xjtulyc/options-pricing

Options Pricing

TL;DR — Price European and American options using Black-Scholes analytics, Monte Carlo simulation with antithetic variates and control variates, and CRR binomial trees. Compute all Greeks, extract implied volatility with Brent's method, and visualize the implied volatility surface and smile.

---

When to Use

| Situation | Recommended Method | |---|---| | European call/put, closed-form price | Black-Scholes formula | | All five Greeks for a European option | Analytical BS Greeks | | Confidence interval around MC price | Monte Carlo + standard error | | Reduce MC variance | Antithetic variates / control variate | | American option (early exercise) | Binomial CRR tree (backward induction) | | Extract market-implied vol from option price | Brent root-finding on BS formula | | Visualize vol smile/surface | Plot IV by strike and maturity |

---

Background

Black-Scholes-Merton Model (1973)

Assumes the underlying S_t follows GBM: dS_t = μ S_t dt + σ S_t dW_t

European call price: C = S N(d₁) - K e^{-rT} N(d₂) d₁ = [ln(S/K) + (r + σ²/2)T] / (σ√T) d₂ = d₁ - σ√T

European put: P = K e^{-rT} N(-d₂) - S N(-d₁)

Greeks

| Greek | Definition | Formula | |---|---|---| | Delta | ∂C/∂S | N(d₁) (call) | | Gamma | ∂²C/∂S² | φ(d₁)/(S σ √T) | | Vega | ∂C/∂σ | S φ(d₁) √T | | Theta | ∂C/∂t | -(S φ(d₁) σ)/(2√T) - rK e^{-rT} N(d₂) | | Rho | ∂C/∂r | K T e^{-rT} N(d₂) (call) |

Monte Carlo Variance Reduction

Antithetic variates: For each path ε_t, also simulate -ε_t. Pair estimates and average. Reduces variance by exploiting negative correlation between paired paths.

Control variate: Use BS analytical price as control: C_CV = C_MC + β (C_BS - E[C_BS]) where β = Cov(C_MC, C_control) / Var(C_control) ≈ 1 for European options.

Effective variance reduction factor can be 5–50x depending on the payoff.

CRR Binomial Tree

Cox-Ross-Rubinstein (1979) recombining tree:

• Up factor: u = e^{σ√Δt}
• Down factor: d = 1/u
• Risk-neutral probability: p = (e^{rΔt} - d) / (u - d)

American option: at each node, take the maximum of intrinsic value and continuation value (backward induction). European option: only backward induction without early exercise check.

Implied Volatility

Given observed market price C_mkt, find σ_IV such that: BS(S, K, T, r, σ_IV) = C_mkt

Solved numerically via scipy.optimize.brentq (Brent's method) over σ ∈ (0.001, 10).

---

Environment Setup

conda create -n options python=3.11 -y
conda activate options
pip install numpy>=1.23 scipy>=1.9 matplotlib>=3.6 pandas>=1.5

python -c "import numpy, scipy, matplotlib; print('All packages OK')"

No external financial data API required. All examples use synthetic parameters.

---

Core Workflow

Step 1 — Black-Scholes Pricing and All Greeks

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from scipy.stats import norm
from scipy.optimize import brentq
import warnings


def bs_price(
    S: float,
    K: float,
    T: float,
    r: float,
    sigma: float,
    option_type: str = "call",
) -> float:
    """
    Black-Scholes-Merton European option price.

    Args:
        S:           Current underlying price.
        K:           Strike price.
        T:           Time to expiry in years.
        r:           Continuously compounded risk-free rate.
        sigma:       Implied/historical volatility (annualized).
        option_type: 'call' or 'put'.

    Returns:
        Option price.
    """
    if T <= 0:
        if option_type == "call":
            return max(S - K, 0)
        return max(K - S, 0)

    d1 = (np.log(S / K) + (r + 0.5 * sigma**2) * T) / (sigma * np.sqrt(T))
    d2 = d1 - sigma * np.sqrt(T)

    if option_type == "call":
        return S * norm.cdf(d1) - K * np.exp(-r * T) * norm.cdf(d2)
    elif option_type == "put":
        return K * np.exp(-r * T) * norm.cdf(-d2) - S * norm.cdf(-d1)
    else:
        raise ValueError(f"option_type must be 'call' or 'put', got '{option_type}'")


def bs_greeks(
    S: float,
    K: float,
    T: float,
    r: float,
    sigma: float,
    option_type: str = "call",
) -> dict:
    """
    Compute all five Black-Scholes Greeks analytically.

    Args:
        S:           Underlying price.
        K:           Strike price.
        T:           Time to expiry in years.
        r:           Risk-free rate (continuously compounded).
        sigma:       Volatility (annualized).
        option_type: 'call' or 'put'.

    Returns:
        Dictionary with keys: delta, gamma, vega, theta, rho.
        Theta is per calendar day (divided by 365).
    """
    d1 = (np.log(S / K) + (r + 0.5 * sigma**2) * T) / (sigma * np.sqrt(T))
    d2 = d1 - sigma * np.sqrt(T)
    phi_d1 = norm.pdf(d1)

    gamma = phi_d1 / (S * sigma * np.sqrt(T))
    vega = S * phi_d1 * np.sqrt(T)

    if option_type == "call":
        delta = norm.cdf(d1)
        theta = (-(S * phi_d1 * sigma) / (2 * np.sqrt(T))
                 - r * K * np.exp(-r * T) * norm.cdf(d2)) / 365
        rho = K * T * np.exp(-r * T) * norm.cdf(d2)
    elif option_type == "put":
        delta = norm.cdf(d1) - 1
        theta = (-(S * phi_d1 * sigma) / (2 * np.sqrt(T))
                 + r * K * np.exp(-r * T) * norm.cdf(-d2)) / 365
        rho = -K * T * np.exp(-r * T) * norm.cdf(-d2)
    else:
        raise ValueError(f"option_type must be 'call' or 'put'")

    return {"delta": delta, "gamma": gamma, "vega": vega,
            "theta": theta, "rho": rho}


def print_greeks_table(
    S: float = 100.0,
    K: float = 100.0,
    T: float = 1.0,
    r: float = 0.05,
    sigma: float = 0.20,
) -> pd.DataFrame:
    """
    Print a formatted table of BS prices and Greeks for call and put.

    Args:
        S, K, T, r, sigma: Standard Black-Scholes parameters.

    Returns:
        DataFrame with rows for call and put, columns for price and Greeks.
    """
    call_price = bs_price(S, K, T, r, sigma, "call")
    put_price = bs_price(S, K, T, r, sigma, "put")
    call_greeks = bs_greeks(S, K, T, r, sigma, "call")
    put_greeks = bs_greeks(S, K, T, r, sigma, "put")

    parity_check = call_price - put_price - (S - K * np.exp(-r * T))
    print(f"Put-Call Parity check: {parity_check:.2e}  (should be ~0)")

    rows = []
    for opt_type, price, greeks in [("call", call_price, call_greeks),
                                     ("put", put_price, put_greeks)]:
        rows.append({
            "type": opt_type,
            "price": round(price, 4),
            "delta": round(greeks["delta"], 4),
            "gamma": round(greeks["gamma"], 4),
            "vega": round(greeks["vega"], 4),
            "theta (per day)": round(greeks["theta"], 4),
            "rho": round(greeks["rho"], 4),
        })

    df = pd.DataFrame(rows)
    print(f"\nBS Option Parameters: S={S}, K={K}, T={T}y, r={r:.2%}, σ={sigma:.2%}")
    print(df.to_string(index=False))
    return df

Step 2 — Monte Carlo with Antithetic and Control Variates

def mc_option_price(
    S: float,
    K: float,
    T: float,
    r: float,
    sigma: float,
    option_type: str = "call",
    n_paths: int = 100_000,
    n_steps: int = 1,
    use_antithetic: bool = True,
    use_control_variate: bool = True,
    seed: int = 42,
) -> dict:
    """
    Monte Carlo European option pricing with variance reduction.

    Args:
        S:                   Initial underlying price.
        K:                   Strike price.
        T:                   Time to expiry (years).
        r:                   Risk-free rate.
        sigma:               Volatility.
        option_type:         'call' or 'put'.
        n_paths:             Number of GBM simulation paths.
        n_steps:             Time steps per path (1 = log-normal terminal price).
        use_antithetic:      Apply antithetic variates variance reduction.
        use_control_variate: Apply control variate (BS price).
        seed:                Random seed.

    Returns:
        Dictionary with keys: price, se, ci_lower, ci_upper, bs_price, variance_reduction.
    """
    rng = np.random.default_rng(seed)
    dt = T / n_steps

    if use_antithetic:
        half = n_paths // 2
        Z = rng.standard_normal((half, n_steps))
        Z_all = np.vstack([Z, -Z])
    else:
        Z_all = rng.standard_normal((n_paths, n_steps))

    # GBM simulation
    log_S = np.log(S) + np.cumsum(
        (r - 0.5 * sigma**2) * dt + sigma * np.sqrt(dt) * Z_all, axis=1
    )
    S_T = np.exp(log_S[:, -1])

    # Payoff
    if option_type == "call":
        payoffs = np.maximum(S_T - K, 0)
    else:
        payoffs = np.maximum(K - S_T, 0)

    # Discount
    payoffs_disc = np.exp(-r * T) * payoffs
    price_plain = float(np.mean(payoffs_disc))

    # Control variate: BS price as control
    if use_control_variate:
        bs_ref = bs_price(S, K, T, r, sigma, option_type)
        cov_matrix = np.cov(payoffs_disc, payoffs_disc)  # Placeholder; use S_T as control
        # Use S_T as a traded asset: known E[e^{-rT} S_T] = S (forward price)
        s_t_disc = np.exp(-r * T) * S_T
        beta = np.cov(payoffs_disc, s_t_disc)[0, 1] / np.var(s_t_disc)
        payoffs_cv = payoffs_disc - beta * (s_t_disc - S)
        price_cv = float(np.mean(payoffs_cv))
        se_cv = float(np.std(payoffs_cv) / np.sqrt(n_paths))
        var_reduction = float(np.var(payoffs_disc) / np.var(payoffs_cv))
        price_final = price_cv
        se_final = se_cv
    else:
        bs_ref = bs_price(S, K, T, r, sigma, option_type)
        var_reduction = 1.0
        price_final = price_plain
        se_final = float(np.std(payoffs_disc) / np.sqrt(n_paths))

    ci_lower = price_final - 1.96 * se_final
    ci_upper = price_final + 1.96 * se_final

    print(f"\nMC {option_type.upper()} Price: {price_final:.4f}  SE: {se_final:.4f}")
    print(f"  95% CI: [{ci_lower:.4f}, {ci_upper:.4f}]")
    print(f"  BS price: {bs_ref:.4f}  |  Variance reduction factor: {var_reduction:.2f}x")

    return {"price": price_final, "se": se_final,
            "ci_lower": ci_lower, "ci_upper": ci_upper,
            "bs_price": bs_ref, "variance_reduction": var_reduction}

Step 3 — Implied Volatility Surface

def implied_vol(
    market_price: float,
    S: float,
    K: float,
    T: float,
    r: float,
    option_type: str = "call",
    tol: float = 1e-8,
) -> float:
    """
    Extract implied volatility from an observed market price using Brent's method.

    Args:
        market_price: Observed option price in the market.
        S:            Current underlying price.
        K:            Strike.
        T:            Time to expiry.
        r:            Risk-free rate.
        option_type:  'call' or 'put'.
        tol:          Solver tolerance.

    Returns:
        Implied volatility, or np.nan if no solution found.
    """
    intrinsic = max(S - K * np.exp(-r * T), 0) if option_type == "call" else max(K * np.exp(-r * T) - S, 0)
    if market_price < intrinsic:
        return np.nan

    def objective(sigma):
        return bs_price(S, K, T, r, sigma, option_type) - market_price

    try:
        iv = brentq(objective, 1e-4, 10.0, xtol=tol, full_output=False)
        return float(iv)
    except (ValueError, RuntimeError):
        return np.nan


def implied_vol_surface(
    S: float = 100.0,
    r: float = 0.03,
    strikes: list = None,
    maturities: list = None,
    sigma_atm: float = 0.20,
    skew: float = -0.02,
    term_slope: float = 0.01,
    output_path: str = None,
) -> pd.DataFrame:
    """
    Generate a synthetic implied vol surface and plot as heatmap.

    Synthetic IV surface: σ(K, T) = σ_ATM + skew × moneyness + term_slope × T

    Args:
        S:           Current price.
        r:           Risk-free rate.
        strikes:     List of strikes. Default: S × [0.8, 0.85, ..., 1.2].
        maturities:  List of maturities in years. Default: [0.25, 0.5, 1.0, 1.5, 2.0].
        sigma_atm:   ATM implied vol level.
        skew:        Slope of vol smile with moneyness (ln(K/S)).
        term_slope:  Term structure slope per year.
        output_path: If provided, save heatmap.

    Returns:
        DataFrame of implied vols: rows = maturities, columns = strikes.
    """
    if strikes is None:
        strikes = [round(S * m, 1) for m in np.arange(0.80, 1.21, 0.05)]
    if maturities is None:
        maturities = [0.25, 0.5, 1.0, 1.5, 2.0]

    iv_matrix = {}
    for T in maturities:
        row = {}
        for K in strikes:
            moneyness = np.log(K / S)
            iv_synth = sigma_atm + skew * moneyness + term_slope * T
            iv_synth = max(0.05, iv_synth)
            # Generate market price and recover IV for self-consistency
            mkt_price = bs_price(S, K, T, r, iv_synth, "call")
            iv_rec = implied_vol(mkt_price, S, K, T, r, "call")
            row[K] = round(iv_rec if not np.isnan(iv_rec) else iv_synth, 4)
        iv_matrix[T] = row

    iv_df = pd.DataFrame(iv_matrix).T
    iv_df.index.name = "Maturity"

    # Plot heatmap
    fig, axes = plt.subplots(1, 2, figsize=(15, 5))

    # Heatmap
    im = axes[0].imshow(iv_df.values, aspect="auto", cmap="RdYlGn_r",
                        vmin=iv_df.values.min() * 0.95, vmax=iv_df.values.max() * 1.05)
    axes[0].set_xticks(range(len(strikes)))
    axes[0].set_xticklabels([str(k) for k in strikes], rotation=45, fontsize=8)
    axes[0].set_yticks(range(len(maturities)))
    axes[0].set_yticklabels([f"{m}y" for m in maturities])
    axes[0].set_xlabel("Strike")
    axes[0].set_ylabel("Maturity")
    axes[0].set_title("Implied Volatility Surface (Heatmap)")
    plt.colorbar(im, ax=axes[0])

    # Vol smile (cross-section at each maturity)
    for T in maturities:
        axes[1].plot(strikes, iv_df.loc[T].values, "o-", label=f"T={T}y", linewidth=1.5)
    axes[1].set_xlabel("Strike")
    axes[1].set_ylabel("Implied Volatility")
    axes[1].set_title("Implied Volatility Smile by Maturity")
    axes[1].legend(fontsize=8)

    fig.tight_layout()
    if output_path:
        fig.savefig(output_path, dpi=150)
        print(f"Saved IV surface to {output_path}")
    plt.show()

    return iv_df

---

Advanced Usage

CRR Binomial Tree for American Options

def binomial_tree_price(
    S: float,
    K: float,
    T: float,
    r: float,
    sigma: float,
    n_steps: int = 200,
    option_type: str = "call",
    american: bool = True,
) -> float:
    """
    CRR binomial tree pricing for European and American options.

    Args:
        S:           Underlying price.
        K:           Strike.
        T:           Time to expiry.
        r:           Risk-free rate.
        sigma:       Volatility.
        n_steps:     Number of time steps (higher = more accurate).
        option_type: 'call' or 'put'.
        american:    If True, allow early exercise (American option).

    Returns:
        Option price.
    """
    dt = T / n_steps
    u = np.exp(sigma * np.sqrt(dt))
    d = 1.0 / u
    p = (np.exp(r * dt) - d) / (u - d)
    disc = np.exp(-r * dt)

    # Terminal stock prices
    j = np.arange(n_steps + 1)
    S_T = S * u ** (n_steps - 2 * j)

    # Terminal payoffs
    if option_type == "call":
        V = np.maximum(S_T - K, 0)
    else:
        V = np.maximum(K - S_T, 0)

    # Backward induction
    for i in range(n_steps - 1, -1, -1):
        V = disc * (p * V[:-1] + (1 - p) * V[1:])
        if american:
            j_i = np.arange(i + 1)
            S_node = S * u ** (i - 2 * j_i)
            if option_type == "call":
                intrinsic = np.maximum(S_node - K, 0)
            else:
                intrinsic = np.maximum(K - S_node, 0)
            V = np.maximum(V, intrinsic)

    return float(V[0])


def early_exercise_premium(
    S: float = 100.0,
    K: float = 100.0,
    T: float = 1.0,
    r: float = 0.05,
    sigma: float = 0.20,
) -> dict:
    """
    Compute early exercise premium: American price - European price.

    For calls on non-dividend paying stocks, this should be ~0 (no early exercise).
    For puts, the premium is positive.

    Args:
        S, K, T, r, sigma: Standard option parameters.

    Returns:
        Dictionary with european_call, american_call, european_put, american_put,
                         call_premium, put_premium.
    """
    eur_call = bs_price(S, K, T, r, sigma, "call")
    eur_put = bs_price(S, K, T, r, sigma, "put")
    amer_call = binomial_tree_price(S, K, T, r, sigma, n_steps=300, option_type="call", american=True)
    amer_put = binomial_tree_price(S, K, T, r, sigma, n_steps=300, option_type="put", american=True)

    call_prem = amer_call - eur_call
    put_prem = amer_put - eur_put

    print(f"\nEarly Exercise Premiums (S={S}, K={K}, T={T}, r={r:.1%}, σ={sigma:.1%}):")
    print(f"  Call: European={eur_call:.4f}  American={amer_call:.4f}  Premium={call_prem:.4f}")
    print(f"  Put:  European={eur_put:.4f}  American={amer_put:.4f}  Premium={put_prem:.4f}")

    return {"european_call": eur_call, "american_call": amer_call,
            "european_put": eur_put, "american_put": amer_put,
            "call_premium": call_prem, "put_premium": put_prem}

---

Troubleshooting

| Error / Issue | Cause | Resolution | |---|---|---| | brentq raises ValueError | Market price below intrinsic or above S | Check arbitrage: intrinsic ≤ mkt_price ≤ S | | IV = NaN for deep OTM options | Very low market price, solver loses bracket | Widen sigma bounds to (1e-5, 20.0) | | MC price converges slowly | Too few paths | Use n_paths=500_000 or apply variance reduction | | Binomial tree slow for large n_steps | O(n²) computation | Reduce to n_steps=200 for daily practice | | BS Theta > 0 for deep ITM put | Theta can be positive for deep ITM puts | This is mathematically correct; not an error | | Gamma → ∞ as T → 0 at-the-money | Near-expiry ATM option | Clip T at minimum 1e-6; expected behavior |

---

External Resources

• Black, F., Scholes, M. (1973). "The Pricing of Options and Corporate Liabilities." Journal of Political Economy, 81(3), 637–654.
• Merton, R.C. (1973). "Theory of Rational Option Pricing." Bell Journal of Economics, 4(1), 141–183.
• Cox, J., Ross, S., Rubinstein, M. (1979). "Option Pricing: A Simplified Approach." Journal of Financial Economics, 7(3), 229–263.
• Hull, J.C. (2022). Options, Futures, and Other Derivatives, 11th Edition. Pearson.
• Glasserman, P. (2004). Monte Carlo Methods in Financial Engineering. Springer.

---

Examples

Example 1 — BS Price + Greeks Parameter Table

greeks_df = print_greeks_table(S=100, K=105, T=0.5, r=0.04, sigma=0.25)

# Vary spot price and plot Delta profile
spots = np.linspace(60, 140, 100)
deltas_call = [bs_greeks(s, 100, 0.5, 0.04, 0.25, "call")["delta"] for s in spots]
deltas_put = [bs_greeks(s, 100, 0.5, 0.04, 0.25, "put")["delta"] for s in spots]

plt.figure(figsize=(9, 4))
plt.plot(spots, deltas_call, label="Call Delta", color="#2980B9")
plt.plot(spots, deltas_put, label="Put Delta", color="#E74C3C")
plt.axvline(100, color="gray", linestyle=":", linewidth=1, label="ATM")
plt.xlabel("Spot Price")
plt.ylabel("Delta")
plt.title("Delta Profile vs Spot")
plt.legend()
plt.tight_layout()
plt.savefig("delta_profile.png", dpi=150)
plt.show()

Example 2 — MC vs BS with Variance Reduction + IV Surface

# Monte Carlo with antithetic + control variate
mc_result = mc_option_price(
    S=100, K=100, T=1.0, r=0.05, sigma=0.20, option_type="call",
    n_paths=100_000, use_antithetic=True, use_control_variate=True,
)
print(f"BS price: {mc_result['bs_price']:.4f}  |  MC price: {mc_result['price']:.4f}")

# American vs European put early exercise premium
premium = early_exercise_premium(S=100, K=110, T=1.0, r=0.05, sigma=0.25)

# Implied vol surface
iv_df = implied_vol_surface(
    S=100, r=0.03, sigma_atm=0.20, skew=-0.03, term_slope=0.01,
    output_path="iv_surface.png"
)
print("\nIV Surface (maturity × strike):")
print(iv_df.round(4))

---

Changelog

| Version | Date | Change | |---|---|---| | 1.0.0 | 2026-03-18 | Initial release — BS pricing, Greeks, MC variance reduction, CRR tree, IV surface |