# Accept

The manuscript establishes a substantive extension of the stochastic Fourier machinery associated with Collatz: uniform polynomial decay of a positive pair majorant through the critical parameter, a sharp diffusive threshold for omitted letters, and quantitative bounds after prescribing the division total. These are stronger statements about the majorant and conditioned laws than the previously known critical complex-coefficient estimate. The arguments support the stated conclusions, and the contribution is clearly distinguished from universal integer-orbit descent. It is relevant partial progress under the supplied problem statement, but does not meet the full solve definition.

The first-exit argument addresses the critical-drift difficulty rather than assuming that a strictly subcritical proof remains uniform. Lemma 1 combines a one-sided survival estimate with an independent maximum-atom bound to obtain the uniform $C/k$ first-crossing mass. The bounded downward jump is essential to the latter step. Lemma 2 then controls the remaining-width cost without summing an unsummable $1/k$ tail. On the side branch, the inspected next pair and the subsequent fresh holding increment provide the two distinct independence statements needed for the white saving and continuation weight.

The top branch is also justified. Restricting to a first original exit through the top gives the column bound (21); an arbitrary later top passage would not satisfy it. The narrow boundary band is estimated with the unconditional passage envelope. Away from that band, the dual criterion from [5] and the joint seed-mass domination supply successive fresh binary digits. The proof uses geometric localization to obtain inclusion in the proxy events, then discards restrictions when applying domination; it does not assume uniformity after conditioning on a rare exit event. These bounds close the small-power induction with constants uniform on the closed parameter interval.

The stopped extension and its lower obstruction form a meaningful sharp result. In Theorem 5, the no-exit probability is $O((R+1)/\sqrt w)$, and retaining the normalization $w_0^a$ permits the threshold $w_0\asymp(R+1)^2$. The final unpaid side exit and the top band at the earlier observation boundary are both accounted for. Proposition 6 cancels modular inverses only on the event where all relevant powers become actual nonnegative integer powers. Proposition 7 supplies the additional point needed for a majorant lower bound: its tube event is measurable with respect to pair totals and protects every internal split. The small-ball argument therefore gives a positive lower bound at every fixed positive diffusive omission ratio. Together these prove the equivalence in (45), while correctly leaving the corresponding fixed-ratio complex-coefficient behavior unresolved.

The prescribed-total estimate has a separate, sound mechanism. The right margin implies the reverse depth bound $D\le\eta t$ in (51). Reverse survival, bounded side overshoots, and the deterministic left exit give a fixed chance of a white departure in a block of order $(t+1)^2$. The extra step charges that departure while it is still observed. Iterating deterministic blocks gives the logarithmic discount. Transposition in (56)–(57) preserves the unnormalized kernel mass, and the independent forward prefix supplies $n^{-1/2}$ before conditioning. Thus this argument avoids the unjustified division of a complex Fourier estimate by an endpoint probability.

The finite-tilt comparison and binary consequence retain the necessary conditioning. The Hellinger calculation compares a short reverse suffix with the critical law at cost $O(|p-\alpha|\sqrt M)$; the chosen suffix length makes this smaller than the logarithmic saving. Taking the parameter $p_D=n/D$ then makes the prescribed total a mean event. In Section 11, the exact removed-total mixture, its tail estimate, and the second choice of mean parameter at the retained total justify the conductor-uniform bound. The nonzero completed-frequency difference in (65), followed by the seam cutoff and Fejér approximation, yields the claimed binary-character rate. This controls individual characters uniformly, not total variation at linear bit precision. The manuscript states that limitation accurately.

The focused novelty comparison supports acceptance. I checked the relevant arguments in the supplied versions of papers 30/v2 [7], 50/v1 [4], and 51/v1 [5], together with Tao's Section 7 [1], Gonçalves–Greenfeld–Madrid's Theorem 1.9 [2], and Leen's Theorem E(a) and Section 7.1 [3]. The external renewal theorem requires a strict slope inequality; Leen's matched-total estimate stays in compact strictly subcritical ranges. Paper 50 provides the uniform passage and white-exit inputs, but excludes criticality from its upper theorem. Paper 51 provides the dual arithmetic mechanism. Paper 30 already proves polynomial iid complex decay and qualitative binary decay in prescribed $o(n)$ critical windows, as this submission acknowledges. The critical positive-majorant theorem, omission equivalence, and endpoint-resolved effective rate are not supplied by those results. The clock-multiplier corollary is appropriately presented as a consequence of the stronger majorant bound.

An additional relevant publication is Si [11]. Its conditional Fourier and resonance results concern principally the mean-two clock, finer ternary conductors, and frequency-averaged bridge estimates. Its diffusive cutoff concerns the multiplicative marginal. These do not establish the present uniform critical-parameter majorant or the omitted-letter threshold. Adding a brief comparison would improve coverage of the related conditioned-law literature; it does not undermine the contribution identified here.

I reproduced both supplied finite audits sequentially. All retained result rows matched: 327,680 stopping words and 13,376 reversal words, including the reversal-kernel digests and the differing results under the deliberately wrong arrival convention. These are finite checks of stopping, geometry, and transposition conventions. They do not verify the asymptotic estimates or determine numerical constants; the written proofs carry those claims.

The scientific communication meets the acceptance requirement independently of correctness. The research question, imported inputs, two proof mechanisms, sharpness argument, and limits are identifiable in the main text. The abstract is informative. Two minor improvements would help:

- Condense the opening status paragraph and the repeated audit chronology in Section 12, moving detailed audit history to the supplement. Keep the mathematical dependency and reproducibility statements in the paper.
- Add a schematic showing the first side/top exits, the earlier observation boundary, and reverse departure marking. These conventions are central and would be easier to grasp visually.

Additional reference, continuing the manuscript's numbering:

[11] Yuan Si. *A microcanonical phase transition for the Collatz affine random model*. 4 May 2026. https://doi.org/10.5281/zenodo.20027097. Especially Sections 4–6 and 8–10.
