Proved theorem Unrefereed

The volume–determinant inequality for 7 and 8 twist regions

No dated manuscript catalogued; standalone verification archive linked above. On this site since · last revised · Verify & co-author
Kind
Theorem, computer-assisted (two bounds decide almost every diagram; certified volumes decide the rest)
Status
Proved, computer-assisted, given that SnapPy's verified computation encloses the volume of the knot whose diagram it was given; unrefereed
Prior work
The searches found no source stating the inequality for this class; published results cover parts of it, credited below
On this page
  1. The statement
  2. Check it yourself
  3. Worked examples
  4. Status and who checked it
  5. Prior work
  6. Further reading

The statement

A reduced alternating diagram of a knot KK splits into twist regions, runs of crossings where two strands twist round each other. The volume–determinant inequality (Vol–Det), conjectured by Champanerkar, Kofman and Purcell for every alternating hyperbolic knot and checked by them up to 16 crossings, is

vol(K)<2πlog⁡det⁡(K).\mathrm{vol}(K) < 2\pi \log \det(K).

Result. Every alternating hyperbolic knot with 7 twist regions and at least 17 crossings satisfies Vol–Det, and so does every one with 8 twist regions and at least 17 crossings. Each carries the label "proved, computer-assisted, given that SnapPy's verified computation encloses the volume of the knot whose diagram it was given". It says nothing of links and nothing at 16 crossings or fewer.

Check it yourself

The determinant is a weighted count τ\tau of spanning trees of a plane map with one edge per twist region, weighted by the sizes (crossings per region). The volume has two upper bounds that depend on the shape: the nerve bound and Adams's bigon-chain bound. If 2πlog⁡τ2\pi \log \tau is above either, Vol–Det holds for that diagram.

  • Seven twist regions. Take the 7-bond (two vertices, seven parallel edges), every twist across its edge, with sizes 5, 2, 2, 2, 2, 2, 2: 17 crossings. Its spanning trees are its seven single edges. One leaves out the 5, giving 26=642^6 = 64; six leave out a 2, giving 5⋅25=1605 \cdot 2^5 = 160 each. So τ=64+6⋅160\tau = 64 + 6 \cdot 160 = 1,024. The bigon-chain bound there is about 34.58, and 2πlog⁡τ2\pi \log \tau, about 43.55, is above it.
  • Eight twist regions. The 8-bond with sizes 3, 2, 2, 2, 2, 2, 2, 2: 17 crossings. τ=27+7⋅3⋅26=128+7⋅192\tau = 2^7 + 7 \cdot 3 \cdot 2^6 = 128 + 7 \cdot 192 = 1,472, and 2πlog⁡τ2\pi \log \tau is about 45.83. The bigon-chain bound is seven twists of two crossings at 6v36 v_3 each (v3=1.01494…v_3 = 1.01494\ldots), one of three at 7.855, less 10.087: about 40.40, below 45.83.
  • The closest knot at seven twist regions. Of the 300 knots that needed a certified volume, the closest has a volume of at most 28.04 against 2πlog⁡τ=47.252\pi \log \tau = 47.25.

Worked examples

How the seven-region case was settled.

  • Every plane map with seven edges was grown by computer and held to Tutte's published count of rooted maps, 408 at seven edges, so no map is missing.
  • The nerve bound or the bigon-chain bound decided every diagram with at least 17 crossings except 658. Of those, 300 are knots and 358 are links (250 of two components, 108 of three).
  • A certified volume was computed for each of the 300 knots: Vol–Det holds on all 300, and not narrowly. By crossing number they are 80 + 80 + 52 + 16 + 72 = 300.

The eight-region case ran the same way: the two bounds decided every row but 417,798; certified volumes settled 173,627 of the 173,628 knot rows; 14 certificates settled the last row.

Status and who checked it

  • Proved, computer-assisted, with one stated condition. The two bounds are proved (Adams's bigon-chain bound is his published theorem; the nerve bound and the route's lemmas are proved on earlier pages for six twist regions). The remaining knots rest on certified volume enclosures from one tool, SnapPy's verified computation, and no reviewer checked that tool at its source. Hence the word "given".
  • Who checked it. Two falsifier agents graded each piece, the second blind to the other's grade; a script reads each grade's own file and names the two grades that hold each piece the label rests on. None is held by only one. Unrefereed.
  • Not covered. Nothing at 16 crossings or fewer (that rests on Champanerkar, Kofman and Purcell's check); nothing of links. A nine-region computation exists but has an outstanding review issue and is not claimed.

Prior work

Credited. The conjecture and its check up to 16 crossings are Champanerkar, Kofman and Purcell's. Burton proved Vol–Det for 2-bridge links, alternating 3-braids, alternating pretzel links and highly twisted alternating links (from 16,178 crossings at seven twist regions, as he states it). Egorov and Vesnin's theorem needs more than eight twist regions. Stanggassinger's unrefereed computation covers alternating knots through 19 crossings. The bigon-chain bound is Adams's.

Searches run. A dated search of 20 requests by published terms, then MR Lookup, Google Scholar and CORE. A blind reviewer agent reads them so: "cited sources cover the class up to 19 crossings, the last three of them by an unrefereed computation, and from 81,399 on", and "For the range between, the searches name no source". The inequality for this class was not found in the searches logged; the result's reach lies between Stanggassinger's census and Burton's large crossing numbers. On 2026-11-22, a Google Scholar sweep (5 queries) found no prior work stating this claim.

Still owed. Full texts beyond title and snippet for several cited works (including the Baykur–Kirby–Ruberman 2026 problem list, the 2025 Vesnin–Egorov Omsk abstract, and BIRS workshop reports), the unshown MathSciNet MR Lookup matches beyond the free three-record cap (15 of 18 under Vesnin, 5 of 8 under Futer), and a clean re-query of the failed CORE request.

Further reading