Keyboard shortcuts

Press ← or → to navigate between chapters

Press ? to show this help

Press Esc to hide this help

Raney Zero-Run Coefficient Families

Abstract

For every admissible Raney sequence, nonzero residues occur arbitrarily far out and every actual maximal zero-run length lies in finitely many affine prime-power families.

Eu, Huang, and Kao’s Conjecture 7.1 asks for finite affine prime-power families containing the tied left-to-right record values of the actual zero-run sequence. The theorem here keeps the literal integral Raney quotient and proves the stronger statement for every actual maximal finite zero interval. Its separate unbounded-support conjunct rules out an infinite zero tail. It neither truncates to finite prefixes nor asserts that every coefficient-family term is realized.

Definition 1.1 (The literal integral Raney number).

Formalization. D5/S1/Recurrence/Raney/ZeroRunFamilies.raneyNumber (✓ std3).

Citation. Sen-Peng Eu, Zai-Ting Huang, and Louis Kao (2026). Zero-Run Spectra of the (3,2) Raney numbers Modulo Primes. URL: https://arxiv.org/html/2609.25742v1.

Commentary.

For natural k,r,n, raneyNumber(k,r,n) is the natural quotient rbinomial(kn+r,n)/(kn+r), exactly as in the source. The theorem assumes positive k and r. Its proof establishes the needed exact division identity before casting to ZMod(p); it does not divide by kn+r in the residue field.

Theorem 1.2 (All actual Raney zero runs have finite coefficient data).

Proof. Machine-checked in Lean as D5/S1/Recurrence/Raney/ZeroRunFamilies.raney_zero_run_coefficient_families (✓ std3). ∎

Resolves. Problems/raney-conjecture-7-1-zero-run-record-families (proved) by D5/S1/Recurrence/Raney/ZeroRunFamilies.raney_zero_run_coefficient_families.

Source. Repository-derived.

Acknowledgement. Sen-Peng Eu, Zai-Ting Huang, and Louis Kao (2026). Zero-Run Spectra of the (3,2) Raney numbers Modulo Primes. URL: https://arxiv.org/html/2609.25742v1.

Acknowledgement. Yann Bugeaud, Dalia Krieger, and Jeffrey Shallit (2009). Morphic and Automatic Words: Maximal Blocks and Diophantine Approximation. URL: https://arxiv.org/abs/0808.2544v2.

Commentary.

Let k,r be positive and p prime with p not dividing kr. First, for every cutoff there is n at least that cutoff with raneyNumber(k,r,n) nonzero modulo p. Second, one finite set C of integer triples (a,b,c), all with c>0, is chosen before arbitrary first and last. Every actual maximal zero interval satisfies c(last+1-first)=a*p^m+b for some member of C and some natural m. Thus all tied record lengths belong to the required finite union.

The proof first derives a guarded adjacent-binomial identity. At n=0 the predecessor term is zero; for n>0 it proves that (k-1)choose(kn+r-1,n-1) is bounded by choose(k*n+r-1,n), and their natural difference equals the literal quotient. This supplies both integrality and the later residue readout.

Set A=max(k-1,r-1), State=Fin(A+1) x Bool, and Alphabet=State -> ZMod(p). The false Boolean component evaluates choose(kn+a,n). The true component evaluates the guarded predecessor choose(kn+a,n-1), with value zero at n=0. For a base-p digit d, the next index (k*d+a)/p remains in Fin(A+1). Lucas’ theorem supplies the normal transition, the positive-digit predecessor transition, and the zero-digit borrow transition. The resulting list of p digit transforms is a p-uniform morphism, and the complete evaluation vector is its pointwise fixed word, including leading zeros.

At state r-1, the ordinary value minus (k-1) times the predecessor value is exactly the literal Raney residue. For sufficiently large j with r-1<p^(j-1), the Lucas readout at n=p^j reduces to k modulo p. Since p does not divide k*r, p does not divide k, so these arbitrarily large values are nonzero. Applying the generic actual-block coefficient theorem with Delta={0} proves the second conjunct.

The two conjuncts have distinct roles: arbitrarily large nonzero residues exclude a terminal infinite zero run, while one finite set C, chosen before the interval endpoints, covers all actual maximal finite zero intervals. This is a containment statement only: it does not assert that every coefficient triple or exponent yields an actual interval, and it supplies no converse realization theorem.

References