Skew Stickiness Ratio

• Skew Stickiness Ratio (SSR) is a metric that quantifies the normalized relationship between system asymmetry (skew) and entrapment (stickiness) across various domains.
• SSR is computed via normalized moments or estimator ratios in contexts like Hamiltonian systems, sticky diffusions, and financial volatility models, providing a robust calibration tool.
• The versatility of SSR in diagnosing trapping structures and asymmetry makes it crucial for both theoretical insights and practical applications in complex system analysis.

The skew stickiness ratio (SSR) is a statistical construct that quantifies the interplay between two distinct but interrelated mechanisms—skewness and stickiness—across several domains, notably dynamical systems, stochastic processes with barriers, and financial stochastic volatility modeling. SSR emerges as a unifying metric for expressing the dynamic coupling of a system's asymmetry (skew) with its tendency to become trapped or to linger (stickiness) in certain states or regions. Its specific mathematical realization, operational interpretation, and implications depend on the underlying context (Hamiltonian recurrence, sticky diffusions, financial implied volatilities), but in each, the SSR serves as a concise diagnostic and calibration tool for identifying, quantifying, and modeling the phenomenon of dynamical stickiness modulated by structural or stochastic asymmetry.

1. Hamiltonian Recurrence Statistics and SSR

In the ergodic theory of low-dimensional Hamiltonian systems, the skew stickiness ratio $S$ is defined as the normalized width of the recurrence time distribution of a chaotic trajectory into small phase-space cells. For a grid-tiling of the phase space $M$ and a sequence of return times $\{\tau_k\}$ of a long chaotic orbit to each cell, the SSR is given by

$S = \frac{\sigma}{\mu},$

where $\mu = \langle \tau \rangle$ is the mean recurrence time and $\sigma$ its standard deviation (Lozej, 2021).

• Poissonian Recurrence Baseline: In the chaotic (strongly mixing) regime, where recurrence times are memoryless and exponentially distributed, $S=1$ exactly.
• Stickiness and Multi-exponential Mixtures: Trapping near islands, cantori, or hierarchical structures introduces multiple return-time scales, leading to mixture-of-exponentials and $S>1$. The deviation from unity directly quantifies the strength of stickiness.
• Visualization: SSR maps (S-plots) spatially resolve regions of high phase-space stickiness, supporting robust, global detection of trapping structures.

This definition is algorithmically tractable via online updating of moments for each grid cell, enabling automation of sticky-region identification without a priori knowledge of regular structures.

2. SSR in Sticky-Threshold and Skew Diffusions

For diffusion processes with "sticky" or "skew" boundary conditions (notably at a singular threshold), the SSR arises as a ratio of estimators for the stickiness and skewness parameters (Anagnostakis et al., 2024):

• Given a sticky–oscillating–skew process with stickiness parameter $\rho$ (quantifying mean sojourn time at the threshold) and skewness $\beta$ (quantifying asymmetry in threshold crossing), consistent estimators $\hat\rho_n, \hat\beta_n$ are constructed via high-frequency occupation and local time functionals.
• The SSR is

$$\mathrm{SSR}_n = \frac{\hat\rho_n}{\hat\beta_n} \to \frac{\rho}{\beta}$$

in probability as sample size grows.

This SSR quantifies the relative dominance of stickiness over skew-induced directional asymmetry, and plays a central role in joint calibration tasks and in theoretical interpretations of boundary phenomena in such diffusions.

3. SSR in Nonlocal Minimal Graphs: Geometric Scaling

In nonlocal minimal graph theory, antimetric (odd) perturbations of boundary data can force persistent "stickiness" at the boundary, manifesting as a robust lower bound for the solution inside the domain (Baronowitz et al., 2021). The "skew stickiness ratio" or skew stickiness exponent is defined as the exponent $\alpha$ characterizing the scaling of the stuck layer's height with respect to the amplitude $\eta$ of the antisymmetric bump:

$$\text{height of sticking} \sim \eta^\alpha, \qquad \alpha = \frac{2+\epsilon_0}{1-s}$$

for $s$-minimal graphs, with $s \in (0,1)$ and negligible constant $\epsilon_0 > 0$.

The result demonstrates that antisymmetric ("skew") stickiness exponents coincide with their symmetric counterparts, indicating that long-range nonlocal interactions maintain stickiness even under skewed boundary inputs.

4. SSR in Stochastic Volatility and Option Pricing Models

In financial mathematics, the SSR is a critical statistic relating the co-movement of spot returns and at-the-money (ATM) implied volatility to the ATM skew of the implied volatility surface (Fukasawa, 5 Feb 2026, Vargas et al., 2013, Jaber et al., 18 Mar 2025).

• General Definition: The continuous-time SSR at time $t$ for maturity $T$ is

$$R_t = \frac{1}{\sigma'_t} \frac{d\langle \sigma^S, \log S \rangle_t}{d\langle \log S \rangle_t}$$

where $\sigma^S = \sigma_t(S_t)$ is the spot-implied volatility and $\sigma'_t$ the ATM skew.

• Model-Free Characterization: Empirically, SSR measures the normalized slope of ATM implied vol versus log-returns—a proxy for "sticky strike" effects.
• Short and Long Maturity Limits: Linear models predict SSR values of $2$ as $T \to 0$ (short maturities) and $1$ as $T \to \infty$, with deviations above $2$ attributed to non-linear leverage mechanisms (e.g., asymmetric GARCH) (Vargas et al., 2013).
• Analytic Representations: Under Bergomi-type models, SSR has explicit characterizations using Malliavin calculus and the Itô–Wentzell/Clark–Ocone formulae (Fukasawa, 5 Feb 2026). The SSR can also be expressed via partial derivatives of the implied vol function in factor models (Jaber et al., 18 Mar 2025).

Table: Empirical and Model SSR Across Maturities in SPX Options (Jaber et al., 18 Mar 2025)

Maturity	Empirical SSR	Model-SSR (Quintic)
1 week	1.90	1.88
1 month	1.45	1.47
6 months	1.20	1.22
2 years	1.00	1.02

SSR exhibits strong maturity dependence—high for short maturities (sticky-strike regime), converging to 1 at long maturities (sticky-delta regime), and is robust to model calibration against observed data.

5. Mathematical Structures and Computational Procedures

The computation and interpretation of SSR depend on the context:

• Recurrence Statistics: Requires estimation of mean and standard deviation of return times for each cell, efficiently implemented using Welford's algorithm for online moment updates (Lozej, 2021).
• Sticky Diffusions: Involves high-frequency path functionals and occupation measures, with consistent estimators derived from sample-path statistics (Anagnostakis et al., 2024).
• Stochastic Volatility Models: SSR calculation can be carried out via Itô–Wentzell and Clark–Ocone representations, or via explicit finite-difference approximations in Markovian factor models (Fukasawa, 5 Feb 2026, Jaber et al., 18 Mar 2025).

For comparison across contexts, SSR plays the role of a dimensionless ratio capturing the "intensity" of stickiness relative to an axis of asymmetry or directional sensitivity.

6. Applications, Interpretation, and Empirical Implications

• Dynamical Systems: SSR isolates and quantifies the structure and strength of trapping regions, supporting phase space transport analysis and visualization.
• Diffusions and Barrier Processes: SSR enables direct calibration and inference of barrier friction versus asymmetrical drift, with immediate applications in boundary-layer physics and stochastic process understanding.
• Financial Engineering: SSR informs model risk diagnostics, the design of robust hedging strategies (via cross-gamma risk), and enables calibration of volatility models to observed market dynamics, particularly in reconciling the response of implied volatilities to spot moves (smile dynamics) (Vargas et al., 2013, Fukasawa, 5 Feb 2026).
• Model Selection and Calibration: Accurate SSR reproduction is now recognized as a minimum requirement for stochastic volatility models aspiring to replicate market-observed smile dynamics and cross-asset risk transmission (Jaber et al., 18 Mar 2025).

7. Theoretical Significance and Limitations

SSR, across its incarnations, is a compact, informative summary statistic capturing the nontrivial interplay between mixing, trapping, and asymmetry in both deterministic and stochastic systems. It is sensitive to fine details of time scale separation, mixture dynamics, nonlocal interactions, and leverage effects.

• Grid-dependence in Hamiltonian systems is monotonic, rendering qualitative features robust to resolution changes (Lozej, 2021).
• Inference Quality in diffusions is tied to occupation and local time estimation; asymptotic normality and rates for SSR estimators are not yet fully established (Anagnostakis et al., 2024).
• Nonlinearity Effects: Empirical values of SSR exceeding predicted linear model bounds indicate the necessity of incorporating nonlinearity, roughness, and higher-moment effects to fully capture realized smile dynamics (Vargas et al., 2013, Jaber et al., 18 Mar 2025).

SSR thus serves as a focal point for dissecting and improving models of stickiness and skew across dynamical, statistical, and financial domains.