Extreme value theorem for geodesic flow on the quotient of the theta group

This paper establishes an extreme value theorem for the geodesic flow on the hyperbolic surface associated with the theta group by introducing a spliced continued fraction algorithm, proving its dynamical equivalence to the flow's first return map, and deriving a Galambos-type extreme value law for maximal cusp excursions via spectral analysis of the transfer operator.

The Gist

Imagine you are watching a tiny, frictionless marble roll forever across a strange, bumpy landscape. This landscape isn't flat; it's a hyperbolic surface, which means it curves away from itself like a saddle or a Pringles chip. Because of this curvature, the marble's path (a geodesic) can get very long and very wild.

Sometimes, the marble zooms off toward the "edges" of this world. In mathematics, these edges are called cusps. Think of a cusp as a deep, infinitely long funnel. The marble might dive down one funnel, come back up, and then dive down another.

The paper by Kim, Lee, and Lim is about predicting the most extreme behavior of this marble. Specifically: How deep will the marble dive into these funnels if we watch it for a very long time?

Here is the breakdown of their discovery using simple analogies:

1. The Problem: Two Different Funnels

The surface they are studying (associated with the "Theta Group") is special because it has two different funnels (cusps).

• Funnels are tricky: In the past, mathematicians had a great way to track marbles diving into one specific funnel using a tool called a "Continued Fraction." Think of a continued fraction as a barcode or a recipe that tells you exactly how the marble moves.
• The Catch: The old recipes only worked for one funnel at a time. If the marble switched from Funnel A to Funnel B, the old recipe broke. It was like trying to read a book written in English while the author suddenly switched to French halfway through.

2. The Solution: The "Spliced" Recipe

The authors' big idea was to splice (stitch together) two different recipes into one super-recipe.

• They took the "Even" recipe (for Funnel A) and the "Odd-Odd" recipe (for Funnel B) and glued them together.
• They created a new Spliced Continued Fraction (SCF).
• The Analogy: Imagine you have two different languages. Instead of switching back and forth, you invent a new language where every other sentence follows the grammar of Language A, and the next follows Language B. This new language can describe the entire journey of the marble, no matter which funnel it dives into.

3. The Connection: From Numbers to Geometry

The paper proves a magical link between this new number recipe and the physical movement of the marble.

• The Digits are the Depth: In their new recipe, the numbers (digits) generated tell you how many "tiles" the marble crosses.
• The Big Number = The Deep Dive: If the recipe generates a huge number (like a 1,000,000), it means the marble is taking a very long, deep dive into a funnel.
• The Translation: They proved that the "largest number" in the recipe over a long time is directly related to the "deepest point" the marble reaches in the physical world.

4. The Prediction: The Extreme Value Theorem

Now, here is the main result. If you watch the marble for a very, very long time, you want to know: What is the probability that it will dive deeper than a certain limit?

The authors found a precise mathematical law (called the Galambos-type Extreme Value Law) that answers this.

• The Analogy: Imagine you are flipping a coin. You know the odds of getting heads. But what if you flip it a million times? What are the odds of getting a streak of 50 heads in a row?
• The Result: They found that the probability of the marble diving deeper than a certain depth follows a specific curve (an exponential decay).
  • If you set a very deep limit, the chance of the marble reaching it is small.
  • If you lower the limit, the chance gets bigger.
  • The formula they found ($e^{-1/y}$) tells you exactly how small that chance is, based on how deep you set your limit.

Why is this a Big Deal?

1. It's the First of Its Kind: This is the first time this kind of prediction has been made for a surface with two funnels. Before this, we could only predict behavior for surfaces with one funnel or for very simple groups.
2. Geometry vs. Symbols: Usually, mathematicians study these deep dives by counting "symbolic windings" (like how many times the marble loops around a pole). This paper is special because it measures the actual physical height (how deep it goes into the funnel), which is much more intuitive for understanding the shape of the universe.
3. The "Splicing" Trick: The method of stitching two different mathematical maps together to create a unified system is a clever new tool that other mathematicians can now use to solve similar problems.

Summary

The authors built a universal translator (the Spliced Continued Fraction) that can read the "barcode" of a marble rolling on a complex, two-funnel surface. They proved that the biggest numbers in this barcode perfectly predict the deepest dives the marble will ever take, giving us a precise formula to calculate the odds of extreme events in this curved world.

Technical Summary

Here is a detailed technical summary of the paper "Extreme Value Theorem for Geodesic Flow on the Quotient of the Theta Group" by Jaelin Kim, Seul Bee Lee, and Seonhee Lim.

1. Problem Statement and Motivation

The paper addresses the statistical behavior of geodesic flows on hyperbolic surfaces, specifically focusing on extreme value theorems (EVT) regarding the maximal excursions of geodesics into the cusps (ends) of the surface.

• Context: For the modular surface $SL(2, \mathbb{Z})\backslash \mathbb{H}^2$, the relationship between continued fraction digits and cusp excursions is well-established (via the Gauss map). Galambos (1972) proved an EVT for continued fraction digits, and Pollicott (2009) translated this to a geometric EVT for the geodesic flow on the modular surface.
• The Gap: The authors study the Theta group $\Theta$, a congruence subgroup of level 2 and index 3 in $SL(2, \mathbb{Z})$. The quotient surface $M = \Theta\backslash \mathbb{H}^2$ is topologically a sphere with two inequivalent cusps (at the orbits of $\infty$ and $1$) and one elliptic point.
• The Challenge: Existing continued fraction algorithms (like the Even Continued Fraction for the cusp at $\infty$ and the Odd-Odd Continued Fraction for the cusp at $1$) only capture excursions into a *single* cusp. No single symbolic coding existed that simultaneously tracked excursions into *both* cusps for a generic geodesic on$M$. Furthermore, previous EVT results for non-essentially free groups or surfaces with multiple cusps were lacking.

2. Methodology

The authors develop a unified symbolic dynamics framework to bridge the gap between the geometry of the surface and the statistics of the geodesic flow.

A. The Spliced Continued Fraction (SCF)

To capture both cusps, the authors construct a new interval map $T: (0, 1) \to [0, 1]$ by splicing two existing maps:

1. Even Continued Fraction (ECF) map ($T_e$): Captures excursions to the cusp $\Theta.\infty$.
2. Odd-Odd Continued Fraction (OOCF) map ($T_o$): Captures excursions to the cusp $\Theta.1$.

The spliced map $T$ is defined piecewise:
$$T(x) = \begin{cases} T_e(x) & x \in (0, 1/2] \\ T_o(x) & x \in (1/2, 1) \end{cases}$$
This generates a Spliced Continued Fraction (SCF) expansion for $x \in (0, 1)$, denoted as $x = [0; (a_1, \varepsilon_1)_{s_1}, (a_2, \varepsilon_2)_{s_2}, \dots]$, where the digits include information about the magnitude ($a_n$), sign ($\varepsilon_n$), and type ($s_n \in \{e, o\}$) of the excursion.

B. Natural Extension and Geometric Coding

• Natural Extension: The authors construct the natural extension of $T$, denoted as $\bar{T}$ on the space $\Omega = [0, 1] \times [\sqrt{3}-2, \sqrt{3}]$. This system is isomorphic to the first return map of the geodesic flow on a carefully chosen cross-section $\pi(X)$ of the unit tangent bundle $T^1M$.
• Dual SCF: To handle the backward endpoint of geodesics, they introduce a "dual" SCF expansion. The natural extension acts as a two-sided shift on the sequence of SCF digits.
• Geometric Correspondence: They prove that the partial quotients $a_n$ of the SCF correspond directly to the number of fundamental quadrilaterals crossed by the geodesic in the tessellation of $\mathbb{H}^2$. Specifically, the magnitude of the digit $a_n$ is asymptotically equivalent to the hyperbolic length (height) of the geodesic excursion into a cusp.

C. Spectral Analysis

To derive the extreme value law, the authors analyze the transfer operator (Ruelle-Perron-Frobenius operator) $\mathcal{L}$ associated with the map $T$.

• They establish that $\mathcal{L}$ acts on the space of functions of bounded variation ($BV$) and possesses a spectral gap.
• This spectral gap implies exponential mixing of the invariant measure $\mu$ and allows for the control of correlations between distant digits in the SCF expansion.

3. Key Results

Theorem 1.1: Extreme Value Theorem for SCF Digits

The authors prove a Galambos-type extreme value theorem for the digits $a_n$ of the spliced continued fraction. For the $T$-invariant probability measure $\mu$:
$$\lim_{N \to \infty} \mu \left\{ x \in (0, 1) : \max_{1 \le n \le N} a_n(x) \le \frac{2Ny}{\log(2 + \sqrt{3})} \right\} = \exp\left(-\frac{1}{y}\right)$$
This result generalizes Galambos' classical theorem for regular continued fractions to the spliced system.

Theorem 1.2: Geometric Extreme Value Theorem

Translating the symbolic result back to geometry, they establish the distribution of maximal cusp excursions for the geodesic flow on $M$. Let $d_M(\pi(i), \gamma_v(t))$ be the distance from a base point to the geodesic $\gamma_v$ at time $t$. There exists an explicit constant $C > 0$ such that for the Liouville measure $m$:
$$\lim_{T \to \infty} m \left\{ v \in T^1M : \max_{0 \le t \le T} d_M(\pi(i), \gamma_v(t)) - \log T \le \log(Cy) \right\} = e^{-1/y}$$
This describes the limiting distribution of the maximal height (distance into the cusp) of geodesics over time $T$.

4. Significance and Contributions

1. First EVT for Non-Essentially Free Groups: The Theta group is not essentially free (it contains torsion/elliptic elements). This is the first extreme value theorem established for a geodesic flow on a quotient by a non-essentially free Fuchsian group.
2. Multiple Cusps: Unlike previous results that focused on single-cusp excursions or the modular surface, this work handles a surface with two distinct cusps simultaneously. The spliced continued fraction is the first symbolic coding capable of tracking excursions into both cusps in a single dynamical system.
3. Geometric vs. Symbolic: While previous works often focused on "symbolic winding numbers" (combinatorial depth), this paper provides an EVT for geometrically measured height (hyperbolic distance). This is a more direct physical interpretation of "how deep" a geodesic goes into a cusp.
4. Methodological Innovation: The construction of the spliced map and its natural extension provides a robust framework for analyzing other congruence subgroups or surfaces with multiple cusps where standard continued fraction algorithms fail to provide a global coding.

Summary

The paper successfully unifies the dynamics of geodesic flows on a specific hyperbolic surface with two cusps by creating a "spliced" continued fraction algorithm. By proving the spectral gap of the associated transfer operator, they derive a precise extreme value law for both the symbolic digits and the geometric heights of geodesic excursions, extending classical number-theoretic results to a more complex geometric setting.