OctoInformBeta

OpenAI preprint, September 2026 · §4. The critical block tree · §4

Lemma 4.2 [Quadratic inverse criterion]: Total mass 11, mean 1/21/2, and finite second moment from an Abel lower bound along a sequence converging to 11

A quadratic inverse criterion for critical block weights.

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

In the Fortuin-Kasteleyn (FK) random planar map model, the map generating function T(z)=∑n≥0anznT(z) = \sum_{n \ge 0} a_n z^n and the block generating function Φ(y)=1+∑m≥1bmym\Phi(y) = 1 + \sum_{m \ge 1} b_m y^m satisfy the quadratic relation T(z)=Φ(zT(z)2)T(z) = \Phi(z T(z)^2). At the critical radius u>0u > 0, the question is whether this relation together with an Abel square-root lower bound on the map series implies that the block offspring random variable ξ\xi of the critical block tree satisfies 𝔼ξ=1\mathbb{E}\xi = 1 and Var⁡(ξ)<∞\operatorname{Var}(\xi) < \infty.

Block Offspring Weights and Generating Series

A block offspring weight distribution ν\nu is a sequence (νk)k≥0(\nu_k)_{k \ge 0} of nonnegative real numbers satisfying ∑k=0∞νk=1\sum_{k=0}^\infty \nu_k = 1.✦

For a block offspring weight distribution ν\nu and x∈ℝx \in \mathbb{R}, the probability generating function Gν(x)G_\nu(x) is Gν(x)=∑k=0∞νkxk.G_\nu(x) = \sum_{k=0}^\infty \nu_k x^k.(1)✦ The Abel mean Aν(x)A_\nu(x) of a block offspring weight distribution ν\nu at x∈ℝx \in \mathbb{R} is the weighted series Aν(x)=∑k=0∞kνkxk.A_\nu(x) = \sum_{k=0}^\infty k \nu_k x^k.(2)✦

Inventory Streams and Block Coefficients

An FK parameter choice PP with probability pp is a parameter choice for the FK model specified by a real probability p∈(1/2,1)p \in (1/2, 1). The flexible order parameter tflex(P)t_{\mathrm{flex}}(P) of an FK parameter choice PP with probability pp is the real number 1+2p/(1−p)1 + 2p/(1 - p). The inventory radius parameter u(P)u(P) of an FK parameter choice PP with probability pp is the real number (1−p)/16(1 - p)/16.

For two burger types, an inventory order is either a fixed order for one of the two types or the flexible order✦, and a burger-order alphabet letter is either a burger of one of the two types or an inventory order.✦ The letter mass function ℓP\ell_P for an FK parameter choice PP with probability pp assigns weight 1/41/4 to each of the two burgers, p/2p/2 to the flexible order, and (1−p)/4(1 - p)/4 to each of the two fixed orders.✦ The letter distribution λP\lambda_P for an FK parameter choice PP is the probability distribution on the burger-order alphabet with probabilities given by ℓP\ell_P✦, and the independent and identically distributed (i.i.d.) inventory stream measure ℙP\mathbb{P}_P for PP on sequences of alphabet letters indexed by a set II is the infinite product measure of λP\lambda_P.

An inventory stack evaluation state is a pair consisting of a finite sequence of pending orders and a finite sequence of burger types stored freshest-first.✦ Starting from an evaluation state ee, the word evaluation of a finite word ww of alphabet letters processes ww step by step by consuming the freshest eligible burger for each order, returning ee when ww is empty✦; the word reduction of ww is its word evaluation from empty pending orders and an empty burger stack.✦ An empty-reducing word ww of alphabet letters is a finite word whose word reduction produces the empty evaluation state with no pending orders and no burgers.✦ For a stream x=(xi)i≥0x = (x_i)_{i \ge 0} of alphabet letters and an integer N≥0N \ge 0, the letter sequence prefix of length NN is the finite sequence (xi)0≤i<N(x_i)_{0 \le i < N}, and the initial word wN(x)w_N(x) is the word (x0,…,xN−1)(x_0, \dots, x_{N-1}) formed from that prefix.✦ The empty-reducing word probability eP(N)e_P(N) at length N≥0N \ge 0 for an FK parameter choice PP is eP(N)=ℙP(wN(x) is empty-reducing),e_P(N) = \mathbb{P}_P\bigl(w_N(x) \text{ is empty-reducing}\bigr),(3) where xx has law ℙP\mathbb{P}_P on ℕ\mathbb{N}. Summing over even lengths gives the total empty word probability τ(P)\tau(P) under PP, τ(P)=∑n=0∞eP(2n).\tau(P) = \sum_{n=0}^\infty e_P(2n).(4)✦

A typed burger-order letter is either a burger or an order carrying one of the two burger types.✦ Given a finite word ww of typed letters and a freshest-first stack ss of burger types, typed word stack execution processes ww letter by letter against ss, returning the updated stack on success or failing upon an illegal closure.✦ Over a finite alphabet Σ\Sigma, the set of words of length N≥0N \ge 0 consists of all sequences (a0,…,aN−1)(a_0, \dots, a_{N-1}) of letters in Σ\Sigma✦; for the typed letter alphabet, the empty-reducing typed word set 𝒲N\mathcal{W}_N consists of those words of length NN whose typed stack execution from an empty stack succeeds and ends with an empty stack.✦

A non-crossing arch matching EE on a finite sequence ll of positions is a finite sequence of ordered pairs from ll built inductively from the empty matching on the empty sequence by skipping an unmatched position or pairing two positions aa and bb nested around non-crossing arch matchings on the interior segment strictly between aa and bb and the suffix strictly after bb.✦ For a finite sequence EE of ordered pairs in a set SS, the arch transposition product ρ(E)\rho(E) is the permutation of SS equal to the identity when EE is empty and to ρ(E′)∘(ab)\rho(E') \circ (a\; b) when EE prepends (a,b)(a, b) to E′E', where (ab)(a\; b) swaps aa and bb.✦ Given a word map ψ\psi from a finite sequence ll in SS to typed letters, a two-stack matching certificate MM for ψ\psi along ll is a pair of non-crossing arch matchings on ll whose concatenated arch endpoints form a permutation of ll, such that the first matching pairs first-type burgers to first-type orders and the second matching pairs second-type burgers to second-type orders.✦ The stack matching mate permutation σM\sigma_M of a two-stack matching certificate MM is the arch transposition product of the concatenated list of its first and second arch matchings.✦

A chord opening endpoint ii of a permutation σ\sigma of {0,…,N−1}\{0, \dots, N - 1\} is an index in {0,…,N−1}\{0, \dots, N - 1\} satisfying i<σ(i)i < \sigma(i). Two chord opening endpoints ii and jj of a permutation σ\sigma of {0,…,N−1}\{0, \dots, N - 1\} satisfy the chord crossing relation if either i<j<σ(i)<σ(j)i < j < \sigma(i) < \sigma(j) or j<i<σ(j)<σ(i)j < i < \sigma(j) < \sigma(i). The chord intersection graph of a permutation σ\sigma of {0,…,N−1}\{0, \dots, N - 1\} is the simple graph on the chord opening endpoints of σ\sigma whose edges join pairs of chord opening endpoints satisfying the chord crossing relation.✦

A connected core word ww is a finite word of typed letters such that there exist a length K≥0K \ge 0, a word map ψ\psi from {0,…,K−1}\{0, \dots, K - 1\} to typed letters whose values (ψ0,…,ψK−1)(\psi_0, \dots, \psi_{K-1}) form ww, and a two-stack matching certificate MM for ψ\psi along (0,…,K−1)(0, \dots, K - 1) such that σM\sigma_M has an opening endpoint at 00 from which every opening endpoint is reachable in the chord intersection graph of σM\sigma_M.✦ The flexible closing count f(w,s)f(w, s) of a finite word ww of typed letters and an initial stack ss of burger types is the number of flexible order closures encountered along valid steps of typed stack execution of ww starting from ss, stopping upon any illegal closure.✦ For t∈ℝt \in \mathbb{R} and an integer m≥0m \ge 0, the core generating coefficient cm(t)c_m(t) is cm(t)=∑w∈𝒲2mcoretf(w,∅),c_m(t) = \sum_{w \in \mathcal{W}_{2m}^{\mathrm{core}}} t^{f(w, \emptyset)},(5)✦ where 𝒲2mcore\mathcal{W}_{2m}^{\mathrm{core}} is the set of connected core words in 𝒲2m\mathcal{W}_{2m} and ∅\emptyset is the empty stack. The block generating coefficient Bm(t)B_m(t) at t∈ℝt \in \mathbb{R} and m≥0m \ge 0 is Bm(t)={1+c0(t),m=0,cm(t),m≥1.B_m(t) = \begin{cases} 1 + c_0(t), & m = 0, \\ c_m(t), & m \ge 1. \end{cases}(6)✦ An FK parameter choice PP and an integer m≥0m \ge 0 define the actual FK block mass ωm(P)\omega_m(P) by ωm(P)=Bm(tflex(P))(u(P)τ(P)2)mτ(P).\omega_m(P) = \frac{B_m\bigl(t_{\mathrm{flex}}(P)\bigr)\, \bigl(u(P)\, \tau(P)^2\bigr)^m}{\tau(P)}.(7)✦

The Claim

Theorem 1. Suppose (πk)k≥0(\pi_k)_{k \ge 0} is a sequence of nonnegative real numbers, and (xn)n≥0(x_n)_{n \ge 0} and (ηn)n≥0(\eta_n)_{n \ge 0} are real sequences converging to 11 with xn∈[0,1]x_n \in [0, 1] and ∑k=0∞πkxnk=ηn\sum_{k=0}^\infty \pi_k x_n^k = \eta_n for all n≥0n \ge 0. Then ∑k=0∞πk=1.\sum_{k=0}^\infty \pi_k = 1.(8)✦

Theorem 2. Let ν\nu be a block offspring weight distribution, let (xn)n≥0(x_n)_{n \ge 0} be a sequence in [0,1)[0, 1) converging to 11, and let C≥0C \ge 0 satisfy 0≤Gν(xn)2−xn≤C(1−xn)20 \le G_\nu(x_n)^2 - x_n \le C(1 - x_n)^2 for all n≥0n \ge 0. Then ∑k=0∞kνk=12and∑k=0∞k2νk<∞.\sum_{k=0}^\infty k \nu_k = \frac{1}{2} \quad \text{and} \quad \sum_{k=0}^\infty k^2 \nu_k < \infty.(9)✦

Theorem 3. Let ν\nu be a block offspring weight distribution, let (πk)k≥0(\pi_k)_{k \ge 0} be a sequence of nonnegative real numbers, let (xn)n≥0(x_n)_{n \ge 0} be a sequence in [0,1)[0, 1) converging to 11 with ∑k=0∞πk(xnGν(xn)2)k=Gν(xn)\sum_{k=0}^\infty \pi_k \bigl(x_n G_\nu(x_n)^2\bigr)^k = G_\nu(x_n) for all n≥0n \ge 0, and let c>0c > 0 satisfy 1−xnAν(xn)≥c\sqrt{1 - x_n}\, A_\nu(x_n) \ge c for all n≥0n \ge 0. Then ∑k=0∞πk=1,∑k=0∞kπk=12,∑k=0∞k2πk<∞.\sum_{k=0}^\infty \pi_k = 1, \qquad \sum_{k=0}^\infty k \pi_k = \frac{1}{2}, \qquad \sum_{k=0}^\infty k^2 \pi_k < \infty.(10)✦

Theorem 4. For any FK parameter choice PP, the actual FK block mass (ωm(P))m≥0\bigl(\omega_m(P)\bigr)_{m \ge 0} satisfies ∑m=0∞ωm(P)=1,∑m=0∞mωm(P)=12,∑m=0∞m2ωm(P)<∞.\sum_{m=0}^\infty \omega_m(P) = 1, \qquad \sum_{m=0}^\infty m\, \omega_m(P) = \frac{1}{2}, \qquad \sum_{m=0}^\infty m^2 \omega_m(P) < \infty.(11)✦

In the paper, the result reads: if TT and Φ\Phi have nonnegative coefficients with constant term 11 and satisfy T(z)=Φ(zT(z)2)T(z) = \Phi(z T(z)^2), if u>0u > 0 satisfies T(u)<∞T(u) < \infty, and if for some c>0c > 0 the bound ∑n≥1nanunxn≥c(1−x)−1/2\sum_{n \ge 1} n a_n u^n x^n \ge c (1 - x)^{-1/2} holds for all x<1x < 1 sufficiently close to 11, then setting τ=T(u)\tau = T(u) and v=uτ2v = u \tau^2 yields Φ(v)=τ\Phi(v) = \tau, 2vΦ′(v)=Φ(v)2 v \Phi'(v) = \Phi(v), and Φ″(v)<∞\Phi''(v) < \infty. In normalized coordinates νn=anun/τ\nu_n = a_n u^n / \tau and πm=bmvm/τ\pi_m = b_m v^m / \tau for m≥1m \ge 1 with π0=1/τ\pi_0 = 1/\tau, and with limits taken along sequences converging to 11, the identities (10) and (11) correspond to those three conclusions of the paper and answer the criticality and finite-variance question for the block offspring distribution.

What Is Checked

The claims (8), (9), (10), and (11) are stated over nonnegative real sequences and block offspring weight distributions ν\nu satisfying generating-series limits along sequences in [0,1][0, 1] or [0,1)[0, 1) converging to 11, quadratic gap bounds or quadratic inverse relations with an Abel square-root lower bound, and, for any FK parameter choice PP with p∈(1/2,1)p \in (1/2, 1), the actual FK block mass ωm(P)\omega_m(P) constructed from the i.i.d. inventory stream probability and connected core word coefficients.