Delta Hedging and Gamma Scalping: How a Hedged Option Makes or Loses Money

Delta hedging means holding an offsetting position in the underlying so that an option’s value does not change, to first order, when the price moves; gamma scalping is running that hedge on a long option position and repeatedly rebalancing, buying low and selling high as the delta shifts. Once direction is hedged away, what is left is a bet on volatility: a hedged long option makes money if the stock actually moves more than the implied volatility it was bought at, and loses if it moves less.

In short: over each short interval, a delta-hedged option earns about one half of gamma times the squared price move, and pays theta for the time that passed. Those two cancel exactly when the stock moves at the implied volatility. So the expected P&L of a hedged option is set by realized volatility minus implied volatility, scaled by vega. In our simulation of ten hedged one-month calls on a $50 stock bought at 30% implied volatility, the average result was about +$287 at 35% realized volatility and -$288 at 25%, but individual outcomes spread several hundred dollars either side, and that spread depends on how often you hedge, what hedging costs, and whether the stock jumps.

This deep dive assumes you know the greeks individually (see delta explained, gamma explained, and theta and time decay). It describes how professionals think about hedged option P&L. It is not a strategy recommendation, and the costs and infrastructure involved are a large part of why this is professional territory.

A common misconception: “a delta-neutral position has no risk”

Delta neutral means no exposure to a small move right now. It does not mean no exposure. A delta-hedged option is fully exposed to how much the stock moves (gamma), to time passing (theta), to changes in implied volatility (vega), and to large gaps that no hedge can follow. Emanuel Derman put the practical point in a footnote to his 1999 Goldman Sachs note “Regimes of Volatility”: “the inevitably inaccurate estimation of future volatility and the discreteness of practical hedging pose significant risks.”

Where the P&L comes from

Black and Scholes priced options by showing that a continuously adjusted mix of stock and cash replicates the option. Run that argument in reverse and you get the P&L of anyone who holds an option and hedges it.

Hold one call and short delta shares. Over a short interval dt, with the stock moving by dS, a Taylor expansion of the option value gives:

change in option   = Delta x dS + 1/2 x Gamma x dS^2 + Theta x dt
change in hedge    = -Delta x dS
hedged P&L         = 1/2 x Gamma x dS^2 + Theta x dt      (plus small financing terms)

The delta terms cancel. What remains is a race between gamma and theta. In the Black-Scholes model, theta is tied to gamma by the pricing equation: ignoring interest, Theta = -1/2 x Gamma x S^2 x sigma_implied^2. Substituting:

hedged P&L over dt = 1/2 x Gamma x S^2 x [ (dS/S)^2 - sigma_implied^2 x dt ]

That one line explains most of what professionals mean by “trading volatility”:

If the squared return over the interval is larger than the implied variance for that interval, the long option holder gains. If smaller, it loses.
The gain or loss is weighted by dollar gamma, 1/2 x Gamma x S^2. Moves matter most when the option has the most gamma, which is near the strike and near expiry.
The seller of the option, hedged, has exactly the opposite P&L.

Over the life of the option, the hedged P&L is approximately the sum of these terms. It depends on realized volatility and on where the stock was when that volatility happened. That second part is the path dependence that makes hedged option P&L noisy even when you are right about volatility.

The breakeven move

Setting the bracket to zero gives the breakeven: the option pays for itself when the typical move over the interval equals implied volatility scaled to that interval. For a daily rebalance, using 252 trading days a year:

breakeven daily move = sigma_implied / sqrt(252)

At 30% implied volatility that is 0.30 / 15.87 = 1.89% a day.

Worked example: one day of gamma against theta

Hypothetical Maple Widgets Inc. trades at $50. A desk buys 10 one-month at-the-money calls (strike $50, 21 trading days to expiry, 30% implied volatility, 4% rate). Each contract covers 100 shares. From Black-Scholes:

Quantity	Per share	Position (10 contracts, 1,000 shares)
Call price	$1.8085	$1,808.53
Delta	0.5326	Short 533 shares to hedge
Gamma (per $1)	0.0918	91.8 shares of delta change per $1
Theta (per trading day)	-$0.0449	-$44.93
Vega (per vol point)	$0.0574	$57.39

Breakeven daily move: 1.89% of $50 = $0.945. For one day, the hedged position’s P&L is approximately 1/2 x 91.8 x (move)^2 minus $44.93:

Stock move in one day	Gamma gain	Theta cost	Net for the long hedged position
$0.25 (0.5%)	$2.87	-$44.93	-$42.06
$0.50 (1.0%)	$11.48	-$44.93	-$33.45
$0.945 (1.89%)	$41.00	-$44.93	about breakeven
$1.50 (3.0%)	$103.30	-$44.93	+$58.37
$2.50 (5.0%)	$286.95	-$44.93	+$242.02

(The breakeven row is slightly negative because theta here includes the interest cost of the option, which the simple formula ignores.) Two things to notice. Gains grow with the square of the move, so one big day can pay for many quiet ones. And the direction does not matter: a $1.50 fall earns the same gamma gain as a $1.50 rise, because the hedge is rebalanced after either.

What “scalping” means mechanically

Suppose the stock rises $1.50. The call’s delta rises by about 91.8 x 1.50 = 138 shares, so the position is now long about 138 shares of delta. To get back to neutral the desk sells 138 shares at the higher price. If the stock then falls back, delta drops and the desk buys shares at the lower price. Every round trip sells high and buys low, and those trading gains are the gamma P&L. The long option is what forces the hedger to trade in that direction. A short option position forces the reverse: buying after rises and selling after falls, which is how hedged option sellers lose when markets move a lot.

Simulation: realized versus implied volatility

One day tells you the mechanics; the distribution tells you the risk. We simulated the position above over 20,000 random price paths (lognormal, zero drift, constant volatility, no jumps), rehedging to the Black-Scholes delta at 30% implied volatility once a day, with cash earning 4%, and holding to expiry. The only thing we varied was the volatility the stock actually realized.

Realized volatility	Average P&L	Std. deviation	5th percentile	95th percentile	Paths with a profit	BS value at realized vol minus price paid
20%	-$576	$274	-$1,065	-$164	0.7%	-$574
25%	-$288	$284	-$793	+$138	13.7%	-$287
30%	$0	$320	-$519	+$532	49.3%	$0
35%	+$287	$391	-$263	+$995	78.9%	+$287
40%	+$574	$493	-$53	+$1,505	93.0%	+$574

Three results stand out.

The average is the volatility gap times vega. At 35% realized, the average of $287 matches the Black-Scholes value at 35% minus the price paid at 30%, and is close to vega times the gap ($57.39 x 5 = $287). The hedged option really is a volatility position.
Being right is not enough. At 35% realized, five points above implied, more than one path in five still lost money. At 25%, almost one in seven made money. Daily hedging leaves a lot of noise.
Unhedged, the same calls are a different instrument. Holding the ten calls without any hedge at 30% realized volatility produced an average of about -$90, a standard deviation of about $2,630, and a profit on only 32% of paths. Hedging cut the standard deviation by about 88%.

Where the noise comes from: gamma-weighted variance

To check the decomposition, we computed for each simulated path the sum of 1/2 x Gamma x S^2 x (daily return^2 - implied variance per day) and compared it with that path’s actual hedged P&L. The correlation across the 20,000 paths was 0.97. Almost all of the spread in outcomes comes from when the big moves happened: a large move with the stock near $50 and a week to go is worth far more than the same move with the stock at $45, where the call has little gamma.

Hedging frequency and costs

Rehedging more often makes the outcome less dependent on any single day. With realized volatility equal to implied (30%), 10,000 paths:

Rehedging frequency	Average P&L	Std. deviation	5th to 95th percentile
Weekly (every 5 trading days)	+$8	$657	-$999 to +$1,135
Daily	-$1	$316	-$514 to +$523
4 times a day	-$2	$161	-$262 to +$265
26 times a day (about every 15 minutes)	$0	$65	-$105 to +$105

The standard deviation roughly halves each time the hedging frequency rises fourfold, which is the square-root rule you would expect from averaging many small independent errors. In the limit of continuous hedging, the Black-Scholes replication argument says the noise would vanish.

It does not vanish in practice, because every rebalance costs money. Charging an illustrative $0.02 per share on each rebalancing trade (half-spread plus fees):

Rehedging frequency	Average P&L at 30% realized	Average P&L at 35% realized	Share of paths profitable at 35%
Daily	-$40	+$244	75.5%
4 times a day	-$70	+$213	88.5%
26 times a day	-$159	+$116	94.5%

More frequent hedging buys certainty with expected profit. At 35% realized volatility, 15-minute hedging made money on 94.5% of paths but kept only $116 of the $287 edge on average; daily hedging kept $244 but lost money on about one path in four. Real desks choose rebalancing rules (time-based, or bands around the target delta) to balance exactly this trade-off, and they net hedges across a whole book so that offsetting positions do not each pay costs.

Which volatility to hedge with

The delta you hedge with depends on the volatility you plug into the formula. Hedging at implied volatility gives the path-dependent P&L above. Hedging at the volatility that is actually realized gives, in theory, a P&L that is locked in from the start: the Black-Scholes value at realized volatility minus the price paid, with no path dependence. The catch is that you do not know realized volatility in advance.

Simulating 10,000 paths with the stock realizing 40% against 30% implied:

Hedging rule	Rehedge frequency	Average P&L	Std. deviation	5th percentile
Delta at implied (30%)	Daily	+$571	$488	-$50
Delta at realized (40%)	Daily	+$574	$422	-$112
Delta at implied (30%)	26 times a day	+$572	$263	+$192
Delta at realized (40%)	26 times a day	+$576	$86	+$437

The theoretical locked-in value is $574. Both rules earn it on average. Hedging at realized volatility strips most of the path dependence out once hedging is frequent, and with 15-minute rebalancing every one of the 10,000 paths made money under either rule. With daily hedging, discreteness dominates and the choice barely matters. In practice, desks hedge at implied volatility, or at a smile-adjusted delta, because it marks consistently with the market; the choice of smile-adjusted delta is discussed in the volatility surface and skew.

Jumps: the risk hedging cannot remove

The simulation so far assumes continuous paths. Real stocks gap on earnings, news, and overnight. A gap moves the price before the hedge can be adjusted, so the position collects or pays the full 1/2 x Gamma x (gap)^2 at once.

We reran the 30% case for a desk that sold the ten calls and hedged daily, and inserted a 10% overnight drop on day 10:

Scenario (short 10 calls, hedged daily)	Average P&L	Std. deviation	Paths profitable
No gap, realized = implied = 30%	$0	$320	50.7%
One -10% gap on day 10	-$1,093	$580	0.3%

The size of the loss is roughly what the formula predicts: 1/2 x 91.8 x ($5)^2 = $1,148 at today’s gamma. The seller collected $1,809 of premium for the calls, and one gap cost about 60% of it. A gap upward would cost a short-gamma book similarly. Robert Merton’s 1976 jump model makes the formal point: when prices can jump, no continuous stock hedge replicates the option, so the price of an option must include compensation for jump risk that hedging cannot remove.

What the evidence says about hedged options

If implied volatility were an unbiased forecast of realized volatility, delta-hedged long options would break even on average. They do not, at least for index options. Gurdip Bakshi and Nikunj Kapadia, in “Delta-Hedged Gains and the Negative Market Volatility Risk Premium” (Review of Financial Studies, 2003, volume 16, pages 527-566), studied delta-hedged positions in S&P 500 index options. They found the hedged long-option strategy loses money on average, with smaller losses for options away from the money and larger losses when volatility is high, and they interpret this as evidence of a negative market volatility risk premium: buyers of index options pay more for volatility than it typically delivers, as compensation to sellers for bearing crash and volatility risk. The broader evidence on that premium is covered in the variance risk premium.

That premium is what option-selling desks are paid to collect, and the gap simulation above is what they are paid to bear. Market makers who take the other side of customer flow manage it with inventory limits, vega and gamma limits, and hedges in other options, as described in how options market makers hedge.

What can go wrong

Realized volatility comes in below implied. The long gamma position bleeds theta every day. At 20% realized against 30% implied, 99.3% of our simulated paths lost money.
Gaps on the short side. One 10% overnight move erased about 60% of the premium a hedged seller collected. Earnings dates, index rebalances, and macro announcements are known gap risks; others are not.
Implied volatility moves before expiry. Our simulations held each position to expiry. A position closed early is marked at the new implied volatility, so vega P&L (about $57 per vol point here) can swamp a month of gamma and theta.
Transaction costs and slippage. Costs scale with hedging frequency, and spreads widen exactly when markets move fast and rehedging is most needed.
Model and smile risk. A delta computed at one volatility, or under one sticky rule, can be materially wrong when the surface shifts.
Pin and expiry risk. Near expiry, gamma for an at-the-money option becomes very large, so small moves around the strike flip delta from near 0 to near 1 and force large hedge trades.
Financing and borrow. Short stock hedges need a borrow, and its cost or availability can change. Long stock hedges need funding. Both were ignored in the simplified figures above except for a flat 4% rate.

The short version

A delta-hedged option turns a directional bet into a volatility bet. Each period it earns half gamma times the squared move and pays theta, and the two balance when the stock moves at implied volatility, which for 30% vol is 1.89% a day. On average the hedged P&L equals vega times the gap between realized and implied volatility, as our simulations reproduce, but the outcome on any one path depends on when the big moves happen relative to the strike. More frequent hedging cuts that noise roughly with the square root of the number of rebalances and costs more in fees and spreads. Jumps cannot be hedged at all. Evidence on S&P 500 options shows hedged long positions lose on average, which is the volatility risk premium sellers earn for bearing exactly those gap risks.