—Corollary 5.12's actual condition η1/η0(3η1/2)<8/9 does imply the needed constant—but the manuscript must be corrected and the chain re-verified. Minor concerns: no code or data accompany the SDP computations, which makes the certificates hard to reproduce; and the smallness conditions are restrictive, as the paper itself notes (Example 1 works only for tiny perturbations, and one cluster in the survey fails).\n\nWho should read this: rigidity theorists, and people working on sphere packings, materials, or molecular geometry. It deserves peer review; I would accept it with major revision, primarily to fix the constant and to make the computational pipeline reproducible. The core contribution is solid enough to justify the time.\n\nBest","headline":"A valuable quantitative extension of prestress stability, with a real but fixable constant-error in Proposition III that needs correction before publication.","tokens_in":47790,"tokens_out":4845,"would_cite":true,"duration_ms":45648,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["52C25"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper gives explicit, semidefinite-programming-checkable conditions under which a flexible framework cannot flex far, and guarantees a nearby rigid configuration.","keywords":["almost-rigidity","framework rigidity","prestress stability","semidefinite programming","rigidity matrix","infinitesimal flex","sphere packing","energy barrier"],"falsifier":"Run numerical path-following or algebraic solving on any framework satisfying Theorem I's inequalities to look for a same-edge-length configuration at distance in $(\\eta_1,\\eta_2)$ from $p$; a single example would refute the theorem. The natural test case is cluster 45601 from the 13-sphere survey, whose computed $D$ exceeds the threshold, or a slight perturbation of it, to see whether it actually flexes beyond $\\eta_1$.","tokens_in":46633,"feed_emoji":"📐","tokens_out":12615,"duration_ms":111478,"temperature":0.7,"pith_summary":"The paper develops a quantitative theory, called almost-rigidity, for frameworks: graphs with vertices in Euclidean space and fixed edge lengths. Its goal is to extract explicit radii from near-critical spectral data: any continuous edge-length-preserving flex stays inside a small ball of radius $\\eta_1$; any other configuration with the same edge lengths is either inside that ball or at least $\\eta_2$ away; moving to a distant configuration forces the squared edge lengths to change by at least $e^*_{\\min}$; and some configuration within $\\eta_1$ is genuinely rigid (prestress stable). These guarantees matter because real and computed frameworks—sphere packs, molecules, engineered structures—are always perturbed, and tiny perturbations can destroy exact rigidity while leaving the structure almost rigid. The conditions are checkable by semidefinite programming, and the paper demonstrates them on a large survey of sphere clusters.","feed_headline":"Flexible frameworks can now be certified to barely move","feed_subtitle":"It gives computable radii and edge-length barriers that turn near-rigid structures into rigorous guarantees.","key_machinery":"The load-bearing object is the artificial energy $H(q)=\\sum_{(i,j)\\in E}\\bigl(\\tfrac12\\kappa(q_{ij}^2-p_{ij}^2)^2+\\omega_{ij}q_{ij}^2\\bigr)$, which models squared edge lengths as springs with stiffness $\\kappa$ carrying tensions $\\omega$. The proof controls three things on $C$: the gradient bound $|H'(p)|\\le2|\\omega^T R(p)|$, the Hessian lower bound $\\tfrac12H''\\ge\\lambda$, and a third-derivative bound written as $\\eta_0(r)=\\tfrac{L}{2}\\bigl(\\bar\\mu^{1/2}+r/L\\bigr)^{-1}$, with $L=(\\lambda/8z\\kappa)^{1/2}$ and $\\bar\\mu=1-\\mu_0/\\lambda$. From those, a generalized second-derivative test for functions that are almost critical produces the discriminant conditions $D<1/2$, (3.12), and $D_{\\rm pss}<1/2$, and the explicit formulas for $\\eta_1,\\eta_2,\\eta_3,e^*_{\\min}$.","core_discovery":"The paper's central discovery is that near-rigidity is certifiable. Starting from a configuration $p$, a subspace $C$ transverse to trivial motions, an almost-flex space $V$, an almost-stress $\\omega$, and parameters $\\lambda,\\kappa$ chosen so that the energy $H(q)=\\sum_{(i,j)\\in E}\\bigl(\\tfrac12\\kappa(q_{ij}^2-p_{ij}^2)^2+\\omega_{ij}q_{ij}^2\\bigr)$ has second derivative at least $2\\lambda$ along unit vectors in $C$, Theorem I proves that if $D=(\\eta_1/L)(\\sqrt{\\bar\\mu}+\\eta_1/L)<1/2$ with $\\eta_1=4|\\omega^T R(p)|/\\lambda$, then every continuous path $q(t)\\in p+C$ preserving edge lengths obeys $|q(t)-p|\\le\\eta_1$. Theorem II gives an outer radius $\\eta_2>\\eta_1$ separating nearby from distant same-edge-length configurations; Theorem III gives a minimum squared-edge-length change $e^*_{\\min}$ along any path to a distant configuration; and Theorem IV, under a stronger $D_{\\rm pss}<1/2$, guarantees a nearby prestress-stable configuration, which is therefore rigid. The paper computes these radii for sphere clusters, isostatic frameworks, Siamese dipyramids, and $K_{3,4}$ examples.","pith_inferences":["Because the conditions are convex, one could optimize over choices of $V$, $W$, and $\\lambda$ to make $\\eta_2$ as large as possible; the paper fixes $\\lambda$ in most examples and reports that $L(\\lambda)$ varies by only about a factor of two, so such tuning could give noticeably larger outer radii.","The edge-change barrier $e^*_{\\min}$ could serve as a rigorous lower bound in molecular or colloidal settings where the question is whether a cluster can reach another geometry without breaking bonds; the paper's energy function already has the form of such interaction potentials, though it does not pursue this application.","Because the certificate failed only for very floppy, high-$\\bar\\mu$ clusters in the 13-sphere survey, a practical screening rule could estimate $\\bar\\mu$ first and only run the full semidefinite test when it is moderate; this is a natural by-product of the paper's constants, not something it states.","One could test how close the constants are to sharp by searching, via numerical continuation, for same-edge-length configurations inside the annulus $(\\eta_1,\\eta_2)$; the Siamese-dipyramid and $K_{3,4}$ examples suggest such close configurations are often absent, so the constants are likely conservative."],"forward_implications":["A framework computed by finite-precision numerical solving can be certified to be nearly rigid: Theorem IV produces a nearby prestress-stable, hence rigid, configuration without assuming that small singular values of the rigidity matrix come from exact zeros.","For first-order rigid frameworks, Corollary 3.4 gives explicit separation constants: any other same-edge-length configuration is at least roughly $0.618\\,\\sigma_0/(2\\sqrt z)$ away, and any path to it must change squared edge lengths by at least roughly $0.528\\,\\sigma_0^2/(2\\sqrt z)$.","The asymptotics in Corollary 3.5 imply that when the exact framework is prestress stable but not first-order rigid, a numerical solution with squared-edge-length error $\\epsilon$ can be only $\\sqrt{\\epsilon}$-close in configuration, so one expects roughly half as many digits of accuracy.","In the survey of 98,529 rigid clusters of 13 unit spheres, all but one satisfy the numerical conditions of Theorems I–II and all but four satisfy the condition of Theorem IV, so the certificates are not empty in the applications that motivated the theory.","With signs on the edges, the same radii extend to tensegrities under the added condition that no edge length changes by more than $2|\\omega_{ij}|/\\kappa$ on any cable or strut."],"supporting_citations":[{"why":"Defines prestress stability and gives the convex characterization plus the existence lemma that the paper generalizes to almost-stresses.","marker":"[8]"},{"why":"Establishes that second-order rigidity implies rigidity, so the isolated energy minimum found in Theorem IV certifies a rigid framework.","marker":"[9]"},{"why":"Establishes the classical fact that first-order rigidity implies rigidity, which underlies the separation radii in Corollary 3.4.","marker":"[3]"},{"why":"Supplies the rigid sphere-packing dataset used for the 10-sphere example and the 98,529-cluster survey.","marker":"[23]"},{"why":"The convex optimization software used to solve the semidefinite programs in the paper's examples.","marker":"[19, 20]"},{"why":"Provides the semidefinite-programming framework behind the efficient tests and the convex programs (2.10) and (3.5).","marker":"[39]"},{"why":"Shows that generic perturbations destroy extra self-stresses, motivating the need for almost-rigidity in the opening examples.","marker":"[41]"}],"fun_headline_variants":["New test certifies frameworks stay almost rigid","Almost-rigidity with computable radii and margins","Flexible frameworks get guaranteed motion bounds","Semidefinite check yields near-rigidity guarantee","Barely move: new framework rigidity proof"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The certificate rests on the computed quantity $D$ (or $D_{\\rm pss}$ for Theorem IV) being below $1/2$ for the chosen stress, subspaces, and $\\lambda$; one of the 98,529 surveyed clusters fails this, so the condition is not automatic.","fun_headline_variants_meta":{"raw":{"variants":["New test certifies frameworks stay almost rigid","Almost-rigidity with computable radii and margins","Flexible frameworks get guaranteed motion bounds","Semidefinite check yields near-rigidity guarantee","Barely move: new framework rigidity proof"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000346,"raw_usage":{"total_tokens":1927,"prompt_tokens":1009,"completion_tokens":918,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":625,"completion_tokens_details":{"reasoning_tokens":848}},"tokens_in":625,"tokens_out":918,"duration_ms":10594,"temperature":1.0,"reasoning_tokens":848,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:02:02.578326+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run numerical path-following or algebraic solving on any framework satisfying Theorem I's inequalities to look for a same-edge-length configuration at distance in $(\\eta_1,\\eta_2)$ from $p$; a single example would refute the theorem. The natural test case is cluster 45601 from the 13-sphere survey, whose computed $D$ exceeds the threshold, or a slight perturbation of it, to see whether it actually flexes beyond $\\eta_1$.","supporting_citations":[{"cited_title":"Second-order rigidity and prestress stability for tensegrity frameworks","cited_arxiv_id":null,"evidence_quote":"Defines prestress stability and gives the convex characterization plus the existence lemma that the paper generalizes to almost-stresses."},{"cited_title":"The rigidity of certain cabled frameworks and the second-order rigidity of arbitrarily triangulated convex surfaces","cited_arxiv_id":null,"evidence_quote":"Establishes that second-order rigidity implies rigidity, so the isolated energy minimum found in Theorem IV certifies a rigid framework."},{"cited_title":"The rigidity of graphs","cited_arxiv_id":null,"evidence_quote":"Establishes the classical fact that first-order rigidity implies rigidity, which underlies the separation radii in Corollary 3.4."},{"cited_title":"Enumerating Rigid Sphere Packings","cited_arxiv_id":null,"evidence_quote":"Supplies the rigid sphere-packing dataset used for the 10-sphere example and the 98,529-cluster survey."},{"cited_title":"White and Walter Whiteley","cited_arxiv_id":null,"evidence_quote":"Shows that generic perturbations destroy extra self-stresses, motivating the need for almost-rigidity in the opening examples."}],"review_version":1}