RESEARCH / Focused target
Expansions of Narayana-Weighted Dyck Paths in an Even Strip
RELEASED FOUNDATIONS / PROPOSED CONNECTION
Research connections
All source anchors (1)
- D5/S3/Combinatorics/NarayanaStrip/CiglerStripExpansion.resultAbsent from this graph
Problem
The question
Johann Cigler, Some sequences and number triangles which are related to Narayana polynomials and to q-Narayana polynomials for q=-1, arXiv:2608.03363v2, Section 4, Conjecture 3 (Conjecture 2 in v1): for m ≥ 1 and n ≥ 0,
C_{n+1}^{(2m)}(t) = Σ_{j=0}^{⌊n/2⌋} C_j^{(m−1)} binom(n, 2j) t^j (1 + t)^{n−2j},
c_{n+1}^{(2m)}(t) = Σ_{j=0}^{⌊n/2⌋} (−1)^j C_j^{(m−1)} binom(⌊n/2⌋, j) t^j (1 + t)^{n−2j}.
Here C_n^{(h)}(t) and c_n^{(h)}(t) are the weighted counts of Dyck paths of semilength n in the strip 0 ≤ y ≤ h, a down-step to height k having weight τ_k with τ = (1, t, 1, t, …) and τ = (1, t, −1, −t, …) respectively, and C_j^{(m−1)} is the number of Dyck paths of semilength j in the strip of height m − 1.
Motivation
Our foothold
The theorem D5/S3/Combinatorics/NarayanaStrip/CiglerStripExpansion.result establishes both identities for every
m ≥ 1 and n ≥ 0 in ℤ[t].
Gap
Missing bridges
Pre-registration issue 12523 records the literature screen: the arXiv record has v1 and v2, both listing the statement as a conjecture, and searches of the arXiv identifier on the web and on GitHub returned no proof outside this repository. This is a bounded negative finding.
Route
Proposed approach
- Remove the first and last steps of a path of semilength n + 1 in the strip of height 2m and pair the remaining 2n steps. Sampling at odd times with coarse height a turns UU and DD into up and down steps and UD, DU into two kinds of flat steps; this is a weight-preserving bijection with weighted Motzkin paths of length n in the strip of height m − 1, the steps DU at coarse height 0 and UD at coarse height m − 1 included.
- Deleting the flat steps leaves a Dyck skeleton of semilength j, counted by C_j^{(m−1)}; the flat steps fill the 2j + 1 gaps of the skeleton with colours.
- With Narayana weights a skeleton with a fixed gap filling, summed over the colours of its flat steps, has weight t^j (1 + t)^{n−2j}, and the gap fillings are counted by binom(n, 2j).
- With the q = −1 weights the sign of a flat step depends on the parity of the index of its gap. An involution that moves a block of flat steps across one skeleton step into the neighbouring gap keeps the skeleton and the colours and reverses the sign of every non-fixed filling; the fixed gap tuples are counted by binom(⌊n/2⌋, j) and carry the sign (−1)^j.
Falsifier
What would falsify this route
The statement would fail if the pairing missed a boundary move at coarse height 0 or m − 1, or if the involution failed to preserve weight or to reverse sign outside its fixed points.
Evidence
Evidence to collect
An independent referee implementation checked both identities and every intermediate bijection exhaustively for m ≤ 6 and n ≤ 16, and both identities coefficientwise by a step-by-step polynomial recursion for m ≤ 10 and n ≤ 40.
Triage
Scope assessment
theorem; the statement is Conjecture 3 of arXiv:2608.03363v2 and is quantified over every m ≥ 1 and n ≥ 0.
- Proved (formalized): both expansion identities for every m ≥ 1 and n ≥ 0.
- Proved (formalized): the pairing of steps is a weight-preserving bijection between strip-2m Dyck paths of semilength n + 1 and weighted Motzkin paths of length n in the strip of height m − 1; the expansions are the skeleton decomposition of these Motzkin paths.
- Open: an analogous closed expansion for odd heights 2m + 1, which the source does not state.
- Open: Conjecture 2 of the same section (the product formula for heights 4m and 4m + 1) is a separate statement and is not implied by these expansions.
ASSUMED-UNVERIFIED
Unverified assumptions
The literature screen is limited to the arXiv record, web and GitHub searches and the repository checks recorded above.