{"id":"32c148fe-4fa1-4435-b6e0-de6a8e4e4b42","arxiv_id":"2505.19168","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"For every rational resistance c/t, a simple planar graph with O(max(t/c, t/(t-c), log t)) vertices exists, and no graph can do better up to a constant.","lead":"Working out how many vertices a planar simple graph needs to realize a given electrical resistance value, the paper proves a universal upper bound and a matching lower bound, up to a constant. The proof routes the problem through continued fractions and a theorem of Bourgain, giving a sharp answer to the inverse problem for planar graphs.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proof of Proposition 1.2's second lower bound applies Lemma 3.4 to the plane dual G*, which need not be simple; for the n-cycle the asserted intermediate inequality (3.7) is false, leaving the optimality claim unproved as written.","rationale":"I read the paper in good faith and checked the main construction in detail. The upper-bound proof, including the use of Theorem 1.4, the marked sum, and the simplification operation, is intricate but internally consistent; I do not see a flaw there. The most serious problem is in Proposition 1.2, which is needed for the paper's claim to determine the smallest planar graph size up to a constant. The proof of the second lower bound explicitly uses the dual G* as if inequality (3.6), which requires simplicity, applied to it. The dual of a simple planar graph is generally a multigraph, and the n-cycle gives a clean numerical witness that the intermediate inequality asserted in (3.7) is false. This is not a matter of an unspecified constant, and it is not the external Bourgain theorem; it is an internal gap in a central argument. The proposition may well be true, and the final bound likely holds, so I am not recommending rejection. However, the paper as written does not rigorously establish optimality, so ACCEPT should be conditioned on supplying a correct proof, for example by proving and citing the bound ρ(G,e) ≤ (|V|−1)/|V| for non-bridge edges, which would repair the argument. The reader's identified weakest assumption is about Bourgain's Theorem 1.4; that is a reasonable external-input concern, but the concrete internal flaw in (3.7) is more definite and more directly load-bearing.","tokens_in":9299,"tokens_out":42264,"duration_ms":395223,"concrete_test":"Plug the n-cycle into the claimed chain (3.7): |F|=2 and ρ(G*,e)=1/n, so the asserted inequality |F|/2 > 1/(2ρ(G*,e)) becomes 1>n/2, which is false for n>2. Then verify the natural repair—the bound ρ(G,e) ≤ (|V|−1)/|V| for every simple planar marked graph—by exact enumeration of all simple planar graphs up to |V|=9 and by checking the equivalent τ(G−e) ≥ τ(G/e)/(|V|−1). If the repair fails on some example, Proposition 1.2 is false; if it holds, the conclusion survives but the printed proof needs correction.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The upper-bound construction (Theorem 1.5 and Lemmas 3.1–3.3) appears sound. The load-bearing weakness is in Proposition 1.2, the second inequality (3.7). The text applies Lemma 3.4 to the dual graph G*=(F,E) and derives |V| > |F|/2 > 1/(2ρ(G*,e)). The first inequality is Euler's formula, but the second would require the analogue of (3.6) for G*, namely |F| > 1/ρ(G*,e), which is justified only when G* is simple. The dual of a simple planar graph need not be simple: for the n-cycle C_n, |F|=2 and G* is two vertices joined by n parallel edges, with ρ(G*,e)=1/n. Then the asserted inequality |F|/2 > 1/(2ρ(G*,e)) reads 1 > n/2, false for n>2. Thus the displayed derivation of |V| ≥ (1/2)·t/(t−c) is invalid. The claimed inequality itself may be true—for C_n it holds with room to spare—but as written Proposition 1.2, which is essential to the claim that the bound is optimal up to a constant, is not proved. A correct proof needs either a different argument or a justification of why multiedges and loops in G* do not break the estimate.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the inverse problem for effective resistance in planar graphs. For a reduced fraction c/t, Theorem 1.1 constructs a simple planar graph G with an edge e such that ρ(G,e)=c/t and |V| ≤ C max(t/c, t/(t−c), log t), and Proposition 1.2 claims a matching lower bound |V| ≥ C′ max(t/c, t/(t−c), log t). The upper-bound construction (Theorem 1.5) represents the target ratio as a sum of rationals via Bourgain's continued-fraction theorem, builds marked graphs for each summand, and combines them by a marked-sum operation followed by a new simplification operation that preserves the spanning-tree ratio. The lower bound uses commute-time inequalities, planar duality, and Euler's formula.","tokens_in":9563,"tokens_out":13761,"duration_ms":101326,"significance":"If correct, the paper resolves the planar inverse effective-resistance problem up to universal constants, identifying the extremal size as max(1/ρ, 1/(1−ρ), log t). The construction is explicit and parameter-free, and the simplification operation (Lemma 2.4) is a clean technical contribution. The upper-bound proof is modular and easy to verify. However, the proof of the matching lower bound contains a genuine gap, so the optimality claim is not currently established.","major_comments":[{"comment":"The proof of the second inequality in Proposition 1.2 applies Lemma 3.4 and then the first inequality (3.6) to the plane dual G*. Inequality (3.6) relies on the graph being simple, since it uses deg(x), deg(y) < |V|. The dual of a simple planar graph need not be simple; for example, for G = C_n, G* has two vertices and n parallel edges, so deg(f) = n = |F|, and the analogue of (3.6) for G* would assert 2 > n, which is false. Indeed the displayed chain |V| > |F|/2 > 1/(2ρ(G*,e)) fails for n > 2 because 1 < n/2. Thus the lower bound |V| ≥ C′ t/(t−c) is not established as written. Since Proposition 1.2 is the matching lower bound that shows the upper bound is optimal up to a constant, this is a load-bearing gap. The proof must either give a direct argument for the second bound that does not require G* to be simple, or demonstrate that the failure of simplicity (multiple edges and loops) does not invalidate the estimate.","section":"§3.3, Eq. (3.7)"}],"minor_comments":[{"comment":"There is a typographical error in Lemma 3.3: 'simple planar marked graph (G, E)' should be '(G, e)', and the formula for |E∘| is printed as '4(|E|−1)+1' but appears as '4|E|−1)+1' in the text.","section":"§3.3 (Lemma 3.3)"},{"comment":"The phrase 'Theorem 1.5 applied to q1 ← (d/c−d′/c′) and k=1' is informal; the substitution should be written explicitly for clarity.","section":"§3.2 (Lemma 3.2)"},{"comment":"The third inequality uses t ≤ τ(G); this follows from the reduced form of c/t and the integrality of τ(G/e), but the justification is omitted and should be stated.","section":"§3.3 (Proposition 1.2)"},{"comment":"The remark that the constant 1/2 in (3.7) can be removed should be revisited after the simplicity issue with the dual is resolved; as written, the main obstacle is not the constant but the validity of the inequality.","section":"§4.3"}],"recommendation":"major_revision","confidential_remarks":"The gap in Proposition 1.2 is load-bearing and specifically located in Eq. (3.7). The upper-bound construction appears sound, with the standard caveat that the constant in Theorem 1.1 depends on the unspecified constant in Bourgain's Theorem 1.4. The paper is well written and the result, if the lower bound is repaired, would be a strong contribution to extremal graph theory."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Main take: the upper-bound construction is the real contribution here, and it looks correct. The new simplification operation (doubling then halving all edges except the marked one) is a genuinely useful trick, and the reduction to Bourgain's continued-fraction decomposition is clean. If that were all, I'd be happy. But the matching lower bound in Proposition 1.2 has a genuine gap, and as written the paper does not resolve the inverse problem up to a constant.\n\nThe upper half: Theorem 1.5 builds a simple planar graph realizing any target sum of continued-fraction values, and Lemmas 3.1–3.3 chain it to Bourgain's theorem to cover all ratios. I checked the algebra in the lemmas; the bookkeeping is fine. The construction is explicit and, modulo the unspecified constant in [Bou12], gives |V| = O(max(t/c, t/(t-c), log t)). That's a clean improvement over the earlier O(log t (log log t)^2) for a restricted range, and the step from multigraphs to simple graphs is nontrivial.\n\nThe soft spot: Proposition 1.2. The proof of the second inequality applies Lemma 3.4 to the dual G*. But G* need not be simple—in an n-cycle the dual is two vertices with n parallel edges. The line |F|/2 > 1/(2ρ(G*,e)) requires |F| > 1/ρ(G*,e), which is the first inequality applied to the dual, and that only holds when the dual is simple. The displayed chain fails; for C_n, (3.7) reads 1 > n/2. So the lower bound |V| ≥ C' t/(t−c) is not proved. This is not a cosmetic issue: without it, the paper only proves an upper bound, and the optimality claim rests on an unproved statement. The authors do mention in §4.3 that the constant 1/2 can be removed, but that's not the problem; the problem is the validity of the inequality itself.\n\nEverything else is honest: the use of Bourgain's theorem is flagged as external with unspecified constant, the small cases are absorbed, and the citation pattern looks appropriate. The upper bound alone is worth publishing, but the current text overstates the result. I'd send it to a referee who can assess whether the lower bound can be fixed; if it can't, the abstract and introduction need a major revision.","headline":"Strong upper-bound construction, but the lower-bound proof in Proposition 1.2 has a real gap, so the 'up to constant' optimality claim is currently unproved.","tokens_in":10170,"tokens_out":7346,"would_cite":false,"duration_ms":55234,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["05C10","11A55","05C30"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper resolves the inverse problem for effective resistance in planar graphs: for every reduced fraction $c/t$ it constructs a simple planar graph with an edge of resistance $c/t$ using $O(\\max(t/c,\\, t/(t-c),\\, \\log t))$ vertices…","keywords":["effective resistance","planar graphs","spanning trees","inverse problem","continued fractions","marked graphs","plane duality","Sedláček problem"],"falsifier":"Compute, for unit fractions $1/t$ with large $t$, the minimum total $S$-height of an expression $1/t = q_1+\\cdots+q_k$ with each $q_i \\in \\mathbb{Q}\\cap(-1,1)$; if this minimum grows faster than $C\\log t$, then the construction of Theorem 1.1 collapses because the $\\log t$ term in (1.3) would be too small. Graph-theoretically, enumerate simple planar graphs up to $n$ vertices, compute all achievable edge resistances, and check whether a planar graph with resistance $1/t$ appears for $n = o(t)$; Proposition 1.2 forbids $n \\le C' t$, so any such graph would falsify the lower bound.","tokens_in":9068,"feed_emoji":"⚡","tokens_out":5476,"duration_ms":42086,"temperature":0.7,"pith_summary":"This paper resolves the inverse problem for effective resistance in planar graphs: given any reduced fraction $c/t$, it determines, up to a universal constant, the smallest number of vertices of a simple planar graph with an edge of effective resistance $c/t$. The extremal size is $\\max(t/c,\\, t/(t-c),\\, \\log t)$ up to constant factors, with a matching construction and lower bound. The result matters because effective resistance is a central graph invariant with electrical-network and random-walk interpretations, and knowing the exact growth of the minimal example makes the achievable values of the invariant concrete. The proof links the graph invariant to continued fractions, showing that rational resistances are realized by assembling small marked graph pieces whose spanning-tree ratios add.","feed_headline":"Every rational resistance has a near-minimal planar graph","feed_subtitle":"For resistance c/t, the smallest planar graph's size is within a constant factor of max(t/c, t/(t-c), log t).","key_machinery":"The load-bearing object is the marked graph $(G,e)$ together with the spanning-tree ratio $\\zeta(G,e) = \\tau(G-e)/\\tau(G/e)$, since deletion–contraction gives $\\rho(G,e)=1/(1+\\zeta(G,e))$. Three operations on marked graphs carry the argument: the marked sum $\\oplus$, under which $\\zeta$ values add; the plane dual, which inverts $\\zeta$ (so $\\rho$ maps to $1-\\rho$); and the simplification operation (doubling every other edge, then subdividing every other edge), which makes a graph simple while preserving $\\zeta$ at the cost of a factor of $4$ in edges. A continued-fraction construction provides, for any rational $q$, a planar marked graph with $\\zeta(G,e)=q$ and edge count $S(q)+1$, where $S$ is the sum of the partial quotients. Bourgain's theorem supplies the arithmetic input: every rational in $[0,1)$ is a sum of rationals of small $S$-height, with total height $O(\\log)$ of the denominator.","core_discovery":"The paper's central claim is Theorem 1.1: for every coprime pair $t > c \\ge 1$ there exists a simple planar graph $G=(V,E)$ and an edge $e$ with $\\rho(G,e) = \\tau(G/e)/\\tau(G) = c/t$ and $|V| \\le C\\max(t/c,\\, t/(t-c),\\, \\log t)$ for a universal constant $C$. Proposition 1.2 gives the matching lower bound $|V| \\ge C'\\max(t/c,\\, t/(t-c),\\, \\log t)$ for any simple planar graph realizing a reduced resistance $c/t$ on a non-bridge edge. Together these resolve the inverse problem for planar effective resistance up to a universal constant. The paper also notes the corollary that for every integer $t\\ge 2$ there is a simple connected planar graph with $\\tau(G)\\equiv 0 \\pmod t$ and $O(\\log t)$ vertices.","pith_inferences":["The same marked-sum and simplification machinery should transfer to other invariants expressible as ratios of spanning-tree counts under deletion and contraction, such as other Tutte-polynomial evaluations.","The result suggests a concrete algorithmic challenge: make Bourgain's decomposition constructive, turning the existence proof into a polynomial-time procedure that outputs the $O(\\log t)$-vertex planar graph realizing a given rational resistance.","The exact form of the extremal bound invites a sharper conjecture: that the true minimal vertex count, before universal constants, is governed by the continued-fraction expansion of $c/t$ itself, not merely by the three-way maximum.","A direct comparison with the poset analogue from the authors' related work suggests that the logarithmic term may be removable in that setting too, since the obstruction there is a technical limitation of the flip-flop construction rather than the arithmetic growth."],"forward_implications":["For every reduced resistance $c/t$ there is a simple planar graph with that resistance on $O(\\max(t/c,\\, t/(t-c),\\, \\log t))$ vertices, and no such graph can be substantially smaller.","The extremal size is governed by the reciprocal of the resistance, the reciprocal of its complement, and the logarithm of the denominator, with the planar-dual symmetry $\\rho \\leftrightarrow 1-\\rho$ reflected in the symmetric term $t/(t-c)$.","For every integer $t\\ge 2$ there exists a simple connected planar graph with $\\tau(G)\\equiv 0 \\pmod t$ and only $O(\\log t)$ vertices.","The result extends earlier bounds that required the resistance to lie in $[1/3,2/3]$ and allowed multiple edges, to all rational resistances with simple graphs.","The construction is effective in principle: given $c/t$, one can assemble the graph from continued-fraction data once Bourgain's decomposition is supplied."],"supporting_citations":[{"why":"Supplies Theorem 1.4, the decomposition of every rational in $[0,1)$ into rationals of small continued-fraction height with total logarithmic height; it powers Lemmas 3.1 and 3.2.","marker":"[Bou12]"},{"why":"Provides the Bourgain–Kontorovich technology from which Theorem 1.4 is derived, and which the authors cite as the main tool behind both this paper and their earlier spanning-tree work.","marker":"[BK14]"},{"why":"Supplies Proposition 2.1 (continued-fraction representation as a planar marked graph) and Lemma 2.2 (marked-sum additivity), the core graph-construction lemmas, and the earlier weaker upper bound.","marker":"[CP24c]"},{"why":"The authors' previous work on Sedláček's inverse problem for spanning trees; its continued-fraction approach and the simplification operation in Section 2.5 are directly reused here.","marker":"[CKP24]"},{"why":"Gives the commute-time interpretation of effective resistance used in Lemma 3.4, the starting point for the lower bound in Proposition 1.2.","marker":"[C+96]"},{"why":"Supplies the commute-time inequality needed to turn Lemma 3.4 into the $t/c$ and $t/(t-c)$ lower bounds in Proposition 1.2.","marker":"[Lov96]"}],"fun_headline_variants":["Planar graphs hit all rational resistances efficiently","Every rational resistance fits a compact planar graph","Optimal planar size for any rational resistance","Tight planar bounds for effective resistance","Inverse problem solved for planar effective resistance"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The upper bound relies on Bourgain's theorem that every rational $d/c$ in $[0,1)$ can be written as a sum of rationals each of small continued-fraction height, with the total height growing only logarithmically in $c+d$.","fun_headline_variants_meta":{"raw":{"variants":["Planar graphs hit all rational resistances efficiently","Every rational resistance fits a compact planar graph","Optimal planar size for any rational resistance","Tight planar bounds for effective resistance","Inverse problem solved for planar effective resistance"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00026,"raw_usage":{"total_tokens":1540,"prompt_tokens":846,"completion_tokens":694,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":462,"completion_tokens_details":{"reasoning_tokens":628}},"tokens_in":462,"tokens_out":694,"duration_ms":6011,"temperature":1.0,"reasoning_tokens":628,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T14:20:10.085418+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute, for unit fractions $1/t$ with large $t$, the minimum total $S$-height of an expression $1/t = q_1+\\cdots+q_k$ with each $q_i \\in \\mathbb{Q}\\cap(-1,1)$; if this minimum grows faster than $C\\log t$, then the construction of Theorem 1.1 collapses because the $\\log t$ term in (1.3) would be too small. Graph-theoretically, enumerate simple planar graphs up to $n$ vertices, compute all achievable edge resistances, and check whether a planar graph with resistance $1/t$ appears for $n = o(t)$; Proposition 1.2 forbids $n \\le C' t$, so any such graph would falsify the lower bound.","supporting_citations":[],"review_version":1}