Abstract

This article explains how to compute and use Gamma and related second- and third-order Greeks in real-world option trading and risk management. Included are:

• How to calculate standard Gamma ($\Gamma$) and interpret it in practice
• Approximations for fast evaluation (Saddle Gamma)
• Normalizing Gamma (GammaP) for position sizing
• Using Gamma symmetry for skew analysis
• Evaluating sensitivity of Gamma to volatility (VommaGamma)
• Using Speed ($\partial \Gamma / \partial S$) for dynamic hedging
• Using Color ($\partial \Gamma / \partial T$) to understand time decay of risk

Throughout, we show numerical examples and discuss how traders and risk managers incorporate these metrics into daily workflows.

1 Introduction

Gamma is a critical measure for understanding how an option’s Delta changes with small moves in the underlying. In practice, high Gamma positions require frequent rebalancing to remain hedged. This guide provides step-by-step formulas, numerical examples, and notes on how to integrate Gamma-related Greeks into trading systems and risk reports.

2 Standard Gamma: Calculation and Example

2.1 Formula

Under Black-Scholes, the Gamma for a European call or put is:

$$\Gamma = \frac{e^{-bT}}{S \sigma \sqrt{2\pi T}} \exp\left(-\frac{d_1^2}{2}\right)$$

where

$$d_1 = \frac{\ln(S/X) + (b + \frac{1}{2}\sigma^2)T}{\sigma \sqrt{T}}, \quad d_2 = d_1 - \sigma \sqrt{T}$$

and

• S: current spot price of the underlying
• X: option strike
• T: time to maturity (in years)
• b: cost-of-carry (for stocks, typically r − q, where q is dividend yield)
• σ: implied volatility (annualized)

2.2 Numerical Example

Suppose:

• Underlying S = 100
• Strike X = 100
• Time to maturity T = 30/365 ≈ 0.0822 years (30-day option)
• Risk-free rate r = 5% annual; no dividends, so b = r = 0.05
• Implied volatility σ = 25%

First compute:

$$d_1 = \frac{\ln(100/100) + (0.05 + 0.5 \times 0.25^2) \times 0.0822}{0.25 \sqrt{0.0822}} \approx 0.092$$

Then:

$$\Gamma = \frac{e^{-0.05 \times 0.0822}}{100 \times 0.25 \times \sqrt{2\pi \times 0.0822}} \exp\left(-\frac{0.092^2}{2}\right)$$

Numerically:

$$e^{-0.05 \times 0.0822} \approx 0.9959$$ $$\sqrt{2\pi \times 0.0822} \approx 1.279$$ $$\exp\left(-\frac{0.092^2}{2}\right) \approx 0.9958$$

So

$$\Gamma \approx \frac{0.9959}{100 \times 0.25 \times 1.279} \times 0.9958 = \frac{0.9917}{31.98} \approx 0.0310$$

Gamma is 0.0310 per $1 move in S. A$1 move in the underlying causes Delta to shift by about 0.031.

2.3 Practical Notes

• Hedging Frequency: With Gamma = 0.031, a $1 move in the stock changes Delta by 3.1$1 million notional position, daily price swings of $1 require adjusting Delta by$31,000 of the underlying to stay hedged.

• Monitoring: Traders track Gamma not just at spot but across a grid of strikes and maturities (a “Gamma surface”) to see where risk is concentrated.

• Real-Time Alerts: Set thresholds for Gamma changes; alerts notify traders when Gamma exceeds risk limits.

3 Saddle Gamma: Fast Approximation

3.1 Motivation

For portfolios with thousands of option positions, computing Black-Scholes Gamma for each can be time-consuming. Saddlepoint approximations offer a faster way to estimate Gamma when extreme moves or non-lognormal features matter.

3.2 Saddle Gamma Formula (Lognormal Case)

In Black-Scholes, the cumulant-generating function is

$$K(q) = bq + \frac{1}{2}\sigma^2 q^2$$

The saddlepoint $q^*$ solves:

$$K'(q^*) = b + \sigma^2 q^* = \frac{\ln\left(\frac{X e^{-bT}}{S}\right)}{T}$$

so

$$q^* = \frac{1}{\sigma^2} \left( \frac{\ln(X/S) + bT}{T} - b \right)$$

Substitute into:

$$\Gamma_{saddle} \approx \frac{e^{-bT}}{S \sqrt{2\pi T (\sigma^2)}} \exp\left[T\left(K(q^*) - q^* K'(q^*)\right)\right]$$

Pre-built libraries (in Python or C++) handle these calculations once parameters are specified.

3.3 When to Use

• Short-Dated Options: As T → 0, Gamma spikes; saddlepoint avoids numerical instabilities.
• Heavy-Tailed Models: For fat-tail returns (e.g., jump-diffusion), saddlepoint captures tail behavior more accurately.
• Speed: Reduces CPU time in risk systems recalculating Greeks for large portfolios.

4 Percentage Gamma (GammaP) for Position Sizing

4.1 Definition and Interpretation

$$\Gamma_P = 100 \times \frac{\Gamma}{S}$$

measured as basis points of Delta per 1% move in the underlying. Traders use GammaP to compare risk across options on different underlyings.

4.2 Example and Use

Continuing the previous example with S = 100, Gamma = 0.0310

$$\Gamma_P = 100 \times \frac{0.0310}{100} = 0.031\%$$

per 1% move.

If a portfolio has $200,000 in option Delta notional at that strike, a 1$200,000 = $62. This helps budget hedging costs.

4.3 Risk Limits

Institutions set limits on aggregated GammaP across all options to cap the total Delta shift for a given market move.

5 Gamma Symmetry: Skew Analysis

5.1 Put-Call Symmetry

Gamma symmetry indicates call Gamma at one strike equals put Gamma at a mirrored strike:

$$\Gamma_{call}(K) = \Gamma_{put}\left(\frac{F^2}{K}\right)$$

where

$$F = S e^{(b-r)T}$$

Traders use this to spot skew: if put Gammas at low strikes exceed call Gammas at mirrored strikes, the market is skewed.

5.2 Application

• Vol Surface Construction: Enforce Gamma symmetry when interpolating to ensure no-arbitrage.
• Skew Monitoring: Compare implied volatilities at K and F²/K; deviations signal directional bias or demand imbalances.

6 VommaGamma: Sensitivity of Gamma to Volatility

6.1 Formula and Calculation

VommaGamma measures how Gamma changes as implied volatility shifts:

$$\frac{\partial \Gamma}{\partial \sigma} = \Gamma \frac{d_1 d_2 - 1}{\sigma}$$

6.2 Numerical Example

Using d1 ≈ 0.092, d2 = 0.0203, ${\Gamma}$ = 0.0310:

$$\frac{\partial \Gamma}{\partial \sigma} = 0.0310 \times 0.092 \times 0.0203 - \frac{1}{0.25} \approx -0.1238$$

A 1% absolute increase in volatility reduces Gamma by about 0.00124.

6.3 Practical Notes

• Vol-of-Vol Risk: Positions with large VommaGamma are sensitive to volatility shifts. Hedge by trading Vega options.
• Risk Reports: Include VommaGamma exposure to assess how volatility surface moves affect hedging.

7 Speed: How Gamma Changes with Spot

7.1 Formula and Interpretation

Speed is the third derivative ${\frac{\partial^3 C}{\partial S^3}}$:

$$\text{Speed} = -\frac{e^{-bT}}{S^2 \sigma \sqrt{2\pi T}} \exp\left(-\frac{d_1^2}{2}\right)\left(2 + \frac{d_1}{\sigma \sqrt{T}} - d_1^2\right)$$

A negative Speed means Gamma decreases as spot moves away from at-the-money.

7.2 Numerical Example

Using $d_1 = 0.092$, $T = 0.0822$, $\sigma = 0.25$, $S = 100$, $b = 0.05$:

$$2 + \frac{d_1}{\sigma \sqrt{T}} - d_1^2 \approx 2 + \frac{0.092}{0.2867} - 0.092^2 \approx 2.0175$$ $$\text{Speed} \approx -\frac{0.9959}{100^2 \times 0.25 \times 1.279} \times 0.9958 \times 2.0175 \approx -0.000626$$

A $0.10 move in spot changes Gamma by approximately −0.0000626.

7.3 Practical Implications

• Dynamic Hedging: Use Speed to estimate additional shares or futures to trade when spot moves a fraction, without recomputing full Gamma.
• Cost Estimates: Estimate transaction costs for small hedge adjustments.

8 Color: Gamma’s Time Decay

8.1 Formula

Color describes ${\frac{\partial \Gamma}{\partial T}}$

$$\text{Color} = \Gamma \left[ b - \frac{1 + d_1 d_2}{2T} \right]$$

Negative Color indicates Gamma decays as time passes.

8.2 Numerical Example

With $\Gamma = 0.0310$, $b = 0.05$, $d_1 = 0.092$, $d_2 = 0.0203$, $T = 0.0822$:

$$1 + d_1 d_2 \approx 1.0019$$ $$\frac{1 + d_1 d_2}{2T} \approx 6.093$$ $$b - \frac{1 + d_1 d_2}{2T} \approx -6.043$$ $$\text{Color} \approx 0.0310 \times (-6.043) \approx -0.1873$$

For a 1-day (0.00274 years) decay, Gamma decreases by 0.1873 × 0.00274 ≈ 0.00051.

8.3 Use Cases

• Hedging Horizon: When Color is large, Gamma erosion is rapid; hedge more often near expiry.
• Margin Forecasting: Gamma affects margin; use Color to project margin requirements.

9 Implementing in a Risk System

9.1 Workflow

1. Market data feed: Ingest live S, implied vols, rates, dividends.
2. Batch Greek computation: Compute Γ, GammaP, VommaGamma, Speed, Color daily or on demand.
3. Risk dashboard: Show aggregated exposures: total GammaP by underlying, VommaGamma by volatility bucket, largest Speed values.
4. Alerts: Notify when Gamma or VommaGamma exceed thresholds or when Color signals rapid Gamma decay.
5. Hedge execution: Use Speed and Color to guide size and timing of Delta hedges.

9.2 Sample Python Pseudocode

# Given S, K, T, r, q, sigma
import math
def compute_greeks(S, K, T, r, q, sigma):
    b = r - q
    d1 = (math.log(S/K) + (b + 0.5*sigma**2)*T) / (sigma*math.sqrt(T))
    d2 = d1 - sigma*math.sqrt(T)
    gamma = math.exp(-b*T)/(S*sigma*math.sqrt(2*math.pi*T)) * math.exp(-0.5*d1**2)
    vomma_gamma = gamma * (d1*d2 - 1)/sigma
    speed = -math.exp(-b*T)/(S*S*sigma*math.sqrt(2*math.pi*T)) * math.exp(-0.5*d1**2) * (2 + d1/(sigma*math.sqrt(T)) - d1**2)
    color = gamma*(b - (1 + d1*d2)/(2*T))
    gamma_p = 100 * gamma/S
    return {"Gamma": gamma, "GammaP": gamma_p,
            "VommaGamma": vomma_gamma, "Speed": speed, "Color": color}

10 Summary and Best Practices

• Compute and monitor Gamma and GammaP daily for all liquid strikes; aggregate by maturity buckets.

• Use VommaGamma to understand how Gamma profiles shift with volatility moves; hedge Vega accordingly.

• Use Speed and Color to forecast hedging needs when spot moves or time passes.

• Incorporate saddlepoint approximations in large-portfolio contexts to save CPU time.

• Validate Greeks by backtesting small price moves to ensure model accuracy in production.