REVIEW 5 major objections 4 minor 9 references
On the Collatz Conjecture: Topological and Ergodic Approach
T0 review · 5 major / 4 minor · reviewed 2026-08-03 · deepseek-v4-flash
Pith's one-line read An ergodic dictionary claims finitely many cycles for Collatz-type maps by moving the problem into thermodynamic formalism.
desk verdict Main theorem is false — f0(n)=n puts infinitely many fixed points in the paper's own family — and the proof's key lemma rests on an invalid inference. 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 key object is the coarsest topology T on N containing {{n,2n}: n in N}, together with its Borel sigma-algebra. It makes the Collatz map measurable, turns every recurrent point into a periodic point via the open set {n,2n}, and makes every periodic orbit open. The load-bearing identity is the pressure formula P(phi)=sup_mu integral phi dmu, since Lemma 12 and Remark 13 force zero entropy for every invariant measure; then Lemma 14 equates finiteness of the set of periodic orbits with every continuous integrable potential having an equilibrium state, and Lemma 16 equates uniqueness of the cycle with uniqueness of equilibrium states for bounded continuous potentials.
What would settle it
Compute cycle-averages under the continuous potential phi(n)=n for a hypothetical infinite family of distinct periodic orbits of a Syracuse map; if the averages stay bounded and no average attains the supremum, then an equilibrium state for phi would not exist and Lemma 14's converse would be refuted. More directly, exhibiting even one new Collatz cycle beyond {1,2,4} would refute the paper's finite-cycles claim.
Extended reading notes
Core claim
On the paper's own terms, the discovery is Theorem B: every map of the form f(n)=f0(n) for odd n, n/2 for even n — in particular the Collatz, Syracuse, and Baker maps — has only finitely many periodic orbits. The route is a 'dictionary': after defining the coarsest topology containing all two-point sets {n,2n}, the Borel structure makes f measurable, periodic orbits are open, every f-invariant probability is supported on a periodic orbit, and entropy vanishes on all of them. Pressure therefore reduces to a supremum of integrals over cycle measures; the author proves that every continuous potential integrable with respect to all invariant probabilities attains its pressure, then invokes Lemma
Load-bearing premise
The load-bearing premise is Lemma 14's converse: that if every continuous potential integrable with respect to all invariant measures has an equilibrium state, then there can be only finitely many periodic orbits; the proof assumes infinitely many cycles would let some continuous unbounded potential take unbounded integrals on the cycle measures, but that potential is never constructed and the implication is not proved.
Editorial extensions
If this is right
- If Theorem B is right, the Collatz function has finitely many cycles even though the full conjecture (one cycle, no divergence) remains open.
- The same conclusion applies to all Syracuse maps and to the Baker map, so the odd part an+b is claimed not to affect finiteness, only the number of cycles.
- The full Collatz conjecture becomes a uniqueness problem: prove that the unique equilibrium state for every bounded continuous potential is the one supported on {1,2,4}.
- All invariant measures have zero entropy, so candidate equilibrium states can be checked by comparing cycle averages of potentials, a finite-dimensional optimization once cycles are known.
- If the asserted no-divergent-orbits claim also holds, the remaining gap to the full conjecture is exactly uniqueness of the cycle.
Reading between the lines
- Lemma 14's converse is the hinge: the paper does not actually construct a continuous unbounded potential whose integrals over infinitely many cycles diverge, so Theorem 18 inherits an unproven equivalence. A reader should treat finite-cycles as conditional on that dictionary step.
- A concrete way to test the dictionary: try to build infinitely many disjoint cycles for a Syracuse map and check whether the cycle averages of a simple potential such as phi(n)=n stay bounded; if they stay bounded without attaining the supremum, Lemma 14's converse would fail.
- The no-divergent-orbits assertion appears stronger than anything proved in the displayed theorems; as printed, the proof would need an additional argument ruling out infinite orbits that never become periodic, not just finiteness of cycles.
- If the equivalence were repaired, the same dictionary could give a numerical strategy: search over potentials phi; the cycle with maximal average would be the cycle selected by the dynamics, making uniqueness testable by comparing averages.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper studies maps of the form f(n)=n/2 for even n and f(n)=f0(n) for odd n, with the Collatz map as a special case. It introduces a coarse topology T on N generated by sets {n,2n} and its Borel σ-algebra, and claims that, in this setting, every recurrent point is periodic, that finiteness of periodic orbits is equivalent to existence of equilibrium states for all continuous integrable potentials, and that uniqueness of periodic orbits is equivalent to uniqueness of equilibrium states for all bounded continuous potentials. On this basis, Theorem B claims finiteness of cycles for all maps in the family (∗), and the standalone abstract further claims that there are no divergent orbits. The paper also sketches an application to Baker and Syracuse maps.
Significance. The idea of translating Collatz-type cycle finiteness into thermodynamic-formalism language is appealing, and the construction of a coarse topology on N is original. If Lemma 14 and Theorem 18 were correct, the finiteness of cycles for the family (∗) would be a notable result. However, the central proof does not hold together: the converse of Lemma 14 is not established, Theorem 18's equilibrium-state argument fails at a key step, and the abstract contains a stronger unproved claim about divergent orbits. The paper does not provide machine-checked proofs or reproducible code. In its current form the manuscript does not support its advertised conclusions.
major comments (5)
- [§3.10, Lemma 14] The converse direction of Lemma 14 is the load-bearing step for Theorem 18, but the proof is invalid. It states: 'the existence of an equilibrium state implies that for every continuous and unbounded potential φ we have sup ∫φ dμ < ∞.' This implication is false. In the family (∗) take f0(n)=n; then every odd n is fixed, so there are infinitely many periodic orbits. The constant potential φ≡0 (or φ≡1 if 0∉N) is continuous and integrable and every invariant measure is an equilibrium state, while φ(n)=oddpart(n) is continuous and satisfies sup∫φ dμ=∞. Thus existence of an equilibrium state for one potential implies nothing about unbounded potentials. To prove the stated biconditional one must assume every continuous potential has an equilibrium state and then derive a contradiction, e.g. from Lemma 10's orbit-sum potential; the proof does not do this. As written, Lemma 14 is unsupported and
- [§4, Theorem 18] The proof of Theorem 18 does not produce an equilibrium state for arbitrary φ. After showing (or attempting to show) sup_i ∫φ dδ_i < ∞, the paper asserts: 'Once the values of the integrals are natural numbers, there must exist some δ_i such that it is an equilibrium state.' But ∫φ dδ_i = (1/#O_i)∑_{x∈O_i} φ(x) is a rational number, not necessarily natural, and a bounded set of rationals need not attain its supremum. Thus no maximum is guaranteed. The preceding boundedness argument is also incomplete: if ∫φdδ_i→∞, one can choose weights a_i = 1/(2^j M_j) on a subsequence with M_j tending to infinity and still have ∑ a_i M_j <∞, so the convex-combination contradiction does not follow as stated. Without an equilibrium state for every continuous φ, Lemma 14 cannot be invoked.
- [Abstract and §1–§4] The standalone abstract claims 'we prove that there is no divergent orbit at all, which is a significant advance to the conjecture itself, proving half of the conjecture.' The body of the paper, including its own abstract, only attempts to prove finiteness of cycles; no theorem about divergent orbits is stated or proved. This is a material overclaim and should be corrected. If the intended contribution is only finiteness of cycles, the abstract must not assert the stronger statement.
- [§3.11, Lemma 16 and Remark 17] Lemma 16 assumes that the characteristic function χ_O of the union of all periodic orbits is continuous because 'every orbit is an open subset.' Openness of an arbitrary periodic orbit for a map in (∗) is not established: a cycle element y may satisfy f(y)=ay+b, which is neither 2y nor y/2, so the basic open sets {n,2n} do not automatically cover the orbit as a union contained in O. Thus the continuity of χ_O is an unproved assumption, and the uniqueness equivalence is unsupported. Remark 17's caveat that continuity depends on X does not repair this gap.
- [§3.4, Lemma 5] The preimage computation in Lemma 5 is incorrect. It claims f^{-1}({(3n+1)/2,3n+1}) = {n,3n+1}. But for the Collatz map, the even number 6n+2 also maps to 3n+1, so 6n+2 belongs to the preimage. Hence the displayed equality and the resulting proof that each odd singleton is open fail as written. This is a concrete topological error, though not the main load-bearing step.
minor comments (4)
- [§2] Typo: 'Themodynamic Formalism' should be 'Thermodynamic Formalism.'
- [§5 and Theorem B] The name 'Baker map' is repeatedly written as 'Bake map.' Also, Theorem B's phrase 'it may exists no periodic orbit' should read 'there may exist no periodic orbit.'
- [§3.9, Lemma 12] The affine identity h((1−t)μ+tν)=(1−t)h(μ)+t h(ν) is used without comment. This is true for convex combinations of invariant measures under standard conditions, but the paper should state the relevant setting, especially since the space is noncompact and nonmetrizable.
- [§3.8, Lemma 11] The proof of the ergodic decomposition claims the support of μ is a forward-invariant subset. This is not automatic for measurable non-continuous maps in an arbitrary topological space; the argument should justify it or avoid relying on it.
Circularity Check
No significant circularity; the derivation is an attempted theorem-proof, with the serious defects being unsupported lemmas and an overbroad conclusion rather than self-referential reasoning.
full rationale
The paper's central inference is Theorem 18, which invokes Lemma 14's claimed equivalence between finiteness of periodic orbits and the existence of equilibrium states for every continuous integrable potential. This equivalence is presented as a theorem to be proved, not as a definition or as a fitted parameter. The proof structure in Theorem 18 — assume infinitely many cycles, attempt to prove the equilibrium-state condition, then apply the converse direction of Lemma 14 to obtain a contradiction — is a standard proof-by-contradiction schema. It would be circular only if the converse of Lemma 14 were itself proved from Theorem 18, or if Lemma 14 encoded finiteness by construction; neither is the case. Lemma 10's special potential does depend on periodic-orbit data, but it is used to establish the non-trivial direction of the equivalence, not to fit the conclusion. There are no load-bearing self-citations: the references are standard external works, and no uniqueness theorem is imported from the authors' own prior work. The main weaknesses are mathematical. Lemma 14's converse is asserted with a one-line sketch rather than a complete construction. The step in Theorem 18 reading 'Once the values of the integrals are natural numbers' is false in general, since an integral over an ergodic measure supported on a periodic orbit is an average and need not be an integer. Moreover, the family (∗) includes f0(n)=n, for which every odd n is a fixed point, so Theorem B's assertion of at most finitely many periodic orbits fails as stated. These are serious correctness risks and indicate that the paper's proof is unsound, but they are not circularity: the conclusion is not identical to an input by construction, and the derivation chain does not reduce to a self-citation or to a fitted value. Under the rubric, the honest finding is no significant circularity.
Assumptions & free parameters
assumptions (4)
- domain assumption Poincaré recurrence applies and yields topological recurrence a.e. for any f-invariant probability (Lemma 2).
- domain assumption Ergodic decomposition holds for the non-metrizable space (N,T) and any invariant probability is a convex combination of measures supported on periodic orbits (Lemma 11).
- ad hoc to paper The characteristic function of the union of all periodic orbits is continuous in T (Lemma 16).
- domain assumption Pressure equals sup∫ϕ dµ because all invariant measures have zero entropy (Lemmas 14 and 16).
invented entities (2)
-
Key topology T generated by sets {n,2n}
-
Dictionary between Collatz cycles and equilibrium states
Cite this review
Pith. "Pith review of On the Collatz Conjecture: Topological and Ergodic Approach." pith.science (2026). https://pith.science/paper/FUEVX4JQ
@misc{pith2026260103297,
author = {Pith},
title = {Pith review of: On the Collatz Conjecture: Topological and Ergodic Approach},
year = {2026},
howpublished = {\url{https://pith.science/paper/FUEVX4JQ}},
note = {Machine review of arXiv:2601.03297}
}
abstract
We study a class of maps having the Collatz function (famously related to the Collatz Conjecture) as an example, under topological and ergodic perspectives, including an approach with thermodynamic formalism. By introducing a key topology and its Borel $\sigma$-algebra we show that recurrence implies periodicity. Moreover, we establish that if every continuous potential with finite pressure possesses some equilibrium state then we have either finiteness of cycles or infinitely many cycles sharing the same period. The existence of some continuous potential with no equilibrium state is equivalent to the unboundedness of periods of cycles. The uniqueness of periodic orbits is equivalent to the uniqueness of equilibrium state for every bounded and continuous potential. We also prove that we have either infinitely many cycles or no divergent orbits. Additionally, by using the dictionary established in the paper, we prove that there is no divergent orbit at all, which is a significant advance to the conjecture itself, proving half of the conjecture. Finally, we apply our technique to the Baker and Syracuse maps, obtaining a similar result for this general class of important maps.
Reference graph
Works this paper leans on
-
[1]
J. H. Conway. Unpredictable iteratios. InProc. 1972 Number Theory Conf., pages 49–52. Univ. Colorado, Boulder, CO, 1972
1972
-
[2]
Einsiedler and T
M. Einsiedler and T. Ward.Ergodic Theory: with a view towards Number Theory. Graduate Texts in Mathematics, Vol. 259. Springer, London, 2011
2011
-
[3]
Krasikov and J
I. Krasikov and J. C. Lagarias. Bounds for the number of integers in the 3x+ 1 problem using difference inequalities.Acta Arithmetica, 109(3):237–258, 2003
2003
-
[4]
J. C. Lagarias. The 3x+1 problem and its generalizations.The American Mathematical Monthly, 92(1):3–23, 1985. 9
1985
-
[5]
J. C. Lagarias (Ed.).The Ultimate Challenge: The3x+ 1Problem. American Mathematical Society, Prov- idence, RI, 2010
2010
-
[6]
Oliveira and M
K. Oliveira and M. Viana.Foundations of Ergodic Theory. Cambridge Studies in Advanced Mathematics, Vol. 151. Cambridge University Press, 2016
2016
-
[7]
T. Tao. Almost all orbits of the Collatz map attain almost bounded values.Forum of Mathematics, Pi, 10:e12, 2022
2022
-
[8]
Walters.An Introduction to Ergodic Theory
P. Walters.An Introduction to Ergodic Theory. Graduate Texts in Mathematics, Vol. 79. Springer-Verlag, New York, 1982
1982
Show all 9 references
-
[9]
G. J. Wirsching.The Dynamical System Generated by the3n+ 1Function. Lecture Notes in Mathematics, Vol. 1681. Springer-Verlag, Berlin, 1998. Universidade Federal de Alagoas, A v. Beira Rio, s/n, 57200-000 Penedo, Brazil. Email address:jemsmath@gmail.com
1998
Reviewed August 3, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.