Part V · Options · Chapter 5C

The Greeks

Delta, gamma, theta, vega and rho

The Greeks are the language of risk in options trading. Each Greek measures how an option's price changes when one specific input moves, holding the others constant, so the Greeks are sensitivities rather than predictions. A skilled trader uses them like the gauges on a car's dashboard, not to predict the road but to understand how the vehicle responds to its inputs.

The five primary Greeks at a glance:

1 · A Greek is a rate of change

Each Greek is a rate: how many dollars the option's value $V$ moves per unit of one input. For a straight line that rate is the slope, the rise over the run, and it is the same everywhere. An option's value curves, so its slope differs from point to point. The slope at a single point comes from steps that shrink toward zero:

Read $\lim_{h \to 0}$ as "the value the ratio settles on as the step $h$ shrinks toward zero". The result is the derivative of $V$ with respect to $S$. The curly $\partial$ marks a partial derivative. The value also depends on time, volatility and the interest rate. A partial derivative moves $S$ alone while it holds all of those fixed.

The $1 step overstates the slope because the value curve bends upward, so its slope grows as spot rises. The rate at which a slope itself changes is a second derivative, written $\partial^2 V/\partial S^2$. For the option's value, the second derivative with respect to spot is gamma.

Together, the first two derivatives approximate the change in value for any modest move $\Delta S$ in spot. This is the Taylor-series idea that Chapter 1A used to approximate $\ln(1 + x)$ by $x - x^2/2$:

The symbol $\Delta$ does two jobs here, as it does in the figures. On its own it is the Greek delta, and in front of a letter it means "the change in".

2 · The Black-Scholes formula

The Greeks are derivatives of the Black-Scholes price, which Chapter 7A derives and Chapter 6B unpacks. For now, treat the formula as a black box whose inputs you know:

• Spot and strike: $S$ is the price of the underlying and $K$ the strike.
• Time: $T$ is the time to expiry in years, so that 30 days is $30/365 = 0.0822$.
• Volatility: $\sigma$ is the implied volatility as a decimal, 0.25 for 25% (Chapter 5E).
• Rates: $r$ is the risk-free interest rate and $q$ the dividend yield, both as decimals.

Two numbers carry most of the work:

Two functions of the standard normal curve turn them into prices. The first, $N(x)$, is the standard normal distribution function, also called the cumulative distribution function (CDF): the area under the curve to the left of $x$. The second, $\varphi(x)$, is the curve's height at $x$, equal to $e^{-x^2/2}/\sqrt{2\pi}$. At zero they take the values $N(0) = 0.5$ and $\varphi(0) = 0.399$. With them, the Black-Scholes values of a call and a put are

3 · Delta: direction

Delta is the change in an option's value for a $1 rise in the underlying, with everything else held fixed. Written as a derivative, it is the slope of the value curve against spot:

If the underlying moves up by a dollar, the option moves by approximately delta dollars. A call's delta lies between 0 and 1, and a put's between −1 and 0. Deep in-the-money options have delta close to ±1 and behave almost like the underlying itself. Deep out-of-the-money (OTM) options have delta close to 0 and barely react. At-the-money (ATM) options have delta near ±0.5. For a European option, Black-Scholes gives the two deltas as

For the reference option, the call's delta is $N(0.0817) = 0.533$ and the put's is $-(1 - 0.533) = -0.467$. A $1 rise in spot adds about $0.53 a share to the call, which is $53 a contract, and takes about $0.47 a share from the put.

Delta as a hedge ratio

Delta is the hedge ratio: the number of shares per contract, divided by 100, that neutralizes the option's first-order exposure to spot. The operative reading is share-equivalent exposure, so a 0.50-delta call contract behaves like 50 shares of the underlying, delta times the multiplier of 100. A dealer who has sold that call hedges by buying 50 share-equivalents of the underlying, as Chapter 6C shows. This share-equivalent reading is what makes delta hedging mechanically possible.

Delta is not a probability

$N(d_1)$ has a geometric reading as a probability, but an unusual one. It is the chance of finishing in the money under a share-numeraire measure, a way of weighting the outcomes that counts each by the price of the underlying in it. It is not the standard risk-neutral probability of finishing in the money: the probability that option prices imply when every asset is assumed to grow at the risk-free rate. For a call, that probability is $N(d_2) = N(d_1 - \sigma\sqrt{T})$.

Because $d_2$ sits exactly $\sigma\sqrt{T}$ below $d_1$, a call's $N(d_2)$ is smaller than $N(d_1)$ by approximately $\varphi(d_1)\,\sigma\sqrt{T}$, which is about $0.4\,\sigma\sqrt{T}$ near the money. For the reference call the gap is $0.5326 - 0.5040 = 0.0286$, close to $0.3976 \times 0.0717 = 0.0285$. For a put the order reverses: its chance of finishing in the money, $1 - N(d_2)$, is larger than its delta suggests.

The retail convention conflates the two, and traders say that "30-delta puts have a 30% chance of being assigned". That is informally close to right for short-dated, near-the-money options, but it is technically incorrect. The gap is a percentage point or two there, and it grows with volatility and time:

The table assumes spot at 6,000, a 4% rate and a 1.5% dividend yield. Even these chances are risk-neutral ones, a pricing device rather than a forecast of what the index does (Chapter 7A). The reliable interpretation of delta is mechanical, as a hedge ratio, not probabilistic.

When traders say "30-delta", they mean a hedge ratio of 0.30, whatever it is informally called.

The probability map draws the standard normal curve with $d_1$ and $d_2$ marked, $\sigma\sqrt{T}$ apart. Its shaded area is the risk-neutral probability of finishing in the money: the area left of $d_2$ for a call, which is $N(d_2)$, and right of it for a put. It opens on an at-the-money SPX call with spot and strike at 6,000, 30 days to expiry, 20% IV, a 4% rate and a 1.5% dividend yield. The table reads $N(d_1) = 0.5257$, a delta of 0.5251 and $N(d_2) = 0.5029$, so delta exceeds the probability by 0.022. Drag the days to expiry toward 365, or raise the IV, and then switch to the put.

spot S

6000

strike K

6000

DTE

30d

IV σ

20%

risk-free r

4.0%

dividend yield q

1.50%

option call put

Quantity	Value	Meaning
N(d₁)	0.5257	share-numeraire probability
Call Δ = e^(−qT)·N(d₁)	0.5251	hedge ratio (shares per contract ÷ 100)
N(d₂) = P(ITM)	0.5029	risk-neutral ITM probability

|Δ| − P(ITM) = +0.0222

absolute gap = 0.0222

|Δ| exceeds P(ITM) by a meaningful margin. Reading delta as the probability of finishing in the money loses accuracy as σ√T, the distance from d₁ down to d₂, grows with vol and time.

As $\sigma\sqrt{T}$ grows, $d_1$ and $d_2$ spread apart and the gap opens. At 365 days the call's delta is 0.580 against a 0.510 chance of finishing in the money. For the put, the readout reports that $|\Delta|$ falls short of the probability instead of exceeding it. Throughout, the call's delta sits slightly below $N(d_1)$ because of the dividend factor $e^{-qT}$.

4 · Gamma: curvature

Gamma is the change in delta for a $1 rise in the underlying. It is the second derivative of the option's value with respect to spot, or equivalently the first derivative of delta:

Positive gamma is favorable for long option holders, because directional exposure accelerates in the right direction as the underlying moves. If you are long a call and the underlying rises, your delta increases, so you take part in further upside faster. Negative gamma is unfavorable: short option holders see their delta move against them as the underlying moves. Black-Scholes gives gamma as a ratio with the bell curve's height on top:

Here $\varphi$ is the height of the bell curve, not the area under it. For the reference option, $\Gamma = 0.3976/(100 \times 0.25 \times 0.28669) = 0.0555$, so a $1 rise in spot lifts the call's delta from 0.533 to about 0.588.

Three features of this formula matter:

1. Gamma is the same for a call and a put with the same strike, which put-call parity (Chapter 5F) ensures. By (5C.8), the call's delta minus the put's is $e^{-qT}$, which does not depend on spot, so the two deltas change with spot at the same rate.
2. Gamma is maximized at the money, because the density $\varphi(d_1)$ peaks when $d_1 = 0$, which happens approximately when spot equals the strike.
3. Gamma is proportional to $1/\sqrt{T}$ at the money, where $\varphi(d_1)$ stays near 0.4. As time to expiry shrinks toward zero, at-the-money gamma diverges to infinity. Cutting the time to a quarter doubles gamma, while going from 30 days to 1 day multiplies it by $\sqrt{30} \approx 5.5$.

The $1/\sqrt{T}$ divergence is not a mathematical curiosity. It is the dominant feature of modern intraday microstructure, because 0DTE options live exactly in this regime (Chapters 6D and 9B).

5 · Theta: time

Theta is the change in an option's value as time passes, with everything else held fixed. Write $t$ for calendar time, which moves forward as the time to expiry $T$ shrinks by the same amount, so that

The formula gives theta per year, and trading screens show it per calendar day, divided by 365. Theta then measures the daily P&L from one calendar day passing, all else constant. Long options have negative theta: every day, all else equal, the option loses value as its remaining time shrinks. Short options have positive theta: every day, the writer accrues a positive bookkeeping gain as the time value decays.

The Black-Scholes formula for theta is messier than the others, with different forms for calls and puts and an extra term when there are dividends. With no dividend, the two are

For the reference call, the first term is $-(100 \times 0.3976 \times 0.25)/(2 \times 0.28669) = -17.34$ a year, and the interest term is $-0.04 \times 100 \times 0.99672 \times 0.5040 = -2.01$ a year. Their sum, $-19.35$ a year, comes to $-0.053$ a day once divided by 365. The call loses about 5.3 cents a share, $5.30 a contract, for each day that passes with nothing else changing. Repricing it with 29 days left instead of 30 gives a fall of 5.34 cents, which agrees.

The structure is the same for calls and puts. Theta is negative for long positions, positive for short ones and largest in size at the money. It grows like $1/\sqrt{T}$ near expiry, the mirror image of gamma's blow-up. The mirror is exact: substitute (5C.10) into $\tfrac12\sigma^2 S^2\,\Gamma$, and one $S$ and one $\sigma$ cancel to leave $S\varphi(d_1)\sigma/(2\sqrt{T})$, the first term of theta. An exception to these signs is a deep in-the-money European put, whose interest term can outweigh the first, which is why such a put can trade below its intrinsic value (Chapter 5B). So for a call with no dividend,

For the reference call, $\tfrac12 \times 0.0625 \times 10{,}000 \times 0.0555$ comes to 17.34, the first term exactly. Most of theta is the rent on gamma, with interest making up the rest. Chapter 7A derives the relation in general.

Chapter 7A derives the formal relationship between theta and gamma, which (5C.13) previews. In the Black-Scholes-Merton (BSM) framework, the long-option holder pays theta in exchange for gamma. The short-option holder receives theta in exchange for accepting negative gamma. This exact trade-off, not an approximate one, is the central mathematical engine behind every dealer-hedging story from Part VI on.

6 · Vega: volatility

Vega is the change in an option's value for a rise in implied volatility, with everything else held fixed:

This formula gives the change per unit of $\sigma$, where $\sigma = 1.0$ means 100% annualized volatility. Most trading platforms instead display vega per vol point, a rise of one percentage point in implied volatility, which is the formula's value divided by 100. A platform showing "Vega = $4.50" means the position gains $4.50 if IV rises from 18.0% to 19.0%, not if $\sigma$ doubles from 1.0 to 2.0. This is a frequent source of beginner error when computing vega by hand.

Vega measures an option's exposure to changes in the market's expected volatility. Long options have positive vega, gaining value when IV rises. Short options have negative vega, losing value when IV rises. This is the direct mechanical link between options trading and the broader volatility market. The Black-Scholes formula, the same for a call and a put, is

For the reference option, $\nu = 100 \times 0.3976 \times 0.28669 = 11.40$ per unit of $\sigma$, or 0.114 per vol point. A rise in IV from 25% to 26% adds about 11.4 cents a share, or $11.40 a contract.

Vega is positive for both long calls and long puts. It is largest for long-dated at-the-money options, because the $\sqrt{T}$ factor grows with time, unlike gamma's $1/\sqrt{T}$. It is smaller for short-dated options, which leave less time for volatility to affect the price.

Operational consequence: a long at-the-money 30-day SPX call has substantial positive vega. At SPX 6,000 and 20% IV, (5C.15) gives about 6.8 index points per vol point, which is about $680 a contract. A long at-the-money 0DTE SPX call has nearly zero vega, because moves in IV at expiry do not change the option's payoff. This is why vol-targeting and volatility surface strategies concentrate in longer-dated options: that is where the vega lives.

7 · Rho: interest rates

Rho is the change in an option's value for a rise in the risk-free rate:

For most short-term equity options rho is negligible, because the value barely changes when rates move by typical daily amounts. Rho becomes relevant for long-dated options (LEAPS, with expiries of 1 to 2 years) and during rate-regime shifts. The 2022-2023 Fed hiking cycle moved option valuations measurably through the rho channel for the first time in over a decade. Black-Scholes gives rho for a call and a put as

For the reference call, rho is $100 \times 0.08219 \times 0.99672 \times 0.5040 = 4.13$ per unit of rate, or 0.041 for a one-point rise. A full percentage point on rates adds about 4 cents a share, less than a single day of theta.

Calls have positive rho because higher rates increase the forward price of the underlying, $S e^{(r - q)T}$, the cost of buying it today and carrying it to expiry (Chapter 4B). Puts have negative rho for the same reason. For near-term SPX options rho is a third-order concern, relevant mostly to how the entire volatility surface reprices during rate-regime changes.

8 · The five Greeks together

The simulator below computes all five Greeks of one option in trading-screen units: theta per calendar day, vega per vol point and rho per 1% of rate. It fixes the rate at 4% with no dividend. Its controls set spot and strike (each from 70 to 130), days to expiry (1 to 180), implied volatility (5% to 80%) and call or put. It opens on the reference call, so its table reads delta 0.533, gamma 0.0555, theta −0.053, vega 0.114 and rho 0.041.

The chart below the table tests (5C.2). For a what-if shift of +$5, delta alone predicts $0.5326 \times 5 = +2.663$, and the gamma term adds $\tfrac12 \times 0.0555 \times 25 = 0.693$, for a total of $+3.356$. The full reprice at spot 105 is $+3.308$, so the delta-gamma estimate misses by $-0.048$, about 1.5% of the change. Drag the shift from −$10 to +$10 and watch the three curves part.

spot S

$100.0

strike K

$100

days to expiry

30d

implied vol σ

25%

option call put

premium = $3.02 per share

Greek	Value	Measures
Δ delta	0.533	∂V/∂S, share-equivalent exposure (× 100 per contract)
Γ gamma	0.0555	∂Δ/∂S, how fast delta changes
Θ theta	−0.053	∂V/∂t, daily P&L from time passing
ν vega	0.114	∂V/∂σ, P&L per 1 vol point
ρ rho	0.041	∂V/∂r, P&L per 1% rate change

what-if spot shift

+$0.0

Δ·ΔS = +$0.000

Δ·ΔS + ½Γ·ΔS² = +$0.000

full reprice = +$0.000

higher-order error = +$0.0000

For small moves delta captures most of the change, and the gamma term captures most of the rest. What remains, with volatility and time held fixed, comes from speed and the higher spot derivatives of Chapter 5D. At +$10 it reaches $-0.50$. It grows with the size of the move, and also as you lower the volatility or the days to expiry, where the value curve bends more sharply. Set 3 days to expiry at the money and watch gamma and theta jump while vega shrinks.

The Greek surface shows how one Greek varies across both spot and time. It draws the Greek of an option struck at $K = 100$, with the rate at 4%, as a wireframe over spot from 70 to 130 and 1 to 90 days to expiry. Theta is per calendar day and vega per vol point. It opens on gamma at 22% IV, where the readout gives 0.0630 at the money with 30 days to go. Turn it with the view angle slider and tilt it with the tilt slider, then switch between the Greeks and between a call and a put. The vanna and charm buttons belong to Chapter 5D.

Greek Δ delta Γ gamma ν vega Θ theta vanna charm

implied vol σ

22%

option call put

view angle

45°

tilt

28°

active Greek = Γ gamma

at ATM, 30 DTE = 0.0630

type = call

On the gamma surface, the hump over the strike stays low and broad at 90 days and collapses into a sharp peak at the money as expiry nears: the $1/\sqrt{T}$ divergence. Theta is its mirror, a trough that deepens at the money near expiry. Vega does the opposite of both, rising toward the long-dated end, while delta turns from a gentle S-curve into a near-step at the strike.

9 · The gamma-theta trade-off

Within the Black-Scholes-Merton framework, gamma and theta are inseparably linked, and you cannot have one without the other. If you are long an option, you have positive gamma, which is favorable, and you pay for it through negative theta. If you are short, you have negative gamma, which is unfavorable, and you are compensated for it through positive theta.

A delta-hedged option is one held against a position in the underlying of the opposite delta, so that small moves in spot cancel to first order (Chapter 6C). Over a short interval $dt$, with $\sigma_I$ the implied volatility at entry, the P&L of a delta-hedged long option is

Read this as: over each interval, a delta-hedged long option earns $\tfrac12\Gamma S^2$ times the difference between the realized squared return and the implied variance. This is the single most important equation in the whole framework, and Chapter 7A derives it properly.

Four consequences follow from (5C.18):

• Long gamma earns when realized volatility exceeds implied. If the underlying makes bigger moves than the IV at entry predicted, the delta-hedged long option scalps gamma. It buys low and sells high in the underlying hedge, and the cumulative scalping gain exceeds the theta paid.
• Short gamma earns when realized volatility falls below implied. If the underlying makes smaller moves than the IV predicted, the short-option position keeps the theta and avoids losing money on the gamma hedges.
• Theta is the "rent" paid for owning gamma. It is the price you pay for having a position that benefits from movement, and selling gamma collects that rent.
• The variance risk premium is the operative bet. Every delta-hedged option position is fundamentally a bet on realized against implied volatility, nothing more and nothing less. The directional exposure is hedged away, and what remains is the spread between realized and implied.

This decomposition is what justifies the whole framework of Part VI onward. The dealer who has sold an option to a customer and delta-hedged it is making exactly this bet: that realized volatility comes in below the IV at which they sold. If realized exceeds implied, they lose money to the customer's gamma, and if it is below implied, they keep the spread. The variance risk premium is, in aggregate, the empirical observation that this trade has paid dealers historically: implied volatility has averaged higher than realized (Chapter 10C).

10 · Check yourself