OctoInformBeta

OpenAI preprint, September 2026 · §3. Ladder cuts and a coefficient lower bound · §3

§3 (setting): ladder pieces and seam intervals: One-sided piece measures from inventory streams and the product-measure identity for empty-reducing seam words at height h≥1h \ge 1 when p∈(1/2,1)p \in (1/2, 1)

Ladder piece laws and seam reduction probability factorization.

Hover over a blue or green term in the text, or a node in the graph, to find it in the other. Select it to read its declaration both in Lean and in words.

The Setting

Consider the Fortuin-Kasteleyn (FK) model on random planar maps with Potts parameter q>4q > 4, where t=qt = \sqrt{q} and p=t/(t+2)∈(1/2,1)p = t / (t + 2) \in (1/2, 1). Through the inventory encoding, a decorated planar map corresponds to a bi-infinite burger-order stream governed by pp.

Splitting the stream at the origin across a positive height creates backward left and forward right excursions that concatenate into a seam word. Given one-sided independent and identically distributed (i.i.d.) streams, what are the laws of the raw and shortened left pieces and of the right pieces, and how does the probability that the seam word reduces to an empty inventory express itself over independent sequences of those pieces?

Inventory Streams and Word Reduction

The set of Boolean burger types is ℬ={false,true}\mathcal{B} = \{\mathrm{false}, \mathrm{true}\}, and an inventory order o∈𝒪o \in \mathcal{O} is either a fixed order fixed(b)\mathrm{fixed}(b) for some b∈ℬb \in \mathcal{B} or the flexible order flex\mathrm{flex}.✦ A burger-order alphabet letter a∈𝒜a \in \mathcal{A} is either a burger token 𝖻b\mathsf{b}_b for b∈ℬb \in \mathcal{B} or an order token 𝗈o\mathsf{o}_o for o∈𝒪o \in \mathcal{O}.✦ Given p∈(1/2,1)p \in (1/2, 1), the letter mass function mp:𝒜→ℝm_p : \mathcal{A} \to \mathbb{R} is defined for each b∈ℬb \in \mathcal{B} by mp(𝖻b)=14,mp(𝗈flex)=p2,mp(𝗈fixed(b))=1−p4.m_p(\mathsf{b}_b) = \frac{1}{4}, \qquad m_p(\mathsf{o}_{\mathrm{flex}}) = \frac{p}{2}, \qquad m_p(\mathsf{o}_{\mathrm{fixed}(b)}) = \frac{1 - p}{4}.(1)✦ For p∈(1/2,1)p \in (1/2, 1), the letter distribution μp\mu_p on 𝒜\mathcal{A} is the probability mass function assigning probability mp(a)m_p(a) from (1) to each a∈𝒜a \in \mathcal{A}.✦ Given an index set II and p∈(1/2,1)p \in (1/2, 1), the i.i.d. inventory stream measure ℙpI\mathbb{P}_p^I on 𝒜I\mathcal{A}^I is the product probability measure μp⊗I\mu_p^{\otimes I} with marginal μp\mu_p; when I=ℕI = \mathbb{N}, write ℙp=ℙpℕ\mathbb{P}_p = \mathbb{P}_p^{\mathbb{N}}.✦

For any set XX, write X*X^* for the set of finite words over XX, with empty word ε\varepsilon. An inventory stack evaluation state e=(𝒐,𝒃)∈𝒪*×ℬ*e = (\mathbf{o}, \mathbf{b}) \in \mathcal{O}^* \times \mathcal{B}^* consists of a list 𝒐\mathbf{o} of pending orders and a list 𝒃\mathbf{b} of Boolean burger types stored freshest-first.✦ Given an evaluation state e∈𝒪*×ℬ*e \in \mathcal{O}^* \times \mathcal{B}^* and a letter a∈𝒜a \in \mathcal{A}, the single-step inventory transition step(e,a)∈𝒪*×ℬ*\mathrm{step}(e, a) \in \mathcal{O}^* \times \mathcal{B}^* is the evaluation state updating ee by aa, where an order consumes the freshest eligible burger. The evaluation eval(w,e)\mathrm{eval}(w, e) of a word w∈𝒜*w \in \mathcal{A}^* from e∈𝒪*×ℬ*e \in \mathcal{O}^* \times \mathcal{B}^* is defined recursively by eval(ε,e)=e\mathrm{eval}(\varepsilon, e) = e and eval(aw,e)=eval(w,step(e,a))\mathrm{eval}(aw, e) = \mathrm{eval}(w, \mathrm{step}(e, a)) for a∈𝒜a \in \mathcal{A} and w∈𝒜*w \in \mathcal{A}^*.✦ The word reduction red(w)\mathrm{red}(w) of w∈𝒜*w \in \mathcal{A}^* is the evaluation state eval(w,(ε,ε))\mathrm{eval}(w, (\varepsilon, \varepsilon)) starting from empty pending orders and an empty burger stack✦, and a word w∈𝒜*w \in \mathcal{A}^* is empty-reducing if red(w)=(ε,ε)\mathrm{red}(w) = (\varepsilon, \varepsilon).

Hitting Times and Piece Constructions

The letter increment Δ:𝒜→ℤ\Delta : \mathcal{A} \to \mathbb{Z} is the map assigning +1+1 to each burger letter 𝖻b\mathsf{b}_b (b∈ℬb \in \mathcal{B}) and −1-1 to each order letter 𝗈o\mathsf{o}_o (o∈𝒪o \in \mathcal{O}).✦ For x=(xi)i≥0∈𝒜ℕx = (x_i)_{i \ge 0} \in \mathcal{A}^{\mathbb{N}} and n∈ℕn \in \mathbb{N}, the letter increment sum Sn(x)∈ℤS_n(x) \in \mathbb{Z} is Sn(x)=∑i=0n−1Δ(xi),S_n(x) = \sum_{i=0}^{n-1} \Delta(x_i),(2)✦ with S0(x)=0S_0(x) = 0. Given a space Ω\Omega, an integer-valued process Z:ℕ×Ω→ℤZ : \mathbb{N} \times \Omega \to \mathbb{Z}, and a point ω∈Ω\omega \in \Omega, the first hitting time of zero τ0(Z)(ω)∈ℕ∪{∞}\tau_0(Z)(\omega) \in \mathbb{N} \cup \{\infty\} is the least n∈ℕn \in \mathbb{N} such that Z(n,ω)=0Z(n, \omega) = 0, or ∞\infty if no such nn exists.✦

For h,n∈ℕh, n \in \mathbb{N} and x∈𝒜ℕx \in \mathcal{A}^{\mathbb{N}}, the upward inventory path Uh(n,x)∈ℤU_h(n, x) \in \mathbb{Z} and downward inventory path Dh(n,x)∈ℤD_h(n, x) \in \mathbb{Z} are defined using (2) by Uh(n,x)=h−Sn+1(x),Dh(n,x)=h+Sn+1(x).U_h(n, x) = h - S_{n+1}(x), \qquad D_h(n, x) = h + S_{n+1}(x).(3) At height h∈ℕh \in \mathbb{N}, the first upward hitting time τh+(x)∈ℕ∪{∞}\tau_h^+(x) \in \mathbb{N} \cup \{\infty\} of x∈𝒜ℕx \in \mathcal{A}^{\mathbb{N}} is τ0(Uh)(x)\tau_0(U_h)(x) for UhU_h in (3).✦ The first downward hitting time τh−(x)∈ℕ∪{∞}\tau_h^-(x) \in \mathbb{N} \cup \{\infty\} of x∈𝒜ℕx \in \mathcal{A}^{\mathbb{N}} is τ0(Dh)(x)\tau_0(D_h)(x).

For a word w=(a0,…,am−1)∈𝒜*w = (a_0, \dots, a_{m-1}) \in \mathcal{A}^*, write rev(w)=(am−1,…,a0)\mathrm{rev}(w) = (a_{m-1}, \dots, a_0) for its reversal, with rev(ε)=ε\mathrm{rev}(\varepsilon) = \varepsilon. For x∈𝒜ℕx \in \mathcal{A}^{\mathbb{N}} and n∈ℕn \in \mathbb{N}, the revealed word ρn(x)∈𝒜*\rho_n(x) \in \mathcal{A}^* of length nn is rev(x0,…,xn−1)\mathrm{rev}(x_0, \dots, x_{n-1}), with ρ0(x)=ε\rho_0(x) = \varepsilon.✦ For a map T:𝒜ℕ→ℕ∪{∞}T : \mathcal{A}^{\mathbb{N}} \to \mathbb{N} \cup \{\infty\} and x∈𝒜ℕx \in \mathcal{A}^{\mathbb{N}}, the truncation T_(x)∈ℕ\underline{T}(x) \in \mathbb{N} of TT at xx is T(x)T(x) when T(x)<∞T(x) < \infty and 00 when T(x)=∞T(x) = \infty, and the stopped word WT(x)∈𝒜*W_T(x) \in \mathcal{A}^* of xx at TT is ρT_(x)+1(x)\rho_{\underline{T}(x)+1}(x).✦ The reversed right piece word WRrev(x)∈𝒜*W_R^{\mathrm{rev}}(x) \in \mathcal{A}^* of x∈𝒜ℕx \in \mathcal{A}^{\mathbb{N}} is Wτ1−(x)W_{\tau_1^-}(x), and for p∈(1/2,1)p \in (1/2, 1) the reversed right piece law νRrev(p)\nu_R^{\mathrm{rev}}(p) on 𝒜*\mathcal{A}^* is the pushforward measure ℙp∘(WRrev)−1\mathbb{P}_p \circ (W_R^{\mathrm{rev}})^{-1}.✦

For any probability measure ν\nu on 𝒜*\mathcal{A}^*, the natural pieces law ν⊗ℕ\nu^{\otimes \mathbb{N}} on (𝒜*)ℕ(\mathcal{A}^*)^{\mathbb{N}} is the infinite product probability measure with common marginal ν\nu.✦ Given a word sequence y=(wk)k≥0∈(𝒜*)ℕy = (w_k)_{k \ge 0} \in (\mathcal{A}^*)^{\mathbb{N}} and n∈ℕn \in \mathbb{N}, the concatenated piece prefix Πn(y)∈𝒜*\Pi_n(y) \in \mathcal{A}^* is defined recursively by Π0(y)=ε\Pi_0(y) = \varepsilon and Πn+1(y)=Πn(y)wn\Pi_{n+1}(y) = \Pi_n(y) w_n.✦ For h∈ℕh \in \mathbb{N}, the piece seam event Eh⊂(𝒜*)ℕ×(𝒜*)ℕE_h \subset (\mathcal{A}^*)^{\mathbb{N}} \times (\mathcal{A}^*)^{\mathbb{N}} is the set of pairs (y1,y2)(y_1, y_2) such that Πh(y1)Πh(y2)\Pi_h(y_1)\,\Pi_h(y_2) is empty-reducing.✦

Backward Survivor Parsing

At step n∈ℕn \in \mathbb{N} of x∈𝒜ℕx \in \mathcal{A}^{\mathbb{N}}, the survivor indicator χ(n,x)∈{0,1}\chi(n, x) \in \{0, 1\} equals 11 if the burger stack of red(ρn+1(x))\mathrm{red}(\rho_{n+1}(x)) is empty and 00 otherwise.✦ The first backward survivor time σ(x)∈ℕ∪{∞}\sigma(x) \in \mathbb{N} \cup \{\infty\} of x∈𝒜ℕx \in \mathcal{A}^{\mathbb{N}} is τ0(χ)(x)\tau_0(\chi)(x).✦ The survivor order count K(x)∈ℕK(x) \in \mathbb{N} of a stream x∈𝒜ℕx \in \mathcal{A}^{\mathbb{N}} is 00 if σ(x)=∞\sigma(x) = \infty, and otherwise equals the number of pending orders in red(ρσ(x)+1(x))\mathrm{red}(\rho_{\sigma(x)+1}(x)).✦ Define the first survivor type η(x)∈ℬ\eta(x) \in \mathcal{B} of x∈𝒜ℕx \in \mathcal{A}^{\mathbb{N}} to be the freshest Boolean burger type in the burger stack of red(ρσ_(x)+1(x))\mathrm{red}(\rho_{\underline{\sigma}(x)+1}(x)), or false\mathrm{false} if that stack is empty.✦ The first survivor data ξ(x)∈ℕ×ℬ\xi(x) \in \mathbb{N} \times \mathcal{B} of a stream x∈𝒜ℕx \in \mathcal{A}^{\mathbb{N}} is the pair (K(x),η(x))(K(x), \eta(x)).✦

For n∈ℕn \in \mathbb{N} and x∈𝒜ℕx \in \mathcal{A}^{\mathbb{N}}, the shifted letter stream θn(x)∈𝒜ℕ\theta_n(x) \in \mathcal{A}^{\mathbb{N}} has entries θn(x)k=xn+k\theta_n(x)_k = x_{n+k} for k∈ℕk \in \mathbb{N}.✦ The first survivor tail Θ(x)∈𝒜ℕ\Theta(x) \in \mathcal{A}^{\mathbb{N}} of x∈𝒜ℕx \in \mathcal{A}^{\mathbb{N}} is the shifted stream θσ_(x)+1(x)\theta_{\underline{\sigma}(x)+1}(x).✦ At parsing step j∈ℕj \in \mathbb{N} of x∈𝒜ℕx \in \mathcal{A}^{\mathbb{N}}, the parsing step survivor data ξj(x)∈ℕ×ℬ\xi_j(x) \in \mathbb{N} \times \mathcal{B} is ξ(Θj(x))\xi(\Theta^j(x)), where Θj\Theta^j is the jj-fold iterate of Θ\Theta.✦ The parsing step defect count Kj(x)∈ℕK_j(x) \in \mathbb{N} at step j∈ℕj \in \mathbb{N} of x∈𝒜ℕx \in \mathcal{A}^{\mathbb{N}} is the first coordinate of ξj(x)\xi_j(x).✦

Given a space Ω\Omega, a process X:ℕ×Ω→ℕX : \mathbb{N} \times \Omega \to \mathbb{N}, a step n∈ℕn \in \mathbb{N}, and x∈Ωx \in \Omega, the defect walk VnX(x)∈ℤV_n^X(x) \in \mathbb{Z} is VnX(x)=∑j=0n−1(1−X(j,x)),V_n^X(x) = \sum_{j=0}^{n-1} (1 - X(j, x)),(4)✦ with V0X(x)=0V_0^X(x) = 0. The defect walk survival event AnX⊂ΩA_n^X \subset \Omega at step n∈ℕn \in \mathbb{N} for XX is defined via (4) by AnX={x∈Ω:VkX(x)≤0 for all k≤n}A_n^X = \{x \in \Omega : V_k^X(x) \le 0 \text{ for all } k \le n\}.✦ For X:ℕ×Ω→ℕX : \mathbb{N} \times \Omega \to \mathbb{N} and x∈Ωx \in \Omega, the defect stopping index NX(x)∈ℕN^X(x) \in \mathbb{N} is 00 if x∈AkXx \in A_k^X for all k∈ℕk \in \mathbb{N}, and otherwise is the least n∈ℕn \in \mathbb{N} with x∉AnXx \notin A_n^X.✦

Writing K•:ℕ×𝒜ℕ→ℕK_\bullet : \mathbb{N} \times \mathcal{A}^{\mathbb{N}} \to \mathbb{N} for the process (j,x)↦Kj(x)(j, x) \mapsto K_j(x), the left parsing step count NL(x)∈ℕN_L(x) \in \mathbb{N} of x∈𝒜ℕx \in \mathcal{A}^{\mathbb{N}} is NK•(x)N^{K_\bullet}(x).✦ Given the nn-th raw chunk word Cn(x)∈𝒜*C_n(x) \in \mathcal{A}^* of x∈𝒜ℕx \in \mathcal{A}^{\mathbb{N}} and the normal word norm(e)∈𝒜*\mathrm{norm}(e) \in \mathcal{A}^* of an evaluation state e∈𝒪*×ℬ*e \in \mathcal{O}^* \times \mathcal{B}^*, the short piece prefix Pnsh(x)∈𝒜*P_n^{\mathrm{sh}}(x) \in \mathcal{A}^* is defined recursively for n∈ℕn \in \mathbb{N} by P0sh(x)=εP_0^{\mathrm{sh}}(x) = \varepsilon and Pn+1sh(x)=norm(red(Cn(x)))Pnsh(x)P_{n+1}^{\mathrm{sh}}(x) = \mathrm{norm}(\mathrm{red}(C_n(x)))\, P_n^{\mathrm{sh}}(x).✦ The short left piece word WLsh(x)∈𝒜*W_L^{\mathrm{sh}}(x) \in \mathcal{A}^* of x∈𝒜ℕx \in \mathcal{A}^{\mathbb{N}} is PNL(x)sh(x)P_{N_L(x)}^{\mathrm{sh}}(x).✦

The Claim

Definition 1. Given p∈(1/2,1)p \in (1/2, 1), the raw left piece law νLraw(p)\nu_L^{\mathrm{raw}}(p) on 𝒜*\mathcal{A}^* is the pushforward measure of ℙp\mathbb{P}_p under Wτ1+W_{\tau_1^+}, νLraw(p)=ℙp∘(Wτ1+)−1.\nu_L^{\mathrm{raw}}(p) = \mathbb{P}_p \circ (W_{\tau_1^+})^{-1}.(5)✦ The right piece law νR(p)\nu_R(p) on 𝒜*\mathcal{A}^* associated with p∈(1/2,1)p \in (1/2, 1) is the pushforward of νRrev(p)\nu_R^{\mathrm{rev}}(p) under word reversal rev\mathrm{rev}, νR(p)=νRrev(p)∘rev−1.\nu_R(p) = \nu_R^{\mathrm{rev}}(p) \circ \mathrm{rev}^{-1}.(6)✦ For p∈(1/2,1)p \in (1/2, 1), the shortened left piece law νLsh(p)\nu_L^{\mathrm{sh}}(p) on 𝒜*\mathcal{A}^* is the law of WLshW_L^{\mathrm{sh}} under ℙp\mathbb{P}_p, νLsh(p)=ℙp∘(WLsh)−1.\nu_L^{\mathrm{sh}}(p) = \mathbb{P}_p \circ (W_L^{\mathrm{sh}})^{-1}.(7)✦ Given h∈ℕh \in \mathbb{N} and (x,y)∈𝒜ℕ×𝒜ℕ(x, y) \in \mathcal{A}^{\mathbb{N}} \times \mathcal{A}^{\mathbb{N}}, the raw seam word Ih(x,y)∈𝒜*I_h(x, y) \in \mathcal{A}^* is Ih(x,y)=Wτh+(x)rev(Wτh−(y)).I_h(x, y) = W_{\tau_h^+}(x)\, \mathrm{rev}(W_{\tau_h^-}(y)).(8)✦

Theorem 1. Let p∈(1/2,1)p \in (1/2, 1) and let h≥1h \ge 1 be an integer. The product measures satisfy (ℙp⊗ℙp)({(x,y)∈𝒜ℕ×𝒜ℕ:Ih(x,y) is empty-reducing})=(νLraw(p)⊗ℕ⊗νR(p)⊗ℕ)(Eh).(\mathbb{P}_p \otimes \mathbb{P}_p)\bigl(\{(x, y) \in \mathcal{A}^{\mathbb{N}} \times \mathcal{A}^{\mathbb{N}} : I_h(x, y) \text{ is empty-reducing}\}\bigr) = \bigl(\nu_L^{\mathrm{raw}}(p)^{\otimes \mathbb{N}} \otimes \nu_R(p)^{\otimes \mathbb{N}}\bigr)(E_h).(9)✦

What Is Checked

The piece laws (5), (6), and (7) along with the raw seam word (8) and the factorization identity (9) are stated over an FK parameter p∈(1/2,1)p \in (1/2, 1), the i.i.d. stream measure ℙp\mathbb{P}_p on 𝒜ℕ\mathcal{A}^{\mathbb{N}}, the hitting times τh+\tau_h^+ and τh−\tau_h^- with stopped words WTW_T and reversed right piece law νRrev(p)\nu_R^{\mathrm{rev}}(p), the short left piece word WLshW_L^{\mathrm{sh}} from backward survivor parsing and the defect stopping index NLN_L, as well as the empty-reducing piece seam event EhE_h under the product of natural pieces laws for h≥1h \ge 1.