A new computation of pairing probabilities in several multiple-curve models

This paper presents a concise new computation of pairing probabilities for multiple chordal interfaces in various critical models by leveraging the convexity and a newly established uniqueness property of local multiple SLE measures.

The Gist

Imagine you are standing in a crowded room where people are holding hands, forming a giant, tangled web of connections. Now, imagine that this web is made of invisible, wiggly strings that are constantly trying to find their way from one side of the room to the other without ever crossing each other.

This paper is about figuring out the odds of how these strings will pair up.

Here is the breakdown of the paper's big ideas, translated from "math-speak" into everyday language:

1. The Setting: The "Dance Floor" of Physics

The author is studying three specific "dance floors" (mathematical models) where these wiggly strings appear:

• The Ising Model: Think of this as a giant checkerboard where every square is either a magnet pointing "Up" (+) or "Down" (-). The lines we are tracking are the borders between the Up magnets and the Down magnets.
• The Harmonic Explorer: Imagine a robot walking on a honeycomb grid. It tries to find the boundary between black and white areas, making random turns but following specific rules.
• The Gaussian Free Field (GFF): This is a bit like a fluctuating, invisible landscape of hills and valleys. The lines we track are the "contour lines" (like on a topographic map) that separate different heights.

In all three cases, if you have $2N$ specific points on the edge of the room, the strings will naturally connect them in pairs. But here's the mystery: Which point connects to which?

2. The Problem: Too Many Possibilities

If you have 4 points (A, B, C, D), there are only a few ways to pair them up without the lines crossing. But if you have 10 points, the number of possible pairings explodes.

For a long time, mathematicians had to do incredibly difficult, specific calculations for each of these three models to figure out the probability of a specific pairing. It was like having to solve a unique, complex puzzle for every single type of dance floor.

3. The Solution: A Universal "Master Key"

The author, Alex Karrila, found a short, universal shortcut. Instead of solving the puzzle for each model individually, he realized that all these models share a hidden "DNA."

He uses a concept called SLE (Schramm–Loewner Evolution). You can think of SLE as the "universal language" of these wiggly lines. No matter if the line is a magnet border, a robot path, or a contour line, they all speak this same language.

The Core Idea (The "Recipe"):
The paper argues that the probability of any specific pairing is just a simple ratio of two numbers:

1. The "Total Score": A number calculated for all possible pairings combined.
2. The "Specific Score": A number calculated for just that one pairing you are interested in.

The Probability = (Specific Score) / (Total Score)

4. The Magic Trick: Convexity and Uniqueness

How did he prove this works for everything? He used two main tools:

• The "Mixing Bowl" (Convexity): Imagine you have a bowl of soup. If you mix a little bit of "Ising soup" with a little bit of "GFF soup," the result is still a valid soup. The author proves that if you mix different pairing scenarios together, the math still holds up perfectly. This allows him to treat the problem as a simple algebraic equation rather than a chaotic mess.
• The "Fingerprint" (Uniqueness): He proves that every specific pairing has a unique mathematical "fingerprint" (called a partition function). If two different pairings had the same fingerprint, the math would break. But since they are all unique, we can simply look at the fingerprints to solve the equation.

5. Why This Matters

Before this paper, if you wanted to know the odds of a specific pattern in the Ising model, you needed a PhD-level proof specific to that model. If you wanted to know the odds for the GFF, you needed a different PhD-level proof.

This paper says: "Stop reinventing the wheel."
Once you identify that your model speaks the language of SLE (which these three do), you can instantly calculate the pairing probabilities using this simple ratio formula. It turns a mountain of complex calculus into a simple arithmetic problem.

Summary Analogy

Imagine you are trying to guess who is married to whom at a massive wedding reception.

• Old Way: You had to interview every single guest, analyze their history, and write a unique report for every possible couple.
• New Way (This Paper): You realize that the seating chart follows a universal rule. You just need to look at the "Total Seating Plan" and the "Specific Couple's Plan." The answer is simply: How much of the Total Plan does this Couple occupy?

The author has handed us the calculator to do this for any "wedding" (model) that follows the universal rules of these wiggly lines.

Technical Summary

Here is a detailed technical summary of the paper "A new computation of pairing probabilities in several multiple-curve models" by Alex M. Karrila.

1. Problem Statement

The paper addresses the problem of computing pairing probabilities for multiple chordal interfaces in critical planar statistical mechanics models. Specifically, it considers models where $2N$marked boundary points on a domain are connected by$N$ non-crossing random curves (interfaces).

• The Challenge: In models like the critical Ising model, the Harmonic Explorer, and the Gaussian Free Field (GFF), the interfaces form a random planar pairing of the $2N$points. There are$C_N = \frac{1}{N+1}\binom{2N}{N}$ possible non-crossing pairings (Catalan number).
• The Goal: Determine the probability $p_\alpha$ that the interfaces connect the boundary points according to a specific pairing configuration $\alpha$.
• Context: Previous proofs for these probabilities (e.g., in [29] for Ising and [28] for GFF) relied on specific martingale arguments tailored to particular values of the SLE parameter $\kappa$ (specifically $\kappa=3$ and $\kappa=4$). These required fine-grained analysis of the SLE process and martingales up to the termination time, which was technically complex and model-specific.

2. Methodology

The author proposes a unified, short, and general proof that relies on the structural properties of local multiple Schramm–Loewner Evolution (SLE) rather than specific martingale calculations for each model.

Core Mathematical Framework

1. Local Multiple SLE: The paper utilizes the definition of local multiple SLE($\kappa$) driven by a partition function $Z(x)$. The driving function $W_t$ of the Loewner equation is modified by a drift term derived from the partition function:
   $$dW_t = \sqrt{\kappa} d\beta_t + \kappa \frac{\partial_j Z}{Z} dt$$
   where $Z$ satisfies a system of partial differential equations (PDEs) and conformal covariance conditions.

2. Convexity of Measures: The set of local multiple SLE measures forms a convex cone. If a random curve model converges to a local multiple SLE, its law can be expressed as a convex combination of measures conditioned on specific pairings.

3. Key Lemma (Lemma 1.1): The central theoretical contribution is a lemma combining convexity and uniqueness:

   • If a measure $P$ is a convex combination of multiple SLE laws $P_\alpha$ (with weights $p_\alpha$), then $P$ is itself a local multiple SLE.
   • Crucially, the partition function $Z$ of the combined measure is uniquely determined (up to a constant) by the weighted sum of the partition functions $Z_\alpha$ of the conditional measures:
     $$Z(x) = C \sum_\alpha p_\alpha \frac{Z_\alpha(V_0)}{Z_\alpha(x)} Z_\alpha(x)$$
     (Note: The paper simplifies this to $Z(x) = \sum p_\alpha Z_\alpha(x)$ under appropriate normalization).
   • Uniqueness: The paper proves (Lemma 3.1) that for a fixed geometry, the law of the driving function uniquely determines the partition function up to a multiplicative constant. This is established using the strong maximum principle for elliptic PDEs and Hörmander's criterion for the absolute continuity of the SDE solution (Lemma 3.2).

The Strategy

The proof strategy for any specific model involves three steps:

1. Convergence: Verify that the discrete model converges weakly to a local multiple SLE (Input i).
2. Identification: Verify that the conditional laws (given a pairing $\alpha$) converge to specific local multiple SLEs with known partition functions $Z_\alpha$ (Input ii).
3. Linear Combination: Verify that the unconditional partition function $Z$ of the model is the sum of the conditional partition functions: $Z(x) = \sum_\alpha Z_\alpha(x)$ (Input iii).

If these hold, the pairing probabilities $p_\alpha$ are simply the ratios of the partition functions evaluated at the boundary points:
$$p_\alpha = \frac{Z_\alpha(V_1^0, \dots, V_{2N}^0)}{Z(V_1^0, \dots, V_{2N}^0)}$$

3. Key Contributions

• Generalization: The proof is independent of the specific value of $\kappa$ and the underlying lattice model, provided the model is identified as a local multiple SLE. This contrasts with previous methods that were specific to $\kappa=3$ (Ising) or $\kappa=4$ (GFF/Harmonic Explorer).
• Simplification: It replaces complex martingale termination analyses with a direct argument based on the convexity of the space of partition functions and the uniqueness of the driving function's law.
• New Uniqueness Lemma: The paper establishes a rigorous uniqueness result (Lemma 3.1) for local multiple SLE measures in a fixed geometry, proving that the partition function is uniquely determined by the driving function's law. This relies on a novel application of Hörmander's criterion to show that the SDE solution has a density with respect to the Lebesgue measure, preventing the ratio of partition functions from being constant unless the functions are proportional.
• Unified Computation: It provides a single formula for pairing probabilities across three distinct models.

4. Results

The author applies the general strategy to compute pairing probabilities for:

1. Critical Ising Model ($\kappa=3$):

   • The partition function $Z$ is given by a Pfaffian formula.
   • The conditional partition functions $Z_\alpha$ correspond to "pure" partition functions derived from global multiple SLE theory.
   • Result: $P[\text{pairing } \alpha] = Z_\alpha / Z$. This recovers the result of [29] but with a simpler proof.

2. Multiple Harmonic Explorer ($\kappa=4$):

   • The partition function $Z$ is a product of differences: $\prod (x_j - x_i)^{(-1)^{j-i}/2}$.
   • Result: $P[\text{pairing } \alpha] = Z_\alpha / Z$. This recovers the result of [15] using the general framework.

3. Gaussian Free Field (GFF) Level Lines ($\kappa=4$):

   • The paper rigorously establishes that GFF level lines with alternating boundary conditions form a collection of disjoint curves corresponding to a global multiple SLE.
   • Result: $P[\text{pairing } \alpha] = Z_\alpha / Z$. This provides a concise proof for the result in [28] and extends it to general boundary conditions with $N$ jumps.

5. Significance

• Conceptual Clarity: The paper clarifies the probabilistic structure underlying multiple SLEs, showing that the "mixing" of different topological pairings is a direct consequence of the convexity of the space of partition functions.
• Technical Advancement: The proof of Lemma 3.1 (uniqueness) and Lemma 3.2 (absolute continuity via Hörmander's criterion) offers new tools for the analysis of SLE-type processes, potentially applicable to other variants of SLE with partition functions.
• Bridge to CFT: The result that pairing probabilities are ratios of partition functions has direct interpretations in Conformal Field Theory (CFT), linking the probabilistic pairing weights to CFT correlation functions, though the paper focuses on the probabilistic derivation.
• Efficiency: By decoupling the computation of probabilities from the specific dynamics of the termination of curves, the method offers a more robust and transferable approach for studying scaling limits in statistical mechanics.

In summary, Karrila provides a powerful, general framework that simplifies the calculation of topological pairing probabilities in critical models by leveraging the convex structure of multiple SLE partition functions and a new uniqueness theorem for local SLE measures.