REVIEW 4 major objections 5 minor 14 references
Untwisting 3-strand torus knots
T0 review · 4 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read Every 3-strand torus knot $T(3,n)$ with $n\ge 4$ and $3\nmid n$ has topological 4-genus exactly $\lceil 2n/3\rceil$, equal to its signature bound and its untwisting number.
desk verdict Proves the topological 4-genus equals the signature bound for all 3-strand torus knots, with a clean asymptotic bonus, but the key local move in Figure 1 is verified only pictorially and should be pinned down algebraically before it goes to press. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The engine is a calculus for positive 3-braids written as $[k_1,\dots,k_n] = a^{k_1}b\,a^{k_2}b\cdots a^{k_n}b$, together with the local move drawn in Figure 1: the braid word $abbaabba$ is carried to $bb$ by one null-homologous twist on four strands followed by one crossing change. Iterating this move, with Lemma 4's presentations of powers of the full twist, produces the untwisting sequences behind Lemma 5's bounds $t(T(3,3k+4))\le 2k+3$ and $t(T(3,3k+5))\le 2k+4$, which are exactly $\lceil 2n/3\rceil$. The passage from untwisting to genus is the general theorem that a null-homologous twist changes the knot by a surface of genus one inside the 4-ball, so the untwisting number is an upper bound for the topological 4-genus. The lower side is supplied by the jump formula for the Levine-Tristram signature, which lets the paper compute $\hat{\sigma}$ from the positions of discontinuities and their signs. For the asymptotic result, the key operation is that a full twist on $2n$ strands can be transformed by one twist into two parallel double full twists on $n$ strands.
What would settle it
Run a braid-group computation on the word in Figure 1: write the null-homologous four-strand twist as an explicit braid word and check whether one further crossing change turns the closure into the unknot, i.e. whether $abbaabba$ really lies one twist plus one crossing change away from $bb$. A direct calculation of the Levine-Tristram signature of the intermediate four-strand twist closure would detect a mismatch: if its signature is not compatible with the claimed trefoil-summand diagram, the local move collapses and Theorem 1 loses its upper bound.
Extended reading notes
Core claim
On the paper's own terms, the discovery is that for $T(3,n)$ the topological 4-genus problem collapses to a braid-word problem. Theorem 1 states that for every natural number $n\ge 4$ with $3\nmid n$, $g_t(T(3,n)) = \hat{\sigma}(T(3,n))/2 = t(T(3,n)) = \lceil 2n/3\rceil$, where $\hat{\sigma}$ is the maximum of the Levine-Tristram signature over the unit circle away from roots of the Alexander polynomial and $t$ is the untwisting number. Because the inequality $\hat{\sigma}/2 \le g_t \le t$ holds for all knots, the two outer quantities are forced into equality by constructing untwisting sequences of exactly the signature bound's length; the minimal genus surfaces are never drawn but are guaranteed to exist by the theorem that null-homologous twists give locally flat surfaces. The paper also establishes upper bounds $n \le g_t(T(4,n)) \le n+1$ and $(3n+1)/2 \le g_t(T(6,n)) \le (3n+3)/2$ within the signature-to-untwisting window, and a limsup bound of $14/27$ for the ratio $g_t/g$ over all torus knots.
Load-bearing premise
Everything rests on the pictorial claim in Figure 1 that one null-homologous twist on four strands plus one crossing change transforms the braid $abbaabba$ into $bb$; the paper verifies that move only by the figure, not by an algebraic sequence of braid relations, and every untwisting upper bound in Lemma 5 is built from it.
Editorial extensions
If this is right
- The whole infinite family of 3-strand torus knots satisfies $g_t = \hat{\sigma}/2 = t = \lceil 2n/3\rceil$, so the signature lower bound is both sharp and attained by untwisting sequences rather than by explicit minimal surfaces.
- Since the untwisting number equals the topological 4-genus for these knots, any algorithm that computes untwisting numbers for them also computes their topological 4-genus.
- For 4- and 6-strand torus knots, the topological 4-genus lies within 1 of the signature bound, so the equality pattern nearly persists for two more strand counts.
- The asymptotic ratio $\limsup g_t/g$ over all torus knots is at most $14/27 \approx 0.519$, improving the previous $2/3$ upper bound and supporting the conjecture that the ratio tends to $1/2$.
- The untwisting construction is inductive and explicit: every $T(3,n)$ admits a prescribed sequence of null-homologous twists, so the genus bound is certified by a finite braid word.
Reading between the lines
- The same braid-calculus framework could be pushed to torus knots with more strands: Proposition 2's 'off by at most 1' for four and six strands suggests that exact equality might hold for fixed $p$ if an analogue of Figure 1's local move exists for $p$-strand full twists.
- A computer search over braid words could test whether the Figure 1 move is the first of an infinite family of moves, and whether all 3-strand torus knots satisfy $t = \lceil 2n/3\rceil$ even when $n$ is divisible by three, where $T(3,n)$ is a 3-component link.
- If the asymptotic ratio really converges to $1/2$, the $14/27$ bound here is not the end; the paper's induction from 3-strand bases, combined with sharper untwisting bounds at higher strand counts, is a natural route to close the gap.
- Untwisting number is a combinatorial, unknotting-type invariant; its equality with a signature bound suggests that topological 4-genus for families of braid closures could be computed without constructing surfaces, which would be useful for computational knot tables.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper studies the topological 4-genus of torus knots. The main result (Theorem 1) asserts that for every n ≥ 4 not divisible by 3, the topological 4-genus of the 3-strand torus knot T(3,n) equals the maximal Levine-Tristram signature bound σ_hat/2 = ceil(2n/3), and that this value is also the untwisting number t(T(3,n)). The proof combines McCoy's inequality g_t ≤ t with explicit sequences of null-homologous twists (upper bound) and Litherland's signature jump formula (lower bound). The paper further claims (Proposition 2) that the same invariants for T(4,n) and T(6,n) are within 1 of the signature bound, and (Theorem 3) improves the asymptotic upper bound on g_t/g for torus knots from 2/3 to 14/27.
Significance. If Theorem 1 holds, it provides the first infinite family of torus knots for which the topological 4-genus is exactly equal to the signature bound, giving a clean topological analogue of the smooth local Thom conjecture; it also identifies the untwisting number with these invariants. The proof is elementary and mostly explicit, building on recent work of McCoy, and the paper gives concrete braid sequences for the upper bounds. The improvement of the asymptotic ratio is a modest but real advance. The main weakness is that the key local move (Figure 1) and several secondary moves (Figure 2, Figure 3) are verified only pictorially, and the lower bounds in Proposition 2 are explicitly left unproved.
major comments (4)
- [Section 3, Lemma 5 and Figure 1] The assertion that the braids abbaabba and bb are related by one null-homologous twist on four strands followed by one crossing change is the linchpin of Lemma 5 and hence of the entire upper bound in Theorem 1; it is verified only by a drawing, not by an algebraic braid computation. This move is used for both infinite families T(3,6k+16) and T(3,6k+19), and a failure of this pictorial claim would invalidate the untwisting upper bound and Theorem 1. Please provide an explicit algebraic sequence of braid relations (or a short computer script) establishing this move.
- [Section 4, Proposition 2] The proof of Proposition 2 states that 'by a similar calculation ... one may prove' the inequalities σ_hat(T(4,n)) ≥ 2n and σ_hat(T(6,n)) ≥ 3n+1, and then 'We note (without proof) that the stated inequalities for σ_hat are in fact equalities.' Since Proposition 2 is stated as a theorem and advertised in the abstract and introduction, these bounds need a proof; if they are omitted intentionally, the proposition should be rephrased as conditional or its proof sketched. The absence of proof is especially notable because the paper itself flags it.
- [Section 4, Figure 2] The 'crucial move' transforming (abc)^12 into (bc)^12 by four twist operations is presented via Figure 2 without algebraic verification, and this move underpins all upper bounds for t(T(4,n)) in Proposition 2. As with Figure 1, a pictorial verification is not sufficient for a load-bearing step; please provide a braid-word or computational check.
- [Section 5, proof of Theorem 3] The proof states (items (1)–(4)) that a single twist transforms T(2n,2n+1) into the 'disjoint union' of copies of T(n,2n+1), whereas the preceding sentence and the analogous argument in Section 4 refer to the connected sum. The untwisting number is defined for knots, and untwisting each component of a split union into an unknot would produce an unlink, not the unknot; the bound on t(T(3·2^k, 3·2^k+1)) therefore requires the connected-sum interpretation. Please reconcile the terminology and make explicit how the untwisting operations are applied to the summands.
minor comments (5)
- [Abstract / general] There is a typo in the abstract: 'off' should be 'off'.
- [Figure 1 caption] The caption reads 'The second twist is a crossing changes'; this should be 'a crossing change'.
- [Section 4] In the paragraph on signature bounds for T(6,n), the expression 'T(6,k+5)' should presumably be 'T(6,6k+5)' (or another integer congruent to 5 modulo 6), since T(6,n) with n coprime to 6 is a torus knot; please clarify the indexing.
- [Section 5] The list in items (1)–(4) uses 'disjoint union' while the surrounding prose and Section 4 use 'connected sum'; please use one consistent term throughout, as discussed in the major comments.
- [Section 3, after Lemma 4] The notation '5k' in 'L[3,5k,4,3,5k,4]' is explained as a sequence of length k, but the same notation is also used for a scalar in Lemma 5; a brief reminder of this convention would help the reader.
Circularity Check
No circularity: the proof sandwiches an independently computed lower bound with an explicit untwisting upper bound.
full rationale
Theorem 1 is proved by establishing hat_sigma(T(3,n))/2 <= g_t(T(3,n)) <= t(T(3,n)) <= ceil(2n/3) = hat_sigma(T(3,n))/2. The lower bound uses Powell's signature bound and an explicit Levine-Tristram computation from Litherland's jump formula; the upper bound is produced by Lemma 5's explicit untwisting sequences via McCoy's theorem. Neither bound is defined in terms of the other: the equalities emerge only after both bounds are independently verified. Lemma 4's braid calculus is a bookkeeping tool, and the reference to the authors' earlier paper [2] supplies a subadditivity principle for extending an asymptotic bound, not the target 4-genus equality. Proposition 2 reduces the 4- and 6-strand cases to the already-proved 3-strand theorem, which is legitimate; its unproved signature equalities are not load-bearing because only lower bounds are used. Figure 1's local twist assertion is a topological verification that could be wrong, but that is a correctness risk, not circularity: no parameter is fitted to the output, and no prediction is identified with the data used to define it.
Assumptions & free parameters
assumptions (6)
- domain assumption McCoy's theorem: the untwisting number is an upper bound for the topological 4-genus.
- domain assumption Powell's signature bound: g_t(K) >= sigma_hat(K)/2.
- standard math Litherland's formula for jumps of the Levine-Tristram signature of torus knots.
- standard math Gordon-Litherland-Murasugi periodicity of signatures.
- standard math Genus formula g(T(p,q)) = (p-1)(q-1)/2.
- standard math Braid group relations on B3 and the calculus of positive braids from [1].
Cite this review
Pith. "Pith review of Untwisting 3-strand torus knots." pith.science (2026). https://pith.science/paper/L5DL3GGR
@misc{pith2026190901003,
author = {Pith},
title = {Pith review of: Untwisting 3-strand torus knots},
year = {2026},
howpublished = {\url{https://pith.science/paper/L5DL3GGR}},
note = {Machine review of arXiv:1909.01003}
}
read the original abstract
We prove that the signature bound for the topological 4-genus of 3-strand torus knots is sharp, using McCoy's twisting method. We also show that the bound is off by at most 1 for 4-strand and 6-strand torus knots, and improve the upper bound on the asymptotic ratio between the topological 4-genus and the Seifert genus of torus knots from 2/3 to 14/27.
Figures
Reference graph
Works this paper leans on
-
[1]
Baader: Positive braids of maximal signature , Enseign
S. Baader: Positive braids of maximal signature , Enseign. Math. 59 (2013), no. 3-4, 351–358
work page 2013
- [2]
-
[3]
M. H. Freedman: The topology of four-dimensional manifolds , J. Differential Geom. 17 (1982), no. 3, 357–453
work page 1982
-
[4]
C. McA. Gordon, R. A. Litherland, K. Murasugi: Signatures of covering links , Canad. J. Math. 33 (1981), 381–394
work page 1981
-
[5]
Ince: The untwisting number of a knot , Pacific J
K. Ince: The untwisting number of a knot , Pacific J. Math. 283 (2016), no. 1, 139–156
work page 2016
-
[6]
R. A. Litherland: Signatures of iterated torus knots , Topology of low-dimensional manifolds (Proc. Second Sussex Conf., Chelwood Gate, 1977), pp. 71–84, Lecture Notes in Math. 722, Springer, Berlin, 1979
work page 1977
-
[7]
C. Livingston: Null-homologous unknotting, arXiv:1902.05405
work page Pith review arXiv 1902
-
[8]
P. B. Kronheimer, T. S. Mrowka: The genus of embedded surfaces in the projective plane, Math. Res. Lett. 1 (1994), no. 6, 797–808
work page 1994
Show all 14 references
-
[9]
McCoy: Null-homologous twisting and the algebraic genus , arXiv:1908.04043
D. McCoy: Null-homologous twisting and the algebraic genus , arXiv:1908.04043
1908 arXiv
-
[10]
McCoy: Gaps between consecutive untwisting numbers, Glasgow Math
D. McCoy: Gaps between consecutive untwisting numbers, Glasgow Math. J. 63 (2021), no. 1, 59–65
2021
-
[11]
Powell: The four-genus of a link, Levine-Tristram signatures and satellites , J
M. Powell: The four-genus of a link, Levine-Tristram signatures and satellites , J. Knot Theory Ramifications 26 (2017), no. 2
2017
-
[12]
Rasmussen: Khovanov homology and the slice genus , Invent
J. Rasmussen: Khovanov homology and the slice genus , Invent. Math. 182 (2010), no. 2, 419–447
2010
-
[13]
Rudolph: Some topologically locally-flat surfaces in the complex projective plane , Comment
L. Rudolph: Some topologically locally-flat surfaces in the complex projective plane , Comment. Math. Helv. 59 (1984), no. 4, 592–599
1984
-
[14]
Rudolph: Quasipositivity as an obstruction to sliceness , Bull
L. Rudolph: Quasipositivity as an obstruction to sliceness , Bull. Amer. Math. Soc. (N.S.) 29 (1993), no. 1, 51–59. Mathematisches Institut, Sidlerstr. 5, 3012 Bern, Switzerland Mathematisches Institut, Sidlerstr. 5, 3012 Bern, Switzerland Uni Regensburg, Fakult¨ at f¨ ur Math...
1993
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.