Conditional Results on the Collatz Conjecture: A Compactification Criterion for Cycle Uniqueness, a Quantitative Divergence Bound, and the Non-Viability of KAM Methods

We establish four results on the Collatz conjecture. Theorem 1: cycle uniqueness follows from a single topological condition (the Compactification Condition) on the Alexandroff compactification of a topology introduced by Santana (arXiv:2601.03297v3, 2026), via a parity argument verified for all cycle structure parameters. Theorem 2: a quantitative divergence bound showing that divergence requires the 2-adic valuation v₂(3n+1) = 1 at asymptotic frequency exceeding p* ≈ 0.7075 among odd steps, when the natural rate is 0.5. Theorem 3: if both conditions hold, the Collatz conjecture is true. Theorem 4 (unconditional): the Syracuse map is non-injective on odd integers, with an explicit infinite family S((2^(2m+2)−1)/3) = 1 for all m ≥ 0, proving that no KAM-type theorem requiring invertibility can apply to the Collatz dynamical system. The paper identifies a specific gap in Santana's claimed proof (the transfer of closedness from the base topology to its compactification in the non-Hausdorff setting), isolates it as a named condition, and provides the conditional framework and divergence analysis that Santana's approach lacks.