OpenFrontiers
← Back to the index
OP-12093OpenGraph theory

Conway thrackle constant (C78C_{78})

Conway thrackle constant: determine C78C_{78} or improve its bounds; the repository records 1 (lower) and 1.393 (upper).

Editorial MSC2020 classification. Original article codes and source evidence are recorded in the literature directory.

Licensed source statementAs of Repository table at commit 2c1968c (2026-09-26); open while the recorded lower and upper bounds differSource & evidence ↓

Problem statement

In topological graph theory, a thrackle is a drawing of a finite graph in the plane in which every pair of edges meets precisely once, either at a common endpoint or at a proper crossing. If t(n)t(n) denotes the maximum number of edges in a thrackle on nn vertices, then Conway's thrackle conjecture states that t(n)=nt(n)=n for every n≥3n\ge 3. [FP2011-def-tn-conj]

For a finite graph GG, write V(G)V(G) for its vertex set and E(G)E(G) for its edge set. We define

C78 := sup⁡{∣E(G)∣∣V(G)∣:G admits a thrackle drawing in the plane}.C_{78} \ :=\ \sup\left\{ \frac{\lvert E(G)\rvert}{\lvert V(G)\rvert} : G \text{ admits a thrackle drawing in the plane} \right\}.

Equivalently,

C78 = sup⁡n≥1t(n)n.C_{78} \ =\ \sup_{n\ge 1}\frac{t(n)}{n}.

Thus Conway's thrackle conjecture is equivalent to the assertion that C78=1C_{78}=1. [FP2011-def-tn-conj]

At present, the best established range is

1 ≤ C78 ≤ 1.393.1\ \le\ C_{78}\ \le\ 1.393.

The lower bound comes from thrackled cycles, and the upper bound from Xu's current record bound m≤1.393(n−1)m\le 1.393(n-1) as recorded in recent survey-style sources. [FP2011-open-sharp] [KSTZ2025-current-best-special]

Lean 4

Not yet attached

No Lean statement or proof is attached yet. Have a formalization? Share its source to add it to this problem.

Background

References (from the repository page)

  • [CN2000] Cairns, Grant; Nikolayevsky, Yury. Bounds for generalized thrackles. Discrete & Computational Geometry 23 (2000), no. 2, 191–206. DOI: 10.1007/PL00009495. Google Scholar

  • [CN2000-ub-lps-cn] loc: Springer PDF p.1, Introduction, file: 10.1007-PL00009495.pdf quote: “Lovász et al. proved: Theorem 1 [LPS]. (a) for thrackles, m≤2n−3m \le 2n - 3, (b) for generalized thrackles, m≤3n−4m \le 3n - 4, (c) a bipartite graph can be drawn as a generalized thrackle if and only if it is planar. We give the following improvement: Theorem 2. (a) for thrackles, m≤32(n−1)m \le \frac{3}{2}(n - 1), (b) for generalized thrackles, m≤2n−2m \le 2n - 2.”

  • [FP2011] Fulek, Radoslav; Pach, János. A computational approach to Conway's thrackle conjecture. Computational Geometry 44 (2011), no. 6–7, 345–355. DOI: 10.1016/j.comgeo.2011.02.001. arXiv PDF: arXiv:1002.3904. Google Scholar

  • [FP2011-def-tn-conj] loc: arXiv PDF p.1, Abstract, file: 1002.3904.pdf quote: “A drawing of a graph in the plane is called a thrackle if every pair of edges meets precisely once, either at a common vertex or at a proper crossing. Let t(n)t(n) denote the maximum number of edges that a thrackle of nn vertices can have. According to a 40 years old conjecture of Conway, t(n)=nt(n) = n for every n≥3n \ge 3. For any ε>0\varepsilon > 0, we give an algorithm terminating in eO((1/ε2)ln⁡(1/ε))e^{O((1/\varepsilon^2)\ln(1/\varepsilon))} steps to decide whether t(n)≤(1+ε)nt(n) \le (1 + \varepsilon)n for all n≥3n \ge 3. Using this approach, we improve the best known upper bound, t(n)≤32(n−1)t(n) \le \frac{3}{2}(n - 1), due to Cairns and Nikolayevsky, to 167117n<1.428n\frac{167}{117}n < 1.428n.”

  • [FP2011-open-sharp] loc: arXiv PDF p.1, §1 Introduction, file: 1002.3904.pdf quote: “A drawing of GG is called a thrackle if every pair of edges meet precisely once, either at a common vertex or at a proper crossing. (A crossing pp of two curves is proper if at pp one curve passes from one side of the other curve to its other side.) More than forty years ago Conway [18, 2, 15] conjectured that every thrackle has at most as many edges as vertices, and offered a bottle of beer for a solution. Since then the prize went up to a thousand dollars. In spite of considerable efforts, Conway’s thrackle conjecture is still open. It is believed to represent the tip of an “iceberg,” obstructing our understanding of crossing patterns of edges in topological graphs. If true, Conway’s conjecture would be tight as any cycle of length at least five can be drawn as a thrackle, see [17].”

  • [FP2019] Fulek, Radoslav; Pach, János. Thrackles: An improved upper bound. Discrete Applied Mathematics 259 (2019), 226–231. DOI: 10.1016/j.dam.2018.12.025. arXiv PDF: arXiv:1708.08037. Google Scholar

  • [FP2019-ub-gx-fp] loc: arXiv PDF p.1, §1 Introduction, file: 1708.08037v1.pdf quote: “The first linear upper bound on the number of edges of a thrackle, in terms of the number of vertices nn, was established in [6 ]. This bound was subsequently improved in [ 1] and [4 ], with the present record, 1.4n1.4n, held by Goddyn and Xu [ 5], which also appeared in the master thesis of the second author [9]. One of the aims of this note is to show that this latter bound is not best possible. Theorem 1. Any thrackle on n>3n > 3 vertices has at most 1.3984n1.3984n edges.”

  • [KSTZ2025] Keszegh, Balázs; Suk, Andrew; Tardos, Gábor; Zeng, Ji. Unavoidable patterns and plane paths in dense topological graphs. Preprint (2025). arXiv:2512.04795. arXiv PDF: arXiv:2512.04795. Google Scholar

  • [KSTZ2025-current-best-special] loc: arXiv PDF p.1, §1 Introduction, file: 2512.04795.pdf quote: “The famous thrackle conjecture of Conway states that a thrackle with nn vertices has at most nn edges, see [21] for a detailed history of the problem. For the special case of geometric graphs, the thrackle conjecture is proved by Erdős, and later, a short proof is given by Perles (see [21]). The proof of Perles also works if the drawing is outerplanar [7], i.e., if the points lie on the boundary of a disk and the edges lie inside this disk. The case that the edges are xx-monotone is settled by Pach and Sterling [21]. In general, after a long series of improvements (Lovász–Pach–Szegedy [17] proved 2n2n, Cairns–Nikolayevsky [6] proved 1.5n1.5n, improved further by Fulek–Pach [12, 13] and Goddyn–Xu [14]), the current best upper bound on the number of edges of an nn-vertex thrackle is 1.393(n−1)1.393(n - 1) by Xu [31].”

Progress & source assessment

Repository table (commit 2c1968c, 26 September 2026): best known lower bound 1; best known upper bound 1.393. Values marked with an asterisk have minimal verification.

Known upper bounds

Bound Reference Comments
22 [CN2000] Lovász–Pach–Szegedy proved m≤2n−3m\le 2n-3, hence C78≤2C_{78}\le 2. [CN2000-ub-lps-cn]
32\frac{3}{2} [CN2000] Cairns–Nikolayevsky improved the linear bound to m≤32(n−1)m\le \frac{3}{2}(n-1), hence C78≤32C_{78}\le \frac{3}{2}. [CN2000-ub-lps-cn]
167117\frac{167}{117} [FP2011] Fulek–Pach proved t(n)≤167117n<1.428nt(n)\le \frac{167}{117}n<1.428n, so C78≤167117C_{78}\le \frac{167}{117}. [FP2011-def-tn-conj]
1.41.4 [FP2019] Previous record, attributed there to Goddyn–Xu: m≤1.4nm\le 1.4n, hence C78≤1.4C_{78}\le 1.4. [FP2019-ub-gx-fp]
1.39841.3984 [FP2019] Fulek–Pach improved the bound to m≤1.3984nm\le 1.3984n, hence C78≤1.3984C_{78}\le 1.3984. [FP2019-ub-gx-fp]
1.3931.393 [KSTZ2025] Current best upper bound: Xu's m≤1.393(n−1)m\le 1.393(n-1) bound, as recorded in recent literature, gives C78≤1.393C_{78}\le 1.393. [KSTZ2025-current-best-special]

Known lower bounds

Bound Reference Comments
11 [FP2011] Any cycle of length at least five can be drawn as a thrackle, and such a cycle has ∣E∣/∣V∣=1\lvert E\rvert/\lvert V\rvert=1. [FP2011-open-sharp]
References
  1. [1]Davis, Damek; Ivanisvili, Paata; Tao, Terence (maintainers). Optimization Constants in Mathematics, github.com/teorth/optimizationproblems (commit 2c1968c, 26 September 2026). — Conway thrackle constant ($C_{78}$) linkApache-2.0 (repository license); constant pages reproduced with attribution; repository references retained.

Recorded progress

Add yours →

Nothing recorded yet. A pointer to a paper the entry is missing counts.

Machine attempts

0 recorded

Attempts run against the statement below and are recorded in full, including the failures. The point is to stop the next model — and the next person — from re-walking a road that is already known to end. Each attempt costs 25 credits.

No machine attempts recorded.

Sign in to run an attempt. New accounts start with enough credits for ten.

Discussion

0 notes

Notes are for clarifying the statement, flagging a missing reference, or recording why an approach fails. Claimed solutions belong in a submission, where they can be reviewed.

No notes yet.

Sign in to add a note.