Gamma Scalping & P&L Attribution

You know all five Greeks, and you know how to short stock against a long option to zero out delta. So here’s the question that pays the rent: once you’re delta-neutral, where does the money actually come from? You’ve cancelled your exposure to the stock’s direction — up or down no longer moves your book to first order. And yet long-gamma desks make money. The trick is that “delta-neutral” doesn’t mean “P&L-neutral.” It means your profit now flows through curvature and time, not direction. This lesson is the machine that turns gamma into cash — and the accountant’s ledger that proves it.

Before you read — take a guess

You're long an at-the-money call and short exactly enough stock to be delta-neutral. The stock then makes a big move — say a sharp drop. To first order (delta), your book is hedged. So where can your profit come from?

From rho — interest rates dominate a delta-neutral bookFrom gamma — the move is big enough that your delta changed, and re-hedging that curvature banks a profitFrom delta — being delta-neutral actually maximizes directional profitNowhere — a delta-neutral position can never make or lose money

Gamma scalping mechanics: the hedge that buys low and sells high

Analogy. Picture a vending machine that, instead of charging you, pays you every time the floor shakes. The harder the building rattles, the more coins drop out. A long-gamma, delta-hedged book is exactly that machine: it converts raw movement — turbulence, chop, wiggle — into cash. It doesn’t care which way the floor tilts, only that it keeps tilting.

Definition. You hold a long option (so you’re long gamma, $\Gamma > 0$ ) and you keep the book delta-neutral by trading the underlying. The mechanism is purely automatic. Recall that gamma is $\Gamma = \partial\Delta/\partial S$ : when the stock moves, your delta moves with it. To stay neutral you must trade back to flat, and the direction of that trade is forced by the sign of gamma:

Stock falls. A long position’s delta drops (gamma is positive, so $\Delta$ moves the same way as $S$ ). You’re now net short delta. To re-neutralize you buy the underlying — and you’re buying it after it fell. You bought low.
Stock rises. Your delta climbs. You’re now net long delta. To re-neutralize you sell the underlying — after it rose. You sold high.

Buy low, sell high — and you never made a decision. The hedge made it for you. Every round-trip (down then up, or up then down) banks a small scalp. This is the whole game: you are being paid, mechanically, to fade every move.

Gamma scalping vs. theta bleedNet P&L: 1.26

Gamma-scalp gainsTheta costNet P&L

Realized volatility30%

Implied volatility (paid)20%

Realized volatility: 30%
Implied volatility (paid): 20%
Gamma-scalp gains: 2.95
Theta cost: -1.69
Net P&L: 1.26

Each day the delta hedge forces you to buy the dip and sell the rip, banking a scalp of about ½·gamma·(price move)². Theta quietly bleeds value the whole time. The net only finishes positive when realized volatility (how much the price actually wiggles) beats the implied volatility you paid. Slide realized above implied and watch the net line lift above zero — then drag it below and watch the same machine bleed.

The writer runs the same machine in reverse

Everything here flips for the option seller. A short option is short gamma: when the stock falls your delta climbs, forcing you to sell low; when it rises your delta drops, forcing you to buy high. The writer’s hedge mechanically buys high and sells low — the worst reflex in trading. That’s why selling options “for the theta” is picking up nickels in front of a steamroller: your hedging dance is structurally a loss every time the stock swings.

When it matters

Gamma scalping is the bread and butter of a long-volatility book — straddles, strangles, anything with positive curvature held delta-hedged. It matters most when gamma is high: near-the-money strikes and short time to expiry, where delta swings hardest per dollar of stock. A deep ITM or OTM option has near-zero gamma, so there’s nothing to scalp — the machine only pays when you’re sitting on the steep part of the curve.

The scalp P&L: ½Γ(ΔS)² per swing

Analogy. Think of each price move as a coin of a fixed size that the machine drops. The catch: the coin’s value grows with the square of how far the stock moved. A move twice as big doesn’t pay double — it pays four times as much. Curvature rewards size disproportionately.

Definition. Each time you re-hedge after a move of $\Delta S$ in the underlying, the long-gamma book banks approximately

\text{scalp} \approx \tfrac{1}{2}\,\Gamma\,(\Delta S)^2 .

This is just the second-order (curvature) term of the Taylor expansion of the option’s value: the delta term is hedged away by your stock, leaving the $\tfrac12\Gamma(\Delta S)^2$ piece as pure profit. It’s always positive for long gamma — squaring kills the sign of the move, so down-moves pay exactly like up-moves of the same size. That’s the deep reason a long-gamma book is direction-blind: only the magnitude of the wiggle counts.

Worked example. Suppose your position gamma is $\Gamma = 0.10$ (your delta shifts by 0.10 per $1 move, per share). The stock swings $2 and you re-hedge:

\tfrac{1}{2}\,\Gamma\,(\Delta S)^2 = \tfrac{1}{2}\times 0.10 \times (2)^2 = \tfrac{1}{2}\times 0.10 \times 4 = 0.20 \text{ dollars per share}.

On a 100-share contract that’s $20 banked on that single swing. If the day delivers, say, eight such $2 swings, you’ve scalped roughly $1.60 per share (that’s $8 \times 0.20$ ), or $160 on the contract — purely from re-hedging the curvature. Now contrast: had the stock made one $8 move instead of four $2 moves, the four-times-squaring rule means the single big lurch alone would scalp $\tfrac12\times0.10\times64 = 3.20$ dollars per share. Big, jumpy days are a long-gamma book’s favorite weather.

You are literally long realized variance

Notice the scalp is proportional to $(\Delta S)^2$ — the squared return, i.e. realized variance. Sum $\tfrac12\Gamma(\Delta S)^2$ over a day and you’ve measured how much the stock actually moved around, in variance units. A delta-hedged long-gamma book is, mathematically, a position that gets paid the realized variance of the underlying. Hold that thought — it’s the punchline of the whole lesson.

Your position gamma is 0.05 per share. On one re-hedge the stock moves $3. Roughly how much does that single swing scalp, per share?

About $0.225 — ½ × 0.05 × 3² = ½ × 0.05 × 9About $0.15 — 0.05 × 3About $0.075 — ½ × 0.05 × 3About $0.45 — 0.05 × 3²

But you paid theta: the cost of admission

Analogy. The vending machine isn’t free. It runs on a meter that ticks down every single day whether or not the floor ever shakes. That meter is theta, and it’s the rent you pay to keep the gamma machine plugged in.

Definition. From the Greeks lesson: long gamma and negative theta are two ends of the same stick. You cannot own positive curvature without paying time decay. Every day the option loses value at a rate of $\Theta$ (negative for a long position), regardless of whether the stock moved. So the long-gamma trader’s daily ledger has exactly two lines:

\text{daily P\&L} \approx \underbrace{\sum \tfrac{1}{2}\Gamma(\Delta S)^2}_{\text{gamma scalps (gain)}} \;+\; \underbrace{\Theta\,dt}_{\text{theta bleed (cost)}} .

The scalps are your income; theta is your fixed cost. A quiet day still costs you the theta — you paid for a machine that nobody shook. A wild day scalps more than enough to cover it. The art of running the book is making sure the building shakes enough to beat the meter.

Worked example. Say your one-day theta is $0.18 per share (you bleed 18 cents a day). From the previous example, eight $2 swings scalped $1.60 per share. Net day:

1.60 - 0.18 = 1.42 \text{ dollars per share profit.}

But on a calm day — one tiny $0.30 swing — the scalp is only $\tfrac12\times0.10\times0.09 \approx 0.0045$ dollars per share, nowhere near the $0.18 theta. Net: a loss of about $0.175 per share. Same position, same Greeks — the only difference is how much the stock actually moved. That is the entire long-gamma business in one sentence.

The master equation: realized vs implied volatility

Analogy. Implied volatility is the rent the landlord quoted you when you signed the lease — the market’s forecast of how much the stock will wiggle, baked into the option’s price. Realized volatility is how much the building actually shook while you lived there. You profit if the place rattled more than the landlord priced in, and you lose if you paid up for turbulence that never came.

Definition. Combine the gamma scalps and the theta bleed and a small miracle happens: for a delta-hedged long option, the expected daily P&L collapses to a single, beautiful expression,

\text{daily P\&L} \approx \tfrac{1}{2}\,\Gamma\,S^2\big(\sigma_{real}^2 - \sigma_{imp}^2\big)\,dt .

Read it slowly. The $\tfrac12\Gamma S^2\,dt$ part is always positive (a scale factor — “dollar gamma” times the time step). The sign of your whole day lives entirely in the bracket: $\sigma_{real}^2 - \sigma_{imp}^2$ . The intuition is exactly the two-line ledger: your scalp gains scale with realized variance $\sigma_{real}^2$ (because each scalp is $\tfrac12\Gamma(\Delta S)^2$ and the moves $\Delta S$ are sized by realized vol), while your theta cost is priced off the implied variance $\sigma_{imp}^2$ you paid when you bought the option (Black–Scholes set theta using the implied vol embedded in the premium). Subtract the rent from the income and only the variance gap survives.

So the entire long-gamma trade reduces to one bet:

Gamma scalping pays if and only if realized volatility exceeds the implied volatility you paid.

You are not really trading the stock. You are trading realized vol against implied vol. Buying a delta-hedged option is “I think the stock will move more than the price implies.” Selling one is “I think it’ll move less.”

You're really trading realized vs implied volatility

Strip away the stock, the strikes, the hedging — at the bottom it’s a single comparison: did the world turn out more volatile (realized) than the price said it would (implied)? Long delta-hedged gamma is a long bet on realized > implied. Short it is the reverse. Every gamma scalper is, knowingly or not, a volatility arbitrageur betting on which number wins.

Worked example — both directions.

Show the realized-beats-implied vs implied-beats-realized math

Take dollar gamma such that $\tfrac12\Gamma S^2 = 5{,}000$ dollars for a one-percentage-point change in variance over the period, and a daily step.

Case A — realized 30% beats implied 20% (you win). You paid for 20% vol but the stock delivered 30%. In variance terms, $0.30^2 - 0.20^2 = 0.09 - 0.04 = 0.05$ . The bracket is positive, so the scalps you bank outrun the theta rent. Your delta-hedge fades a string of big moves, harvesting more $\tfrac12\Gamma(\Delta S)^2$ than the decay costs. Net: profit — you bought cheap movement and the world delivered expensive movement.

Case B — realized 10% trails implied 20% (you lose). Same option, but the stock went quiet: it only delivered 10% against the 20% you paid for. Now $0.10^2 - 0.20^2 = 0.01 - 0.04 = -0.03$ . The bracket is negative: the few tiny scalps you collect can’t cover the theta meter, which was set by the richer 20% implied. Net: loss — you rented an earthquake-proofing machine for a city that never trembled.

The Greeks were identical in both cases. The hedging discipline was identical. The only thing that changed was how much the stock actually moved versus the move you paid for. That is the master equation made tangible.

When it matters

This equation is the foundation of volatility arbitrage and how options market-makers think about every position. It’s also why “implied vol is cheap/rich” is the first thing a vol trader checks: it’s literally the strike price of the realized-vs-implied bet. The approximation assumes you re-hedge frequently and that gamma/spot don’t move too much within the period — over a single huge gap, higher-order terms (and the residual we meet next) start to matter.

You buy a delta-hedged straddle at an implied volatility of 25%. Over its life, the stock realizes only 18% volatility. Ignoring frictions, what happened to your gamma-scalping P&L?

A profit — any realized volatility above zero guarantees a long-gamma profitA loss — realized (18%) fell short of the implied (25%) you paid, so theta rent outran your scalpsBreakeven — delta-hedging always nets to zeroA profit, because lower realized volatility means lower theta

P&L attribution: decomposing the day into Greeks

Analogy. At day’s end your book moved by some dollar amount. P&L attribution is the itemized receipt for that number: not just “I made $4,200 today” but “$3,000 of it was the stock move, $900 was gamma harvesting the chop, −$300 was theta, $700 was vol popping, $50 was rates, and $-150 is rounding error the model couldn’t explain.” A trader who can’t itemize the receipt doesn’t actually know why they made money — and can’t tell skill from luck.

Definition. Take the second-order Taylor expansion of the option’s value in all its inputs. The change in value over a period decomposes, Greek by Greek, into named contributions:

\Delta V \approx \underbrace{\Delta\,dS}_{\text{delta}} + \underbrace{\tfrac{1}{2}\Gamma\,dS^2}_{\text{gamma}} + \underbrace{\Theta\,dt}_{\text{theta}} + \underbrace{\mathcal{V}\,d\sigma}_{\text{vega}} + \underbrace{\rho\,dr}_{\text{rho}} + \underbrace{\varepsilon}_{\text{residual}} .

Each term is a measured number: take the Greek you held at the start of the period and multiply by the market move that actually happened. The residual $\varepsilon$ is whatever the full, exact reprice does that the first-and-second-order terms missed — higher-order curvature, cross-terms, and the fact that the Greeks themselves drift during the period.

Where the day's P&L came fromGreek total: $3,099

10 × Call · K 100 · S 100

Stock move dS ($)5

Vol change dσ (pts)0

Days passed1

Rate change dr (pts)0

Delta P&L$2,844
Gamma P&L$278
Theta P&L−$23
Vega P&L$0
Rho P&L$0
Residual (reprice − Greeks)−$11
Greek total$3,099

A day's option P&L breaks into its Greeks: delta P&L from the stock's move, gamma P&L from the move squared (always helping a long position), theta bleeding time value, vega from the vol change, and rho from rates. Sum them and you nearly recover the full reprice — the leftover residual is the higher-order error the linear-plus-gamma sketch misses, and it grows fast on big moves. Drag the sliders and watch the bars rearrange the story of the day.

Worked example. You’re long 10 call contracts (1,000 shares). At the open your Greeks are: delta $\Delta = 0.55$ , gamma $\Gamma = 0.04$ (per share), theta $\Theta = -0.05$ per share per day, vega $\mathcal V = 0.20$ per share per vol point, rho $\rho = 0.12$ per share per rate point. Over the day: the stock rose $3 ( $dS = +3$ ), implied vol rose $d\sigma = +1$ point, one day passed $dt = 1$ , and rates moved $dr = +0.10$ point. All values below are scaled to the 1,000-share position.

Term	Formula	Arithmetic (per share → ×1,000)	P&L
Delta	$\Delta\,dS$	$0.55 \times 3 = 1.65$	+$1,650
Gamma	$\tfrac12\Gamma\,dS^2$	$0.5 \times 0.04 \times 9 = 0.18$	+$180
Theta	$\Theta\,dt$	$-0.05 \times 1 = -0.05$	−$50
Vega	$\mathcal V\,d\sigma$	$0.20 \times 1 = 0.20$	+$200
Rho	$\rho\,dr$	$0.12 \times 0.10 = 0.012$	+$12
Greek total	sum		+$1,992

Suppose the exact full reprice (Black–Scholes at the new $S$ , $\sigma$ , $t$ , $r$ ) shows the position actually gained $2,010. Then the residual is $2{,}010 - 1{,}992 = +18$ dollars — the higher-order error the five terms missed. Small here because the moves were modest; on a violent day it balloons, and a large residual is itself a signal (your linear-plus-gamma picture is breaking down).

Why desks do this

Three hard reasons, every one of them about survival:

Risk control. If $1,650 of today’s profit was delta P&L on a book that’s supposed to be delta-neutral, your hedge has a leak — you were accidentally taking a directional bet. Attribution catches it before it compounds.
Skill vs luck. A vol trader’s job is the vega and gamma lines — those are the bets they meant to make. If the day’s profit was all delta and rho, they got lucky on stuff they don’t control, not skilled at what they do. Attribution separates the two so a desk pays for skill, not coin-flips.
Model error. A persistently large residual means your Greeks (and the model producing them) don’t describe the position you actually hold. The unexplained term is the model raising its hand.

Attribution catches the lucky-vs-skilled difference

Two traders both book +$50k today. One made it all on the vega and gamma they were paid to trade — that’s repeatable skill. The other made it on a delta leak and a rho windfall they never intended — that’s luck that will reverse. From the P&L number alone you can’t tell them apart. From the attribution, it’s obvious. This is why desks itemize every day: the receipt, not the total, is where the truth lives.

P&L term → Greek → driver

Each line of the attribution maps a profit contribution to the Greek that measures it and the market move that lit it up:

P&L term	Which Greek	What market move drove it
Delta P&L	$\Delta$ (delta)	the stock’s directional move $dS$
Gamma P&L	$\Gamma$ (gamma)	the squared move $dS^2$ — the size of the chop
Theta P&L	$\Theta$ (theta)	the passage of time $dt$ (the clock, not the market)
Vega P&L	$\mathcal V$ (vega)	the change in implied volatility $d\sigma$
Rho P&L	$\rho$ (rho)	the change in interest rates $dr$
Residual	(higher-order)	big moves, cross-terms, drifting Greeks

Match each P&L attribution term to the market move that drives it.

Pick a term, then click its definition.

Sort each statement about gamma scalping and P&L attribution as TRUE or FALSE.

Place each item in the right group.

Theta P&L requires the stock to move
A long-gamma hedge buys the stock after it falls and sells after it rises
A large attribution residual can signal model error
Each scalp is approximately ½·Γ·(ΔS)²
A delta-hedged long option profits if realized vol exceeds implied vol
Gamma P&L can be negative for a long-gamma position

On a delta-neutral long-gamma book, today's attribution shows a large positive DELTA P&L line. What's the most useful read?

The gamma scalping worked perfectly, as intendedThe book wasn’t actually delta-neutral — there’s a hedging leak that took an unintended directional betIt proves the trader is skilled at picking directionImplied volatility must have risen sharply

Putting it together

A delta-neutral long-gamma book is a machine that gets paid for movement. The hedge mechanically buys low and sells high, banking $\tfrac12\Gamma(\Delta S)^2$ per swing; theta is the daily rent for owning that machine; and the net collapses to the master equation $\tfrac12\Gamma S^2(\sigma_{real}^2 - \sigma_{imp}^2)\,dt$ — you win exactly when the stock moves more than the implied vol you paid for. At the close, P&L attribution writes the itemized receipt: delta, gamma, theta, vega, rho, plus a residual, each a Greek times the move that drove it. The receipt is how a desk controls risk, separates skill from luck, and catches model error. You don’t trade the stock — you trade realized against implied, and you keep the books to prove which one won.

Big picture

Gamma scalping & P&L attribution

Gamma scalping & attribution
- Scalping mechanics
  - Long option, delta-hedged
  - Stock falls: delta drops, BUY low
  - Stock rises: delta climbs, SELL high
  - Hedge forces buy-low/sell-high
- The scalp P&L
  - Each swing banks about half gamma times move squared
  - Move enters squared: bigger pays disproportionately
  - Always positive for long gamma
- Theta is the rent
  - Long gamma equals negative theta
  - Bleeds daily whether or not stock moves
  - Cost of admission to the machine
- Master equation
  - P&L tracks realized minus implied variance
  - Win if realized vol beats implied vol
  - You trade realized vs implied, not the stock
- P&L attribution
  - Delta, gamma, theta, vega, rho, residual
  - Each = Greek times its market move
  - Risk control: catch hedging leaks
  - Skill vs luck; residual flags model error

The long-gamma machine, the rent it pays, the bet it really is, and the receipt that proves it.

Recap: gamma scalping & P&L attribution

Question 1 of 60 correct

A long-gamma, delta-hedged book re-hedges after the stock falls. Which trade does the hedge force?

Buy the underlying — delta dropped, so you buy back to neutral, after the fallBuy puts to add more gammaSell the underlying — to lock in the lossDo nothing — delta-neutral books never re-hedge

Check your answer to continue.

That closes the topic. You can now see a delta-neutral options book the way a desk does: not as a frozen premium but as a live machine that scalps movement, pays theta rent, and lives or dies on the realized-versus-implied vol gap — with an itemized Greek receipt proving, every single day, exactly where the money came from.