voly 0.0.145__py3-none-any.whl → 0.0.147__py3-none-any.whl

voly/core/rnd.py

@@ -5,7 +5,7 @@ fitted volatility models and converting to probability functions.
 
 import numpy as np
 import pandas as pd
-from typing import Dict, List, Tuple, Optional, Union, Any
+from typing import Dict, List, Tuple, Optional, Union, Any, Callable
 from voly.utils.logger import logger, catch_exception
 from voly.exceptions import VolyError
 from voly.models import SVIModel
@@ -13,83 +13,240 @@ from voly.formulas import bs, d1, d2, get_domain
 from scipy import stats
 
 
-# Breeden-Litzenberger Method
 @catch_exception
-def breeden(domain_params, s, r, o, t, return_domain):
-    LM = get_domain(domain_params, s, r, o, t, 'log_moneyness')
-    M = get_domain(domain_params, s, r, o, t, 'moneyness')
-    R = get_domain(domain_params, s, r, o, t, 'returns')
-    K = get_domain(domain_params, s, r, o, t, 'strikes')
-    D = get_domain(domain_params, s, r, o, t, 'delta')
+def _prepare_domains(domain_params, s, r, o, t):
+    """
+    Calculate domain arrays for different representations (log_moneyness, moneyness, etc.).
 
-    c = bs(s, K, r, o, t, option_type='call')
-    c1 = np.gradient(c, K)
-    c2 = np.gradient(c1, K)
+    Parameters:
+    -----------
+    domain_params : tuple
+        (min_log_moneyness, max_log_moneyness, num_points)
+    s : float
+        Spot price
+    r : float
+        Risk-free rate
+    o : ndarray
+        Implied volatility array
+    t : float
+        Time to expiry in years
 
-    rnd_k = np.maximum(np.exp(r * t) * c2, 0)
-    rnd_lm = rnd_k * K
+    Returns:
+    --------
+    dict
+        Dictionary containing arrays for different domains
+    """
+    domains = {}
+    domains['log_moneyness'] = get_domain(domain_params, s, r, o, t, 'log_moneyness')
+    domains['moneyness'] = get_domain(domain_params, s, r, o, t, 'moneyness')
+    domains['returns'] = get_domain(domain_params, s, r, o, t, 'returns')
+    domains['strikes'] = get_domain(domain_params, s, r, o, t, 'strikes')
+    domains['delta'] = get_domain(domain_params, s, r, o, t, 'delta')
+
+    # Precompute differentials for integration
+    domains['dx'] = domains['log_moneyness'][1] - domains['log_moneyness'][0]
+
+    return domains
+
+
+@catch_exception
+def _normalize_density(pdf_values, dx):
+    """
+    Normalize a probability density function to integrate to 1.
+
+    Parameters:
+    -----------
+    pdf_values : ndarray
+        Array of PDF values
+    dx : float
+        Grid spacing
+
+    Returns:
+    --------
+    ndarray
+        Normalized PDF values
+    """
+    total_area = np.sum(pdf_values * dx)
+    if total_area <= 0:
+        logger.warning("PDF area is negative or zero, using absolute values")
+        total_area = np.sum(np.abs(pdf_values) * dx)
+
+    return pdf_values / total_area
+
+
+@catch_exception
+def _transform_to_domains(rnd_k, domains):
+    """
+    Transform density from strike domain to other domains.
+
+    Parameters:
+    -----------
+    rnd_k : ndarray
+        PDF in strike domain
+    domains : dict
+        Domain arrays
+
+    Returns:
+    --------
+    dict
+        Dictionary of PDFs in different domains
+    """
+    LM = domains['log_moneyness']
+    M = domains['moneyness']
+    K = domains['strikes']
+    R = domains['returns']
+    dx = domains['dx']
+
+    # Calculate PDF in different domains
+    rnd_lm = rnd_k * K  # Convert to log-moneyness domain
+    pdf_lm = _normalize_density(rnd_lm, dx)
 
-    dx = LM[1] - LM[0]
-    total_area = np.sum(rnd_lm * dx)
-    pdf_lm = rnd_lm / total_area
+    # Transform to other domains
     pdf_k = pdf_lm / K
-    pdf_m = pdf_k * s
+    pdf_m = pdf_k * domains['strikes'][0]  # s = K[0] * M[0]
     pdf_r = pdf_lm / (1 + R)
 
-    pdf_d1 = stats.norm.pdf(d1(s, K, r, o, t, option_type='call'))
-    dd_dK = pdf_d1 / (o * np.sqrt(t) * K)
+    # For delta domain, need special handling due to non-monotonicity
+    pdf_d1 = stats.norm.pdf(d1(domains['strikes'][0], K, 0, domains['delta'][0], 1, option_type='call'))
+    dd_dK = pdf_d1 / (domains['delta'][0] * np.sqrt(1) * K)
     pdf_d = pdf_k / dd_dK
 
-    cdf = np.cumsum(pdf_lm) * dx
-    cdf = cdf / cdf[-1]
-
-    if return_domain == 'log_moneyness':
-        x = LM
-        pdf = pdf_lm
-        moments = get_all_moments(x, pdf)
-        return pdf, cdf, x, moments
-    elif return_domain == 'moneyness':
-        x = M
-        pdf = pdf_m
-        moments = get_all_moments(x, pdf)
-        return pdf, cdf, x, moments
-    elif return_domain == 'returns':
-        x = R
-        pdf = pdf_r
-        moments = get_all_moments(x, pdf)
-        return pdf, cdf, x, moments
-    elif return_domain == 'strikes':
-        x = K
-        pdf = pdf_k
-        moments = get_all_moments(x, pdf)
-        return pdf, cdf, x, moments
-    elif return_domain == 'delta':
+    # Calculate CDF
+    cdf = np.cumsum(pdf_lm * dx)
+    cdf = np.minimum(cdf / cdf[-1], 1.0)  # Ensure max value is 1
+
+    return {
+        'log_moneyness': pdf_lm,
+        'moneyness': pdf_m,
+        'returns': pdf_r,
+        'strikes': pdf_k,
+        'delta': pdf_d,
+        'cdf': cdf
+    }
+
+
+@catch_exception
+def _select_domain_results(pdfs, domains, return_domain):
+    """
+    Select results for the requested domain.
+
+    Parameters:
+    -----------
+    pdfs : dict
+        PDFs in different domains
+    domains : dict
+        Domain arrays
+    return_domain : str
+        Requested domain
+
+    Returns:
+    --------
+    tuple
+        (pdf, cdf, x, moments)
+    """
+    if return_domain == 'delta':
+        # Special handling for delta domain due to potential non-monotonicity
+        D = domains['delta']
+        pdf_d = pdfs['delta']
         sort_idx = np.argsort(D)
         x = D[sort_idx]
         pdf = pdf_d[sort_idx]
-        moments = get_all_moments(x, pdf)
-        return pdf, cdf, x, moments
+    else:
+        x = domains[return_domain]
+        pdf = pdfs[return_domain]
+
+    moments = get_all_moments(x, pdf)
+    return pdf, pdfs['cdf'], x, moments
+
+
+@catch_exception
+def breeden(domain_params, s, r, o, t, return_domain):
+    """
+    Breeden-Litzenberger method for RND estimation.
+
+    Parameters:
+    -----------
+    domain_params : tuple
+        (min_log_moneyness, max_log_moneyness, num_points)
+    s : float
+        Spot price
+    r : float
+        Risk-free rate
+    o : ndarray
+        Implied volatility array
+    t : float
+        Time to expiry in years
+    return_domain : str
+        Domain for results ('log_moneyness', 'moneyness', 'returns', 'strikes', 'delta')
+
+    Returns:
+    --------
+    tuple
+        (pdf, cdf, x, moments)
+    """
+    # Prepare domain arrays
+    domains = _prepare_domains(domain_params, s, r, o, t)
+    K = domains['strikes']
+
+    # Calculate option prices and derivatives
+    c = bs(s, K, r, o, t, option_type='call')
+    c1 = np.gradient(c, K)
+    c2 = np.gradient(c1, K)
+
+    # Calculate RND in strike domain and apply discount factor
+    rnd_k = np.maximum(np.exp(r * t) * c2, 0)
+
+    # Transform to other domains
+    pdfs = _transform_to_domains(rnd_k, domains)
+
+    # Return results for requested domain
+    return _select_domain_results(pdfs, domains, return_domain)
 
 
-# Rookley's Method
 @catch_exception
 def rookley(domain_params, s, r, o, t, return_domain):
-    LM = get_domain(domain_params, s, r, o, t, 'log_moneyness')
-    M = get_domain(domain_params, s, r, o, t, 'moneyness')
-    R = get_domain(domain_params, s, r, o, t, 'returns')
-    K = get_domain(domain_params, s, r, o, t, 'strikes')
-    D = get_domain(domain_params, s, r, o, t, 'delta')
+    """
+    Rookley method for RND estimation, using volatility smile derivatives.
+
+    Parameters:
+    -----------
+    domain_params : tuple
+        (min_log_moneyness, max_log_moneyness, num_points)
+    s : float
+        Spot price
+    r : float
+        Risk-free rate
+    o : ndarray
+        Implied volatility array
+    t : float
+        Time to expiry in years
+    return_domain : str
+        Domain for results ('log_moneyness', 'moneyness', 'returns', 'strikes', 'delta')
+
+    Returns:
+    --------
+    tuple
+        (pdf, cdf, x, moments)
+    """
+    # Prepare domain arrays
+    domains = _prepare_domains(domain_params, s, r, o, t)
+    M = domains['moneyness']
+    K = domains['strikes']
 
+    # Calculate volatility derivatives with respect to moneyness
     o1 = np.gradient(o, M)
     o2 = np.gradient(o1, M)
 
+    # Precompute common terms
     st = np.sqrt(t)
     rt = r * t
     ert = np.exp(rt)
 
-    n_d1 = (np.log(M) + (r + 1 / 2 * o ** 2) * t) / (o * st)
+    # Calculate Black-Scholes d1 and d2 terms
+    n_d1 = (np.log(M) + (r + 0.5 * o ** 2) * t) / (o * st)
     n_d2 = n_d1 - o * st
 
+    # Calculate various derivatives needed for the density
     del_d1_M = 1 / (M * o * st)
     del_d2_M = del_d1_M
     del_d1_o = -(np.log(M) + rt) / (o ** 2 * st) + st / 2
@@ -98,226 +255,163 @@ def rookley(domain_params, s, r, o, t, return_domain):
     d_d1_M = del_d1_M + del_d1_o * o1
     d_d2_M = del_d2_M + del_d2_o * o1
 
+    # Complex second derivatives
     dd_d1_M = (
             -(1 / (M * o * st)) * (1 / M + o1 / o)
             + o2 * (st / 2 - (np.log(M) + rt) / (o ** 2 * st))
             + o1 * (2 * o1 * (np.log(M) + rt) / (o ** 3 * st) - 1 / (M * o ** 2 * st))
     )
+
     dd_d2_M = (
             -(1 / (M * o * st)) * (1 / M + o1 / o)
             - o2 * (st / 2 + (np.log(M) + rt) / (o ** 2 * st))
             + o1 * (2 * o1 * (np.log(M) + rt) / (o ** 3 * st) - 1 / (M * o ** 2 * st))
     )
 
-    d_c_M = stats.norm.pdf(n_d1) * d_d1_M - 1 / ert * stats.norm.pdf(n_d2) / M * d_d2_M + 1 / ert * stats.norm.cdf(n_d2) / (
-                M ** 2)
+    # Call price derivatives with respect to moneyness
+    d_c_M = (
+            stats.norm.pdf(n_d1) * d_d1_M
+            - (1 / ert) * stats.norm.pdf(n_d2) / M * d_d2_M
+            + (1 / ert) * stats.norm.cdf(n_d2) / (M ** 2)
+    )
+
     dd_c_M = (
             stats.norm.pdf(n_d1) * (dd_d1_M - n_d1 * d_d1_M ** 2)
             - stats.norm.pdf(n_d2) / (ert * M) * (dd_d2_M - 2 / M * d_d2_M - n_d2 * d_d2_M ** 2)
             - 2 * stats.norm.cdf(n_d2) / (ert * M ** 3)
     )
 
+    # Convert from moneyness to strike derivatives
     dd_c_K = dd_c_M * (M / K) ** 2 + 2 * d_c_M * (M / K ** 2)
 
+    # Calculate RND in strike domain and apply discount factor
     rnd_k = np.maximum(ert * s * dd_c_K, 0)
-    rnd_lm = rnd_k * K
 
-    dx = LM[1] - LM[0]
-    total_area = np.sum(rnd_lm * dx)
-    pdf_lm = rnd_lm / total_area
-    pdf_k = pdf_lm / K
-    pdf_m = pdf_k * s
-    pdf_r = pdf_lm / (1 + R)
+    # Transform to other domains
+    pdfs = _transform_to_domains(rnd_k, domains)
 
-    pdf_d1 = stats.norm.pdf(d1(s, K, r, o, t, option_type='call'))
-    dd_dK = pdf_d1 / (o * np.sqrt(t) * K)
-    pdf_d = pdf_k / dd_dK
+    # Return results for requested domain
+    return _select_domain_results(pdfs, domains, return_domain)
 
-    cdf = np.cumsum(pdf_lm) * dx
-    cdf = cdf / cdf[-1]
-
-    if return_domain == 'log_moneyness':
-        x = LM
-        pdf = pdf_lm
-        moments = get_all_moments(x, pdf)
-        return pdf, cdf, x, moments
-    elif return_domain == 'moneyness':
-        x = M
-        pdf = pdf_m
-        moments = get_all_moments(x, pdf)
-        return pdf, cdf, x, moments
-    elif return_domain == 'returns':
-        x = R
-        pdf = pdf_r
-        moments = get_all_moments(x, pdf)
-        return pdf, cdf, moments
-    elif return_domain == 'strikes':
-        x = K
-        pdf = pdf_k
-        moments = get_all_moments(x, pdf)
-        return pdf, cdf, x, moments
-    elif return_domain == 'delta':
-        sort_idx = np.argsort(D)
-        x = D[sort_idx]
-        pdf = pdf_d[sort_idx]
-        moments = get_all_moments(x, pdf)
-        return pdf, cdf, x, moments
 
-'''
 @catch_exception
 def get_all_moments(x, pdf, model_params=None):
-    mean = np.trapz(x * pdf, x)  # E[X]
-    median = x[np.searchsorted(np.cumsum(pdf * np.diff(x, prepend=x[0])), 0.5)]  # Median (50th percentile)
-    mode = x[np.argmax(pdf)]  # Mode (peak of PDF)
-    variance = np.trapz((x - mean) ** 2 * pdf, x)  # Var[X] = E[(X - μ)^2]
-    std_dev = np.sqrt(variance)  # Standard deviation
-    skewness = np.trapz((x - mean) ** 3 * pdf, x) / std_dev ** 3  # Skewness
-    kurtosis = np.trapz((x - mean) ** 4 * pdf, x) / std_dev ** 4  # Kurtosis
-    excess_kurtosis = kurtosis - 3  # Excess kurtosis (relative to normal dist.)
-    q25 = x[np.searchsorted(np.cumsum(pdf * np.diff(x, prepend=x[0])), 0.25)]  # 25th percentile
-    q75 = x[np.searchsorted(np.cumsum(pdf * np.diff(x, prepend=x[0])), 0.75)]  # 75th percentile
-    iqr = q75 - q25  # Inter-quartile range
-    entropy = -np.trapz(pdf * np.log(pdf + 1e-10), x)  # Differential entropy (avoid log(0))
-
-    # Full Z-score areas
-    dx = np.diff(x, prepend=x[0])
-    z = (x - mean) / std_dev
-    o1p = np.sum(pdf[(z > 0) & (z < 1)] * dx[(z > 0) & (z < 1)])
-    o2p = np.sum(pdf[(z >= 1) & (z < 2)] * dx[(z >= 1) & (z < 2)])
-    o3p = np.sum(pdf[(z >= 2) & (z < 3)] * dx[(z >= 2) & (z < 3)])
-    o4p = np.sum(pdf[z >= 3] * dx[z >= 3])
-    o1n = np.sum(pdf[(z < 0) & (z > -1)] * dx[(z < 0) & (z > -1)])
-    o2n = np.sum(pdf[(z <= -1) & (z > -2)] * dx[(z <= -1) & (z > -2)])
-    o3n = np.sum(pdf[(z <= -2) & (z > -3)] * dx[(z <= -2) & (z > -3)])
-    o4n = np.sum(pdf[z <= -3] * dx[z <= -3])
-
-    moments = {
-        'mean': mean,
-        'median': median,
-        'mode': mode,
-        'variance': variance,
-        'std_dev': std_dev,
-        'skewness': skewness,
-        'kurtosis': kurtosis,
-        'excess_kurtosis': excess_kurtosis,
-        'q25': q25,
-        'q75': q75,
-        'iqr': iqr,
-        'entropy': entropy,
-        'o1p': o1p,
-        'o2p': o2p,
-        'o3p': o3p,
-        'o4p': o4p,
-        'o1n': o1n,
-        'o2n': o2n,
-        'o3n': o3n,
-        'o4n': o4n
-    }
-
-    # Add model parameters if provided
-    if model_params is not None:
-        moments.update(model_params)
+    """
+    Calculate statistical moments and other distributional properties.
 
-    return moments
-'''
+    Parameters:
+    -----------
+    x : ndarray
+        Domain values
+    pdf : ndarray
+        Probability density values
+    model_params : dict, optional
+        Additional model parameters to include in the results
 
+    Returns:
+    --------
+    dict
+        Dictionary of calculated moments and properties
+    """
+    # Skip calculation for invalid inputs
+    if len(x) != len(pdf) or len(x) < 3:
+        logger.warning("Invalid inputs for moment calculation")
+        return {}
 
-@catch_exception
-def get_all_moments(x, pdf, model_params=None):
-    # Precompute dx for integration
+    # Compute dx for integration
     dx = np.diff(x, prepend=x[0])
 
+    # Ensure the PDF integrates to 1
+    pdf_normalized = pdf / np.trapz(pdf, x)
+
     # Raw Moments (μ_k = E[X^k])
-    raw_0 = np.trapz(pdf, x)           # Zeroth (~1)
-    raw_1 = np.trapz(x * pdf, x)       # First (mean)
-    raw_2 = np.trapz(x**2 * pdf, x)    # Second
-    raw_3 = np.trapz(x**3 * pdf, x)    # Third
-    raw_4 = np.trapz(x**4 * pdf, x)    # Fourth
-    raw_5 = np.trapz(x**5 * pdf, x)    # Fifth
-    raw_6 = np.trapz(x**6 * pdf, x)    # Sixth
-
-    mean = raw_1
-    variance = np.trapz((x - mean)**2 * pdf, x)  # m_2
-    std_dev = np.sqrt(variance)
+    raw_moments = {
+        'raw_0': np.trapz(pdf_normalized, x),  # Zeroth (~1)
+        'raw_1': np.trapz(x * pdf_normalized, x),  # First (mean)
+        'raw_2': np.trapz(x ** 2 * pdf_normalized, x),  # Second
+        'raw_3': np.trapz(x ** 3 * pdf_normalized, x),  # Third
+        'raw_4': np.trapz(x ** 4 * pdf_normalized, x),  # Fourth
+        'raw_5': np.trapz(x ** 5 * pdf_normalized, x),  # Fifth
+        'raw_6': np.trapz(x ** 6 * pdf_normalized, x),  # Sixth
+    }
+
+    # Derived statistics
+    mean = raw_moments['raw_1']
+    variance = np.trapz((x - mean) ** 2 * pdf_normalized, x)
+    std_dev = np.sqrt(max(variance, 1e-10))  # Prevent division by zero
 
     # Central Moments (m_k = E[(X - μ)^k])
-    cent_0 = raw_0                     # Zeroth (~1)
-    cent_1 = np.trapz((x - mean) * pdf, x)  # First (~0)
-    cent_2 = variance                  # Second (variance)
-    cent_3 = np.trapz((x - mean)**3 * pdf, x)  # Third
-    cent_4 = np.trapz((x - mean)**4 * pdf, x)  # Fourth
-    cent_5 = np.trapz((x - mean)**5 * pdf, x)  # Fifth
-    cent_6 = np.trapz((x - mean)**6 * pdf, x)  # Sixth
+    cent_moments = {
+        'cent_1': 0,  # Theoretically zero
+        'cent_2': variance,  # Second (variance)
+        'cent_3': np.trapz((x - mean) ** 3 * pdf_normalized, x),  # Third
+        'cent_4': np.trapz((x - mean) ** 4 * pdf_normalized, x),  # Fourth
+        'cent_5': np.trapz((x - mean) ** 5 * pdf_normalized, x),  # Fifth
+        'cent_6': np.trapz((x - mean) ** 6 * pdf_normalized, x),  # Sixth
+    }
 
     # Standardized Moments (m̄_k = E[((X - μ)/σ)^k])
     z = (x - mean) / std_dev
-    std_0 = np.trapz(pdf, x)           # Zeroth (~1)
-    std_1 = np.trapz(z * pdf, x)       # First (~0)
-    std_2 = np.trapz(z**2 * pdf, x)    # Second (~1)
-    std_3 = np.trapz(z**3 * pdf, x)    # Skewness
-    std_4 = np.trapz(z**4 * pdf, x)    # Kurtosis
-    std_5 = np.trapz(z**5 * pdf, x)    # Fifth
-    std_6 = np.trapz(z**6 * pdf, x)    # Sixth
-
-    # Extra statistics
-    cdf = np.cumsum(pdf * dx)
-    median = x[np.searchsorted(cdf, 0.5)]  # Median
-    excess_kurtosis = std_4 - 3
-    q25 = x[np.searchsorted(cdf, 0.25)]    # 25th percentile
-    q75 = x[np.searchsorted(cdf, 0.75)]    # 75th percentile
+    std_moments = {
+        'std_3': np.trapz(z ** 3 * pdf_normalized, x),  # Skewness
+        'std_4': np.trapz(z ** 4 * pdf_normalized, x),  # Kurtosis
+        'std_5': np.trapz(z ** 5 * pdf_normalized, x),  # Fifth
+        'std_6': np.trapz(z ** 6 * pdf_normalized, x),  # Sixth
+    }
+
+    # Calculate CDF for quantiles
+    cdf = np.cumsum(pdf_normalized * dx)
+    cdf = cdf / cdf[-1]  # Normalize
+
+    # Quantiles and other statistics
+    mode_idx = np.argmax(pdf)
+    mode = x[mode_idx] if 0 <= mode_idx < len(x) else mean
+
+    # Find percentiles
+    q25_idx = np.searchsorted(cdf, 0.25)
+    q50_idx = np.searchsorted(cdf, 0.50)
+    q75_idx = np.searchsorted(cdf, 0.75)
+
+    q25 = x[q25_idx] if 0 <= q25_idx < len(x) else np.nan
+    median = x[q50_idx] if 0 <= q50_idx < len(x) else np.nan
+    q75 = x[q75_idx] if 0 <= q75_idx < len(x) else np.nan
     iqr = q75 - q25
-    entropy = -np.trapz(pdf * np.log(pdf + 1e-10), x)
-
-    # Z-score areas
-    o1p = np.sum(pdf[(z > 0) & (z < 1)] * dx[(z > 0) & (z < 1)])
-    o2p = np.sum(pdf[(z >= 1) & (z < 2)] * dx[(z >= 1) & (z < 2)])
-    o3p = np.sum(pdf[(z >= 2) & (z < 3)] * dx[(z >= 2) & (z < 3)])
-    o4p = np.sum(pdf[z >= 3] * dx[z >= 3])
-    o1n = np.sum(pdf[(z < 0) & (z > -1)] * dx[(z < 0) & (z > -1)])
-    o2n = np.sum(pdf[(z <= -1) & (z > -2)] * dx[(z <= -1) & (z > -2)])
-    o3n = np.sum(pdf[(z <= -2) & (z > -3)] * dx[(z <= -2) & (z > -3)])
-    o4n = np.sum(pdf[z <= -3] * dx[z <= -3])
-
-    # Combine results as flat columns
-    moments = {
-        'raw_0': raw_0,
-        'raw_1': raw_1,
-        'raw_2': raw_2,
-        'raw_3': raw_3,
-        'raw_4': raw_4,
-        'raw_5': raw_5,
-        'raw_6': raw_6,
-        'cent_0': cent_0,
-        'cent_1': cent_1,
-        'cent_2': cent_2,
-        'cent_3': cent_3,
-        'cent_4': cent_4,
-        'cent_5': cent_5,
-        'cent_6': cent_6,
-        'std_0': std_0,
-        'std_1': std_1,
-        'std_2': std_2,
-        'std_3': std_3,
-        'std_4': std_4,
-        'std_5': std_5,
-        'std_6': std_6,
-        'median': median,
+
+    # Information theory measures
+    entropy = -np.trapz(pdf_normalized * np.log(pdf_normalized + 1e-10), x)
+
+    # Z-score areas (probability mass in standard deviation regions)
+    z_areas = {
+        'o1p': np.sum(pdf_normalized[(z > 0) & (z < 1)] * dx[(z > 0) & (z < 1)]),
+        'o2p': np.sum(pdf_normalized[(z >= 1) & (z < 2)] * dx[(z >= 1) & (z < 2)]),
+        'o3p': np.sum(pdf_normalized[(z >= 2) & (z < 3)] * dx[(z >= 2) & (z < 3)]),
+        'o4p': np.sum(pdf_normalized[z >= 3] * dx[z >= 3]),
+        'o1n': np.sum(pdf_normalized[(z < 0) & (z > -1)] * dx[(z < 0) & (z > -1)]),
+        'o2n': np.sum(pdf_normalized[(z <= -1) & (z > -2)] * dx[(z <= -1) & (z > -2)]),
+        'o3n': np.sum(pdf_normalized[(z <= -2) & (z > -3)] * dx[(z <= -2) & (z > -3)]),
+        'o4n': np.sum(pdf_normalized[z <= -3] * dx[z <= -3]),
+    }
+
+    # Common statistic names
+    common_stats = {
+        'mean': mean,
+        'variance': variance,
         'std_dev': std_dev,
-        'excess_kurtosis': excess_kurtosis,
+        'skewness': std_moments['std_3'],
+        'kurtosis': std_moments['std_4'],
+        'excess_kurtosis': std_moments['std_4'] - 3,
+        'median': median,
+        'mode': mode,
         'q25': q25,
         'q75': q75,
         'iqr': iqr,
         'entropy': entropy,
-        'o1p': o1p,
-        'o2p': o2p,
-        'o3p': o3p,
-        'o4p': o4p,
-        'o1n': o1n,
-        'o2n': o2n,
-        'o3n': o3n,
-        'o4n': o4n
     }
 
+    # Combine all statistics
+    moments = {**raw_moments, **cent_moments, **std_moments, **z_areas, **common_stats}
+
+    # Add model parameters if provided
     if model_params is not None:
         moments.update(model_params)
 
@@ -328,70 +422,90 @@ def get_all_moments(x, pdf, model_params=None):
 def get_rnd_surface(model_results: pd.DataFrame,
                     domain_params: Tuple[float, float, int] = (-1.5, 1.5, 1000),
                     return_domain: str = 'log_moneyness',
-                    method: str = 'rookley') -> Dict[str, np.ndarray]:
+                    method: str = 'rookley') -> Dict[str, Any]:
     """
-    Generate RND surface from vol smile parameters.
-
-    Works with both regular fit_results and interpolated_results dataframes.
+    Generate risk-neutral density surface from volatility surface parameters.
 
     Parameters:
-    - model_results: DataFrame from fit_model() or interpolate_model(). Maturity names or DTM as Index
-    - domain_params: Tuple of (min, max, num_points) for the x-domain grid
-    - return_domain: Domain for x-axis values ('log_moneyness', 'moneyness', 'returns', 'strikes', 'delta')
-    - method: 'rookley' or 'breeden'
+    -----------
+    model_results : pd.DataFrame
+        DataFrame from fit_model() or interpolate_model() with SVI parameters
+    domain_params : tuple
+        (min_log_moneyness, max_log_moneyness, num_points)
+    return_domain : str
+        Domain for results ('log_moneyness', 'moneyness', 'returns', 'strikes', 'delta')
+    method : str
+        Method for RND estimation ('rookley' or 'breeden')
 
     Returns:
-    - Tuple containing:
-      - pdf_surface: Dictionary mapping maturity/dtm names to PDF arrays of their requested domain
-      - cdf_surface: Dictionary mapping maturity/dtm names to CDF arrays
-      - x_surface: Dictionary mapping maturity/dtm names to requested x domain arrays
-      - moments_df: DataFrame with moments of the distributions using model_results index
+    --------
+    dict
+        Dictionary containing pdf_surface, cdf_surface, x_surface, and moments
     """
-    # Check if required columns are present
+    # Validate inputs
     required_columns = ['s', 'a', 'b', 'sigma', 'm', 'rho', 't', 'r']
     missing_columns = [col for col in required_columns if col not in model_results.columns]
     if missing_columns:
         raise VolyError(f"Required columns missing in model_results: {missing_columns}")
 
+    # Validate method
+    if method not in ['rookley', 'breeden']:
+        raise VolyError(f"Invalid method: {method}. Must be 'rookley' or 'breeden'")
+
+    # Validate return_domain
+    valid_domains = ['log_moneyness', 'moneyness', 'returns', 'strikes', 'delta']
+    if return_domain not in valid_domains:
+        raise VolyError(f"Invalid return_domain: {return_domain}. Must be one of {valid_domains}")
+
+    # Select method function
+    rnd_method = rookley if method == 'rookley' else breeden
+
+    # Initialize result containers
     pdf_surface = {}
     cdf_surface = {}
     x_surface = {}
     all_moments = {}
 
-    # Process each maturity/dtm
+    # Process each maturity/expiry
     for i in model_results.index:
-        # Calculate SVI total implied variance and convert to IV
-        params = [
-            model_results.loc[i, 'a'],
-            model_results.loc[i, 'b'],
-            model_results.loc[i, 'sigma'],
-            model_results.loc[i, 'rho'],
-            model_results.loc[i, 'm']
-        ]
-        s = model_results.loc[i, 's']
-        r = model_results.loc[i, 'r']
-        t = model_results.loc[i, 't']
-
-        # Calculate implied volatility
-        LM = np.linspace(domain_params[0], domain_params[1], domain_params[2])
-        w = np.array([SVIModel.svi(x, *params) for x in LM])
-        o = np.sqrt(w / t)
-
-        if method == 'rookley':
-            pdf, cdf, x, moments = rookley(domain_params, s, r, o, t, return_domain)
-        else:
-            pdf, cdf, x, moments = breeden(domain_params, s, r, o, t, return_domain)
-
-        pdf_surface[i] = pdf
-        cdf_surface[i] = cdf
-        x_surface[i] = x
-        all_moments[i] = moments
-
-    # Create a DataFrame with moments using the same index as model_results
+        try:
+            # Extract SVI parameters for this maturity
+            params = [
+                model_results.loc[i, 'a'],
+                model_results.loc[i, 'b'],
+                model_results.loc[i, 'sigma'],
+                model_results.loc[i, 'rho'],
+                model_results.loc[i, 'm']
+            ]
+            s = model_results.loc[i, 's']
+            r = model_results.loc[i, 'r']
+            t = model_results.loc[i, 't']
+
+            # Calculate implied volatility surface from SVI parameters
+            LM = np.linspace(domain_params[0], domain_params[1], domain_params[2])
+            w = np.array([SVIModel.svi(x, *params) for x in LM])
+            o = np.sqrt(w / t)
+
+            # Calculate RND using the selected method
+            pdf, cdf, x, moments = rnd_method(domain_params, s, r, o, t, return_domain)
+
+            # Store results
+            pdf_surface[i] = pdf
+            cdf_surface[i] = cdf
+            x_surface[i] = x
+            all_moments[i] = moments
+
+        except Exception as e:
+            logger.warning(f"Failed to calculate RND for maturity {i}: {str(e)}")
+
+    # Check if we have any valid results
+    if not pdf_surface:
+        raise VolyError("No valid densities could be calculated. Check your input data.")
+
+    # Create DataFrame with moments
     moments = pd.DataFrame(all_moments).T
 
-    # Ensure the index matches the model_results index
-    moments.index = model_results.index
+    logger.info(f"RND surface calculation complete using {method} method")
 
     return {
         'pdf_surface': pdf_surface,