From ABC to Effective Roth and Ridout Constants for Cubic Roots
Karsten Müller, Michael Taktikos · 2026
Read From ABC to Effective Roth and Ridout Constants for Cubic Roots by Karsten Müller, Michael Taktikos (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-11376.txt.
Introduction
Karsten Muller and Michael Taktikos make a deep number theory bound effective. Their 2026 preprint links the ABC conjecture to Roth's theorem. Roth's theorem bounds how well irrationals can be approximated. Bombieri showed in 1994 that ABC would imply Roth's result.
Van Frankenhuysen gave a full proof of that link in 1999. The new work builds on an explicit formula by Bombieri. Van der Poorten shares credit for that formula on coefficients. The formula covers continued-fraction coefficients of algebraic numbers. The team narrows it to cubic roots as the test case. From it they derive a hard bound for a Roth-type constant. The bound holds if an effective form of ABC is true.
Catalog record
| CREATOR | Karsten Müller, Michael Taktikos |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 21 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-11376
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
From ABC to Effective Roth and Ridout Constants for Cubic Roots Karsten Müller, Michael Taktikos Abstract Enrico Bombieri showed conditionally (1994) that the ABC conjecture implies Roth’s theorem, and Van Frankenhuysen (1999) later provided a complete proof. Building on Bombieri’s and Van der Poorten’s explicit formula for continued-fraction coefficients of algebraic numbers—specialized to cubic roots—we derive an effective bound for a Roth-type constant assuming an effective form of ABC. Roth’s original argument establishes existence but does not yield an explicit value; our approach makes the dependence on the ABC parameters explicit and also gives an explicit bound in the corresponding special case of Ridout’s theorem. We then introduce the notion of approximation gain as a refinement of the “quality” of an abc-triple. For c in a large computational range (verified up to c < 2 63 c<2^{63} ), the approximation gain remains below a strikingly small threshold, motivating the conjecture that the approximation gain is always smaller than 3 2 \frac{3}{2} . This suggests a potential strategy for attacking ABC by bounding approximation gain and power gain separately. 1 1 1 The authors are grateful to Benne de Weger, Joshua Lampert, Timm Lampert, Enrico Bombieri, Wadim Zudilin, Michel Waldschmidt, Preda Mihailescu, Paul Vojta, Stephane Fischler, Robin Zhang, Abderrahmane Nitaj, Sam Chow, Noah Lebowitz-Lockard, Ingo Althöfer, ChatGPT and Ulrich Tamm for helpful comments. 1 Introduction Roth’s theorem is famously deep, and its original proof is both long and ineffective (see Roth 1955 ). In particular, it guarantees the existence of a constant in the Diophantine approximation bound, but it does not provide an explicit value or a practical way to compute one. Bombieri proved conditionally in 1994 that Roth’s theorem follows from the ABC conjecture, and Van Frankenhuysen later gave a full proof in 1999 ( Van Frankenhuysen 1999 ). In this paper we revisit that implication with an additional goal: making the constants explicit in a natural family of cases. We focus on cubic roots and use an explicit continued-fraction formula of Bombieri and Van der Poorten ( Bombieri & Van der Poorten 1975 ) for the coefficients of regular continued fractions of algebraic numbers. This yields a concrete relationship between the \varepsilon -parameters in Roth-type inequalities and the corresponding \varepsilon in ABC, and it leads to an explicit upper bound for the inverse of the Roth constant in our setting. In our earlier paper ( Sibbertsen 2022 ), we showed that a weakened form of ABC follows from Roth’s theorem in certain special cases. The present work uses essentially the same vocabulary and techniques, but in the opposite direction: assuming (an effective form of) ABC, we extract explicit Diophantine approximation bounds. We also introduce the approximation gain, which is always lower or equal to the quality of an ABC hit. 2 Approximation gains for ABC 2.1 Roth’s theorem In the following, the coefficients of the regular continued fraction are called b_{n} and the approximants \frac{p_{n}}{q_{n}} . Theorem 1 Roth’s Theorem: Let a be an algebraic number. Then for every \varepsilon>0 , there exists a constant C dependent on the algebraic number a and \varepsilon such that for all positive integers p and q : |a-\frac{p}{q}|>\frac{C}{q^{2+\varepsilon}}. (1) Proof 1 See Roth 1955 . Van Frankenhuysen 1999 discusses Roth’s Theorem and its implication by the ABC conjecture in detail and investigates effective versions of Mordell’s conjecture. Furthermore, Granville and Tucker (2002, p.1229) show that a generalized version of Roth’s theorem implies the ABC conjecture effectively. With the help of the inequality for regular continued fractions from this follows b_{n+1}\leq\frac{q_{n}^{\varepsilon}}{C} 2.2 Formula by Bombieri and Van der Poorten Our main technical input is the explicit formula of Bombieri and Van der Poorten for the continued-fraction coefficients of algebraic numbers by ( Bombieri & Van der Poorten 1975 , p.151, Theorem 3, formula (13)) , in which {f(x)} is the minimal polynomial of the algebraic number. The formula links two quantities—call them d_{n} and b_{n} — and is therefore well suited to bridge ABC-type information (which naturally arises from integer identities involving d_{n} with Diophantine approximation data encoded by b_{n} ). \frac{f^{\prime}(x)}{f(x)} leads to d_{n} in the denominator. This holds because \frac{f^{\prime}(x)}{f(x)} results in \frac{f^{\prime}(\frac{p_{n}}{q_{n}})}{f(\frac{p_{n}}{q_{n}})}=\frac{q_{n}^{s}sp_{n}^{s-1}}{d_{n}q_{n}^{s-1}} for the convergents with f(x)=x^{s}-k being the minimal polynomial of the s th root of k. For the cubic case the error term of the formula ( Bombieri & Van der Poorten 1975 , p.151, Theorem 3, formula (13)) can be ignored due to the effective bound by Liouville’s theorem. 2.3 ABC Conjecture To formulate the ABC Conjecture we need the following definition: Definition 1 For a positive integer a , rad(a) is the product of the distinct prime factors of a . Definition 2 (ABC Conjecture) For every positive real number \varepsilon , there exists a constant K_{\varepsilon} such that for all triples (a,b,c) of coprime positive integers, with a+b=c : c<K_{\varepsilon}\cdot rad(abc)^{1+\varepsilon}. Note that this version of the ABC Conjecture is not effective. This is so because only the existence of a constant K_{\varepsilon} is claimed without specifying how to calculate K_{\varepsilon} for given \epsilon . 2.4 Resulting ABC equations We look at 3rd roots of an integer k which is not a perfect cube. The coefficients of the regular continued fraction are called b_{n} and the approximants \frac{p_{n}}{q_{n}} . We call the equations with positive numbers p_{n}^{3}=kq_{n}^{3}+d_{n} resp kq_{n}^{3}=p_{n}^{3}+d_{n} resulting equations. To use ABC we have to get rid off the cofactors of the resulting equations. For this purpose we observe that gcd( kq_{n}^{3},d_{n} ) is the natural greatest common divisor to divide out. This will be called gcd in the following. It is larger or equal to 1 and smaller or equal to k as gcd( q_{n},p_{n} ) = 1. So with \frac{p_{n}^{3}}{gcd}=\frac{kq_{n}^{3}}{gcd}+\frac{d_{n}}{gcd} we get an ABC equation. 2.5 Construction of the effective bound for the inverse of Roth’s constant Theorem 2 From an effective version of the ABC conjecture Roth’s Theorem effective for resulting equations of 3rd roots follows: Let a be the 3rd root of k. Then for every \varepsilon>0 , there exists a constant C dependent on the algebraic number a and \varepsilon such that for all positive integers p and q : |a-\frac{p}{q}|>\frac{C}{q^{2+\varepsilon}}. (2) And the inverse of Roth’s constant C is smaller or equal to K_{\varepsilon}^{\frac{1}{1+\varepsilon}}3k(\frac{p_{1}}{q_{1}})^{\frac{3\varepsilon}{1+\varepsilon}} Proof: Let \varepsilon be the ABC \varepsilon and K_{\varepsilon} the ABC constant. The ABC conjecture implies \frac{p_{n}^{3}}{gcd}\leq K_{\varepsilon}Rad(p_{n}q_{n}k(d_{n}gcd^{-1}))^{1+\varepsilon} This impiles p_{n}^{3}\leq K_{\varepsilon}(p_{n}q_{n}d_{n}k)^{1+\varepsilon}\frac{gcd}{gcd^{1+1\varepsilon}} In the cubic case, Bombieri–Van der Poorten’s explicit continued-fraction formula (Bombieri and Van der Poorten (1975), Theorem 3, formula (13)) applies without the final error term, thanks to effective Liouville bounds; compare also their worked example for \sqrt[3]{2} (loc. cit., p. 152). In our case their algorithm is also always reduced after the first step. For n>=4 an error term can arise and the method can not always be started with the first step. This is the reason, why we restrict ourselves to 3rd roots here. From ( Bombieri & Van der Poorten 1975 , p.151, Theorem 3, formula (13)) , we have the inequality d_{n}\leq\frac{3p_{n}^{2}}{q_{n}b_{n+1}} Using that inequality we get b_{n+1}\leq K_{\varepsilon}^{1/(1+\varepsilon)}3kp_{n}^{\frac{3\varepsilon}{1+\varepsilon}}({\frac{gcd}{gcd^{3+3\varepsilon}})^{1/(1+\varepsilon)}} Roth’s theorem implies the existence of a constant C depended only on \varepsilon_{Roth} and 3rd root of k with b_{n+1}\leq\frac{q_{n}^{\varepsilon_{Roth}}}{C} So as relationship between the epsilons we get