Recognition: unknown
A Ceiling Continued Fraction Approach to the ErdH{o}s-Straus Conjecture: Heuristic finiteness of counterexamples
Pith reviewed 2026-05-08 15:56 UTC · model grok-4.3
The pith
A ceiling continued fraction approach provides heuristic evidence that the Erdős-Straus conjecture has only finitely many counterexamples.
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
We introduce the Ceiling Continued Fractions (FCT) framework for constructing three-term Egyptian fraction representations in the Erdős-Straus conjecture. The approach exploits divisor structures of shifted integers p+i rather than congruence-based techniques. Computational tests on 10^9 primes in ranges around 10^17 and 10^52, and 10^7 primes around 10^131, show no counterexamples with very small search depth. We derive a super-polynomial upper bound on the failure probability; its convergence, together with the Borel-Cantelli lemma, provides heuristic evidence that counterexamples, if any exist, form a finite set.
What carries the argument
The Ceiling Continued Fractions (FCT) framework for constructing three-term Egyptian fraction representations by exploiting divisor structures of shifted integers rather than congruences.
If this is right
- Extensive searches of 10^9 primes near 10^17 and 10^52 plus 10^7 near 10^131 find no counterexamples.
- The super-polynomial bound ensures the sum of failure probabilities over primes converges.
- Borel-Cantelli then implies only finitely many counterexamples occur almost surely.
Where Pith is reading between the lines
- The FCT method might adapt to other fixed-numerator Egyptian fraction problems.
- An effective version of the bound could reduce the conjecture to a finite computational check.
- Similar divisor-exploiting continued fraction techniques could bound exceptions in related Diophantine representation questions.
Load-bearing premise
The super-polynomial upper bound on failure probability derived from the FCT framework is tight enough and that the failure events across primes satisfy the conditions needed for the Borel-Cantelli lemma to conclude finiteness.
What would settle it
A single counterexample prime p for which 4/p has no three-term Egyptian fraction representation, or a demonstration that the failure probability bound fails to be super-polynomial.
read the original abstract
We introduce the Ceiling Continued Fractions (FCT) framework for constructing three-term Egyptian fraction representations in the Erdos-Straus conjecture. The approach exploits divisor structures of shifted integers p+i rather than congruence-based techniques. Computational tests on 10^9 primes in ranges around 10^17 and 10^52, and 10^7 primes around 10^131, show no counterexamples with very small search depth. We derive a super-polynomial upper bound on the failure probability; its convergence, together with the Borel-Cantelli lemma, provides heuristic evidence that counterexamples, if any exist, form a finite set.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The manuscript introduces the Ceiling Continued Fraction (FCT) framework for constructing three-term Egyptian fraction representations of 4/p for primes p in the Erdős-Straus conjecture. It reports no counterexamples found in computational searches over 10^9 primes near 10^17 and 10^52 and 10^7 primes near 10^131 using small search depths, derives a super-polynomial upper bound on the FCT failure probability, and invokes the Borel-Cantelli lemma to provide heuristic evidence that any counterexamples form a finite set.
Significance. The large-scale computational searches constitute a notable strength, extending verification to primes around 10^131. If the super-polynomial bound on failure probability is rigorously derived and independent of internal modeling choices, and if FCT captures all representations, the Borel-Cantelli application would supply useful heuristic support for the conjecture having at most finitely many exceptions. The approach via divisor structures of p+i offers a fresh constructive perspective distinct from congruence methods.
major comments (3)
- [FCT framework and probability bound derivation] The manuscript does not establish completeness of the FCT framework: no theorem shows that every possible triple (a,b,c) yielding 4/p = 1/a + 1/b + 1/c arises from some ceiling continued fraction expansion exploiting divisors of p+i. Without this, FCT failure controls only the success rate of this particular search procedure and does not equate to the non-existence of any representation, undermining the direct application of the failure-probability bound to counterexamples (central to the finiteness claim via Borel-Cantelli).
- [Section deriving the failure probability bound] The explicit construction and assumptions underlying the super-polynomial upper bound on failure probability are not provided in sufficient detail to verify independence from FCT modeling choices or to confirm the bound's tightness. This is load-bearing because the convergence needed for Borel-Cantelli rests on this estimate.
- [Application of Borel-Cantelli lemma] The conditions required for the Borel-Cantelli lemma (e.g., independence or pairwise independence of failure events across distinct primes, or the precise form of the sum of probabilities) are not verified explicitly for the sequence of primes considered.
minor comments (2)
- [Computational tests section] The precise search algorithm, including how search depth is determined, how candidate divisors are enumerated, and the exact ranges and selection criteria for the tested primes, should be stated to permit independent reproduction of the computational results.
- Clarify the notation for the ceiling continued fraction expansions and the precise definition of 'failure' in the FCT procedure to avoid ambiguity in the probability estimates.
Simulated Author's Rebuttal
We thank the referee for the careful reading and constructive comments on our manuscript. The points raised help clarify the scope of our heuristic approach. We respond to each major comment below and indicate the revisions we will make.
read point-by-point responses
-
Referee: The manuscript does not establish completeness of the FCT framework: no theorem shows that every possible triple (a,b,c) yielding 4/p = 1/a + 1/b + 1/c arises from some ceiling continued fraction expansion exploiting divisors of p+i. Without this, FCT failure controls only the success rate of this particular search procedure and does not equate to the non-existence of any representation, undermining the direct application of the failure-probability bound to counterexamples (central to the finiteness claim via Borel-Cantelli).
Authors: We agree that the FCT framework is a specific constructive procedure based on ceiling continued fractions that exploit divisors of p+i, and the manuscript does not contain (nor claim to contain) a completeness theorem asserting that every possible triple arises in this way. The failure probability bound and the computational searches therefore apply strictly to the success rate of this method. The Borel-Cantelli invocation is presented as heuristic evidence that counterexamples, if any, are finite, under the modeling assumption that the FCT method is representative of the general difficulty of finding representations. We will revise the text to state this limitation more explicitly and to frame the finiteness conclusion as heuristic rather than direct. revision: partial
-
Referee: The explicit construction and assumptions underlying the super-polynomial upper bound on failure probability are not provided in sufficient detail to verify independence from FCT modeling choices or to confirm the bound's tightness. This is load-bearing because the convergence needed for Borel-Cantelli rests on this estimate.
Authors: We will expand the relevant section to give the explicit construction of the bound, list the modeling assumptions (including the divisor distribution model and the treatment of ceiling choices), and demonstrate that the super-polynomial decay is independent of the specific internal parameters of the FCT search. This revision will make the derivation verifiable and will confirm that the resulting sum converges. revision: yes
-
Referee: The conditions required for the Borel-Cantelli lemma (e.g., independence or pairwise independence of failure events across distinct primes, or the precise form of the sum of probabilities) are not verified explicitly for the sequence of primes considered.
Authors: We will add an explicit paragraph verifying the application of the Borel-Cantelli lemma: we state the precise form of the probability sum (which converges by the super-polynomial upper bound), note that the events are treated as approximately independent for large, distinct primes (a standard heuristic assumption in this area), and clarify that the lemma is invoked heuristically rather than under a fully rigorous independence proof. revision: partial
Circularity Check
No circularity: derivation chain is independent of its target claim
full rationale
The manuscript introduces the FCT framework as a new constructive method based on ceiling continued fractions and divisor structures of p+i. It reports direct computational searches over large prime ranges that found no counterexamples, then states a derived super-polynomial upper bound on the probability that FCT fails to find a representation. The bound is combined with the standard Borel-Cantelli lemma to obtain a heuristic statement about finiteness. No quoted equation or step shows the bound reducing to a fitted parameter, a self-definition, or a self-citation chain whose only justification is internal to the paper. The derivation therefore remains self-contained; the heuristic applies to the success rate of the presented search procedure rather than being tautological with the conjecture itself.
Axiom & Free-Parameter Ledger
axioms (3)
- domain assumption The ceiling continued fraction construction yields a valid three-term Egyptian fraction representation for 4/p whenever the divisor condition on p+i holds.
- domain assumption A super-polynomial upper bound on the probability that a random prime fails the representation can be derived from the FCT analysis.
- standard math The Borel-Cantelli lemma applies directly to the sequence of failure events indexed by primes.
Reference graph
Works this paper leans on
-
[1]
Cram´ er,On the order of magnitude of the difference between consecutive prime numbers, Acta Arithmetica2(1936), no
H. Cram´ er,On the order of magnitude of the difference between consecutive prime numbers, Acta Arithmetica2(1936), no. 1, 23–46
1936
-
[2]
C. Elsholtz and T. Tao,The number of solutions of4/n= 1/x+ 1/y+ 1/z, Math. Comp.82 (2013), 1737–1773.https://arxiv.org/abs/1201.3173arXiv:1201.3173
-
[3]
Khrushchev,Orthogonal Polynomials and Continued Fractions, Encyclopedia of Mathe- matics and its Applications, vol
S. Khrushchev,Orthogonal Polynomials and Continued Fractions, Encyclopedia of Mathe- matics and its Applications, vol. 122, Cambridge University Press, 2008, p. 159
2008
-
[4]
Moree and J
P. Moree and J. Cazaran,On a claim of Ramanujan in his first letter to Hardy, Expositiones Mathematicae17(1999), 289–311
1999
-
[5]
N. J. A. Sloane,The On-Line Encyclopedia of Integer Sequences, Primes of the formx 2+840y2, Sequence A139665.https://oeis.org/A139665
-
[6]
R. C. Vaughan,On a problem of Erd˝ os, Straus and Schinzel, Mathematika17(1970), 193–198
1970
-
[7]
Ventas,Relaci´ on entre series infinitas, fracci´ ons continuas teito e constantes
A. Ventas,Relaci´ on entre series infinitas, fracci´ ons continuas teito e constantes. Aplicaci´ ons (Parte I), 2025.https://retallosdematematicas.blogspot.com/2025/04/ relacion-entre-series-infinitas.html Email addres: aventas.avp@gmail.com 10
2025
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.