Volatility Strategies Advanced Neutralising directional exposure Continuous / path-dependent

Delta Hedging

The hedge that removes the risk you can see and hands you the ones you cannot.

Quick answer: Delta hedging is the practice of holding a position in the underlying that offsets the directional exposure of an option, so that small moves in the spot price no longer change the combined value — a neutrality that Black–Scholes assumes is continuous and free, and that every real desk pays for in spreads and re-hedges.

In simple words

An option's value moves when the underlying moves, and delta is how much: a call with a delta of 0.50 gains about ₹0.50 for every ₹1 NIFTY rises. If you have sold that call and you do not want the directional bet, you buy 0.50 units of NIFTY exposure against it. Now a small move up costs you on the option and pays you on the futures, and the two roughly cancel. That is a delta hedge. The word 'roughly' is where the entire subject lives, because delta itself changes as the underlying moves, so a hedge that is perfect at 24,000 is already wrong at 24,100.

The textbook says you should adjust the hedge continuously, re-buying and re-selling the underlying every instant so the offset is always exact. In the real world you cannot trade continuously, and every trade costs a bid-ask spread, so you re-hedge in steps — when the delta has drifted far enough to be worth the cost. Between those steps you are carrying a directional bet you did not choose to take. Hedging more often makes those bets smaller and the trading bill larger; hedging less often does the reverse. Choosing where to stand between the two is not a detail of the job — it is the job.

Not to be confused with: Gamma hedging, which is a separate and harder problem. Delta hedging fixes the first-order directional exposure at one price; gamma is the rate at which that delta itself changes, and it is precisely what forces you to re-hedge at all. A book can be delta-neutral and still lose or gain sharply on a large move, because its delta was only zero at the price you last hedged. Removing gamma requires trading other options, not the underlying.

The step hedge versus the smooth true delta

Every vertical gap is unhedged risk

The smooth curve is an option's true Black–Scholes delta as NIFTY moves; the step function is the hedge actually held by a trader who only re-adjusts every 400 points.

The staircase can never sit on the curve except at the instants it is redrawn. Between re-hedges the held delta and the true delta diverge, and the vertical distance between them is a directional position the trader is carrying without intending to. Drawing the steps closer together shrinks every gap and multiplies the number of trades — this picture is the trade-off, not an illustration of it.

Professional explanation

The assumption Black–Scholes makes, and the bill reality sends

The derivation of the Black–Scholes equation requires that the hedge be rebalanced continuously and at no cost. Under those two assumptions the directional risk of an option can be removed entirely, and what remains is a deterministic relationship between the option's price, its time decay and its gamma. Neither assumption is true. Continuous rebalancing is impossible — there is a smallest interval at which a human or a machine can act, and below it the market has already moved. Costless rebalancing does not exist — every adjustment crosses a bid-ask spread, pays exchange and clearing fees, and in India pays securities transaction tax. So the clean result becomes a messy one: the risk is not removed, it is reduced, and the reduction is bought with a stream of transaction costs that has no upper bound if you hedge too eagerly.

The trade-off has no free corner

State the two forces plainly. Re-hedging more frequently keeps the held delta close to the true delta, so the unhedged directional gaps stay small — but each adjustment costs a spread, and the costs accumulate with the number of trades. Re-hedging less frequently saves those costs but lets the gaps grow, so the book carries larger unintended directional bets between adjustments, and the variance of the final profit and loss rises. There is no rebalancing frequency at which both the hedging error and the transaction cost are zero, because the two move in opposite directions in the same variable. Every practitioner is therefore choosing a point on a curve that trades expected cost against variance, and there is no setting that dominates all others. Pretending the choice does not exist does not remove it; it just makes it for you, badly.

Hedging bands: the formal version of 'when it is worth it'

The academic treatment of this trade-off produces a no-trade band around the target delta. Inside the band you do nothing; only when the delta drifts to the edge do you rebalance, and then only back to the edge, not all the way to the centre. The Leland approach adjusts the volatility used in the hedge to account for transaction costs directly. The Whalley–Wilmott result derives the band width from a utility trade-off and finds it scales with the cube root of transaction costs divided by risk aversion and gamma — wider bands where costs are high, gamma is low, or the trader is more cost-averse. The practical lesson survives even if the exact formula does not: the optimal hedge is not the most frequent hedge, it is the least frequent hedge whose remaining risk you are willing to hold, and that width should be wider when gamma is small and narrower when gamma is large.

In India there is one instrument, and it is the futures

To delta-hedge a NIFTY option you need a liquid, low-cost way to take a position of the opposite sign in NIFTY itself. There is no tradeable cash instrument on the index — you cannot buy 'the NIFTY' the way you can buy a single share — so the only practical hedge is the NIFTY future. It is the right choice for reasons beyond availability: the futures market is deep, its impact cost is small relative to a basket of fifty stocks, and securities transaction tax on futures is a fraction of what a delta-equivalent basket of cash equities would attract. A trader who tried to replicate the index with its constituents to hedge would pay the transaction cost fifty times over and chase fifty separate spreads. The futures is not merely convenient; for retail-scale and most institutional purposes it is the only route that survives its own costs.

The delta from a flat-vol model is the wrong delta when skew exists

Black–Scholes assumes a single volatility for all strikes. Real markets do not, and the volatility smile means that when the spot moves, the implied volatility attached to a given strike tends to move too. A delta computed as if volatility were fixed ignores that second channel. The minimum-variance delta corrects it: Δ_MV = Δ_BS + ν·(∂σ/∂S), adding the vega multiplied by how the strike's implied volatility responds to spot. The sign of the correction depends on the regime. Under sticky-strike behaviour, where each strike's implied volatility stays roughly put as spot moves, the correction is small. Under sticky-delta or the empirically common negative spot-vol correlation of equity indices, a fall in spot raises implied volatility, so ∂σ/∂S is negative, and the minimum-variance delta of a call is lower than its Black–Scholes delta. A desk that hedges on the raw Black–Scholes delta in an index systematically over-hedges, and the error is not random — it leans one way.

Formula

The hedge, and the delta that actually minimises variance

Position delta = Δ_option × contracts × lot; hedge = −(position delta) units of NIFTY futures

A long call has positive delta, so it is hedged by a short futures position, and vice versa. Neutrality holds only at the spot price where the delta was evaluated; as spot moves, gamma changes the delta and the hedge must be adjusted. The lot converts a per-unit delta into a whole-contract exposure.

Δ_optionThe option's delta per unit of underlying — the sensitivity of its price to a ₹1 move in spot, between 0 and 1 for a call and 0 and −1 for a put.
contractsNumber of option contracts held; negative if the position is short.
lotContract multiplier — 75 for NIFTY, 30 for BANKNIFTY — converting a per-unit delta into units of the underlying.
position deltaTotal directional exposure of the option position, in units of the underlying, before hedging.

The minimum-variance delta under skew

Δ_MV = Δ_BS + ν × (∂σ/∂S)

The Black–Scholes delta assumes implied volatility is fixed as spot moves. When skew exists, the strike's implied volatility itself moves with spot, and the second term captures that. In an equity index, spot and implied volatility are negatively correlated, so ∂σ/∂S is typically negative, making the minimum-variance delta of a call lower than its Black–Scholes delta.

How a desk delta-hedges an option in practice

Compute the option's current delta from a pricing model, using the mid-market implied volatility rather than a stale last-traded price.
Multiply by the number of contracts and the lot size to get the position delta in units of the underlying.
Take the opposite position in NIFTY futures of the same size — short futures against a positive delta, long futures against a negative one.
Decide a no-trade band around zero net delta rather than re-hedging on every tick, and set it wider where gamma is small and narrower where gamma is large.
Re-hedge only when the net delta reaches the edge of the band, and adjust back to the edge, not all the way to the centre, to avoid paying spreads on noise.
If the position sits in an equity index with pronounced skew, consider hedging on the minimum-variance delta rather than the raw Black–Scholes delta, because the two differ systematically.
Track the running cost of the hedging programme separately from the option's own value, because the transaction bill is the price of the neutrality and it is easy to overlook until it has eaten the position.

Practical example

NIFTY worked example

You are short one NIFTY 24,000 call, 30 days out, at an implied volatility of 12.8%. Its delta is about 0.52, so with a lot of 75 your position delta is −0.52 × 75 ≈ −39 units of NIFTY: a short call is short delta, meaning you lose if NIFTY rises. To neutralise it you buy 39 units of NIFTY exposure — a bit over half a futures lot in delta terms — so a small rise now pays on the futures what it costs on the call. Suppose NIFTY climbs 400 points to 24,400 before you next look. The call's delta has risen toward 0.60, so your true position delta is now about −45, but your hedge is still sized for −39. The six-unit gap is a short directional bet you have been carrying, unhedged, for the whole 400-point move — roughly 400 × 6 ≈ ₹2,400 of drift on this single contract that had nothing to do with any view. Re-hedge more often and that number shrinks while your spread bill grows. That tension is the entire exercise, and it never resolves.

BANKNIFTY worked example

BANKNIFTY makes the cost side of the trade-off vivid because it moves further and faster. Take a short BANKNIFTY 52,000 call with a delta of 0.50 and a lot of 30: the position delta is −15 units, hedged with 15 units of BANKNIFTY futures. BANKNIFTY routinely travels 600–800 points in a session, and its gamma near the strike is large, so the delta that was 0.50 in the morning can be 0.58 by the afternoon. A trader who re-hedges tightly to keep the gap small will trade several times a day, each time paying the BANKNIFTY futures spread, which is wider than NIFTY's. A trader who re-hedges loosely to save that bill carries larger directional gaps through a faster-moving index. The same instrument that makes the hedge cheap to place — the deep, single BANKNIFTY future — makes it expensive to place often. Neither trader has removed the risk; they have chosen different mixtures of hedging error and cost.

Lot sizes used above (NIFTY 75, BANKNIFTY 30) are those in force at the time of writing; NSE revises them periodically. Figures exclude brokerage, STT, exchange charges, stamp duty and GST. Examples are teaching scenarios built on round numbers — they are not historical quotes, not backtests and not trade calls.

Risk note. A perfectly delta-hedged book is not a riskless book. It has exchanged the directional risk you understood for gamma, vega and hedging-error risk, all of which are larger and less intuitive. A delta-neutral short-option position still loses on a large move because it is short gamma; it still loses when implied volatility rises because it is short vega; and it still leaks value through discrete re-hedging even when nothing else goes wrong. Delta-neutral does not mean loss-neutral, and treating it as though it does is how neutral books blow up.

Advantages & limitations

What it is good for

It isolates the exposure you actually want. Stripping out direction leaves a position whose profit and loss depends on volatility, gamma and time rather than on whether the market went up or down, which is the entire point of trading options as volatility instruments.
It is mechanical and auditable. The delta of a book is a computable number, the required hedge is a computable number, and a risk manager can check at any moment whether the two agree — unlike a discretionary directional view, which cannot be measured against a target.
In India it has a single, deep instrument. The NIFTY and BANKNIFTY futures are liquid enough that the hedge can be placed at small impact cost, which is not true of hedging a fifty-name basket, so the mechanics are unusually clean for retail-scale sizes.
It scales continuously. A hedge can be any size, adjusted in small increments, so exposure can be dialled toward zero rather than switched on and off, which lets a desk carry an inventory of options and manage the net rather than each line.
It converts an unmanageable problem into a managed one. You cannot forecast direction reliably, but you can measure and offset it, so delta hedging replaces a bet you cannot win with a cost you can budget.

Where it breaks down

It only holds at the price where the delta was evaluated. Gamma changes the delta the instant the spot moves, so a hedge that is exact at 24,000 is already stale at 24,050, and the neutrality is a snapshot, not a state.
It assumes you can trade continuously, and you cannot. Between re-hedges the book carries an unintended directional position whose size grows with the move and with the time since the last adjustment — the vertical gaps in the chart on this page.
It assumes trading is costless, and it is not. Every adjustment pays a bid-ask spread and, in India, securities transaction tax, so hedging more accurately costs more money, and there is no frequency that makes both the error and the cost small.
The delta it uses is model-dependent. A delta from a flat-volatility model ignores that implied volatility moves with spot under skew, so it systematically over- or under-hedges an equity index unless corrected to the minimum-variance delta.
It removes only the first-order risk. A delta-neutral book is still exposed to gamma, to vega, to changes in the skew, and to the accumulated error of the hedging programme itself, none of which the delta hedge touches.
It breaks down in a gap. If the underlying jumps overnight past several re-hedging points at once — an event, a shock, a limit move — there was no opportunity to adjust, and the discrete hedge behaves nothing like the continuous one the theory assumes.

Common mistakes

Believing a delta-neutral book is a safe book. Neutral means the first-order directional term is offset at one price; it says nothing about gamma or vega, and a short-gamma neutral book loses on exactly the large move the trader felt protected against.
Re-hedging on every tick. Chasing the delta back to zero continuously converts small market noise into a stream of spread payments, and the transaction bill quietly consumes the position while the trader congratulates themselves on tight risk control.
Never re-hedging. Setting the hedge once and leaving it lets gamma pull the delta far from zero, so the book becomes a large directional bet in disguise — the opposite failure, and just as common.
Hedging on the raw Black–Scholes delta in a skewed index and ignoring the minimum-variance correction. The error is not random noise that averages out; it leans one direction, so it accumulates rather than cancels.
Hedging a NIFTY option with a proxy basket of stocks to save on futures. This multiplies the transaction cost across fifty names and fifty spreads, and the tracking error between the basket and the index adds a new risk on top of the one being hedged.
Forgetting that theta pays for the gamma. A long-gamma hedged book that re-hedges profitably on movement is paying for that privilege in time decay every day, and a trader who counts the hedging gains without the theta bill is reading half the ledger.

Professional usage

A market-making desk does not hold a view on direction at all; it quotes options to customers and inherits whatever delta the flow leaves behind, then hedges that delta in the futures continuously through the session to keep the book close to neutral while it earns the bid-ask spread on the options themselves. The hedging is not the trade — it is the cost of being allowed to run the trade, and the desk's skill is in hedging just enough to control risk without letting the futures spread bill overwhelm the option edge. Sophisticated desks hedge on a minimum-variance or model-calibrated delta rather than the raw Black–Scholes number, precisely because in an index the two differ in a predictable direction and the difference compounds over thousands of re-hedges.

Volatility arbitrage desks delta-hedge for a different reason: they have a view on volatility, not direction, and hedging is how they remove the direction so that only the volatility view remains in the profit and loss. There the hedging frequency is itself a research question — hedge too often and transaction costs swamp a thin volatility edge, hedge too rarely and the path-dependent hedging error swamps it instead — so the desk studies the optimal band width for the specific instrument, and treats the hedging programme as part of the strategy rather than an afterthought to it.

Key takeaways

Delta hedging offsets an option's directional exposure by holding the opposite position in the underlying, but the neutrality is exact only at the price where the delta was evaluated.
Black–Scholes assumes continuous, costless re-hedging; reality is discrete and pays a spread on every adjustment, so the risk is reduced rather than removed and the reduction is bought with transaction costs.
The frequency of re-hedging trades variance against cost, and there is no setting at which both are zero — choosing the point on that curve is the substance of the job, formalised by no-trade hedging bands.
In India the only practical hedge for an index option is the futures, because there is no cash index instrument and a constituent basket multiplies costs and adds tracking error.
A delta computed from a flat-volatility model is wrong under skew; the minimum-variance delta Δ_MV = Δ_BS + ν·(∂σ/∂S) corrects it, and in an equity index the correction lowers a call's delta.
A delta-neutral book is not a riskless book — it has swapped directional risk for gamma, vega and hedging-error risk, which are larger and less intuitive.

Read delta hedging as a cost, not a cure. It removes the one risk you could already see — direction — and in doing so hands you three you find harder to see, then charges you a running fee for the privilege. The honest description of a delta-neutral options book is not that it is riskless but that it has traded a familiar risk for unfamiliar ones and pays a spread to keep doing so. The interesting decisions in options trading almost all live in that trade, and a trader who thinks delta hedging has made them safe has usually just stopped watching the parts that can hurt them.

Frequently asked questions

What is delta hedging?

Delta hedging is holding a position in the underlying that offsets the directional exposure of an option, so that small moves in the spot price no longer change the combined value. A short call with delta 0.52 is hedged by holding about 0.52 units of long underlying exposure per unit of the contract.

Why is delta hedging not risk-free?

Because it only neutralises the first-order directional term, and only at the price where the delta was measured. As soon as the underlying moves, gamma changes the delta so the hedge is stale, and the book remains exposed to gamma, vega and the accumulated error of hedging in discrete steps.

How often should you re-hedge a delta-neutral position?

There is no universally right frequency, because re-hedging more often shrinks the directional gaps and raises transaction costs, while re-hedging less often does the reverse. The formal answer is a no-trade band: rebalance only when the delta drifts to the edge, with the band wider where gamma is small.

What instrument is used to delta-hedge a NIFTY option in India?

The NIFTY future, because there is no tradeable cash instrument on the index itself. The futures market is deep, its impact cost is low, and its securities transaction tax is far smaller than hedging with a basket of the fifty constituent stocks would incur.

Why can't you hedge continuously as Black–Scholes assumes?

Because there is a smallest interval at which anyone can actually trade, and below it the market has already moved, so the hedge is always slightly stale. Continuous, costless rebalancing is a mathematical idealisation that makes the derivation work, not a description of any real desk.

What is the minimum-variance delta?

The minimum-variance delta is the hedge ratio that accounts for implied volatility moving with spot under skew: Δ_MV = Δ_BS + ν·(∂σ/∂S). The ordinary Black–Scholes delta assumes volatility is fixed, so it ignores the second channel and mis-hedges a skewed index.

Does delta hedging remove gamma?

No. Delta hedging with the underlying cannot change gamma at all, because the underlying has zero gamma. Gamma is precisely what makes the delta drift and forces re-hedging, and removing it requires trading other options, not more of the underlying.

Why does a delta-neutral short-option book still lose on a big move?

Because it is short gamma. Delta-neutral means the position does not move for a small change in spot, but gamma means the delta itself shifts against you as the move grows, so a large or fast move produces a loss that the delta hedge was never able to prevent.

What are hedging bands?

Hedging bands are a no-trade region around the target delta: inside the band you do nothing, and only at the edge do you rebalance, and then only back to the edge. They formalise the trade-off between hedging error and transaction cost rather than re-hedging on every tick.

What does the Whalley–Wilmott result say about band width?

It derives the optimal band width from a utility trade-off and finds it scales with the cube root of transaction costs divided by risk aversion and gamma. In plain terms, the band should be wider when costs are high or gamma is low, and narrower when gamma is large.

Why is the Black–Scholes delta wrong when there is skew?

Because it holds implied volatility fixed as spot moves, but with skew the volatility attached to a strike moves with spot too. Ignoring that second effect means the hedge is systematically the wrong size, and in an equity index the error leans one direction rather than averaging out.

What is the sign of the skew correction under sticky-strike?

Under sticky-strike behaviour, where each strike's implied volatility stays roughly fixed as spot moves, ∂σ/∂S for a given strike is small, so the minimum-variance correction to the delta is small. The correction matters far more under sticky-delta or negative spot-vol correlation.

Does hedging with the futures introduce basis risk?

Yes, a small amount. The futures price can diverge from the spot index through the cost of carry and short-term supply and demand, so the hedge is not perfect even ignoring gamma. In practice the basis is small and predictable relative to the risks the hedge removes.

Can I delta-hedge without a pricing model?

Not properly, because delta is a model output — you need spot, strike, time, rate and an implied volatility to compute it. You can approximate with the option chain's displayed delta, but that number is itself produced by a model, so you are trusting someone else's assumptions rather than removing them.

Is theta the price of a long-gamma hedge?

Yes, in effect. A long-gamma hedged book profits when the underlying moves because re-hedging buys low and sells high, but it pays for that structure through time decay every day, so the gamma gains and the theta losses are two sides of the same position.

Why does re-hedging too often lose money?

Because every adjustment crosses a bid-ask spread and pays transaction taxes and fees, so chasing the delta back to zero on market noise converts that noise into a steady stream of costs. The neutrality improves marginally while the trading bill grows without a matching benefit.

What happens to a delta hedge in an overnight gap?

It fails to behave like the continuous hedge the theory assumes, because there was no chance to rebalance through the move. The book jumps past several re-hedging points at once, so a short-gamma position takes the full gap loss that frequent hedging was meant to avoid.

Is delta hedging the same as being market-neutral?

Only in the narrowest first-order sense. A delta-hedged book is neutral to small directional moves at the current price, but it is not neutral to volatility, to large moves, or to changes in the skew, so calling it market-neutral overstates what has actually been removed.

Why do market makers delta-hedge?

Because they quote options to earn the bid-ask spread and inherit whatever delta the customer flow leaves behind, not because they have a directional view. Hedging that delta in the futures lets them keep the option spread while shedding the direction they never wanted.

Does a higher gamma mean I should hedge more often?

Generally yes, because higher gamma makes the delta drift faster, so the unhedged gap grows more quickly between adjustments and a narrower band is justified. Near expiry and near the strike, gamma spikes, which is exactly when hedging becomes most demanding and most expensive.

Can delta hedging make an option position lose money that would otherwise have profited?

Yes. If you buy an option expecting a move and delta-hedge it, you have removed the directional payoff, so a large favourable move no longer helps you — your profit now depends on realised volatility and the hedging path, which can leave you worse off than the naked option would have.

What is delta-hedging error?

Delta-hedging error is the difference between the profit and loss of a real discrete hedge and the deterministic result the continuous-hedging theory predicts. It arises from the gaps between re-hedges and is path-dependent, so two identical positions hedged on different schedules end with different outcomes.

Voice search & related questions

Natural-language questions people ask about delta hedging.

How does delta hedging work in plain English?

You hold enough of the underlying to cancel the option's directional exposure, so a small move up or down roughly nets to nothing between the two. Because the option's delta keeps changing as the market moves, you have to keep adjusting the amount of underlying you hold.

If I hedge my delta am I safe now?

No. You have removed the directional bet at today's price, but you are still exposed to gamma on a big move, to a rise in implied volatility, and to the cost and error of adjusting the hedge over time. Delta-neutral is not the same as loss-neutral, and treating it as safe is a common and expensive mistake.

Why do I have to keep adjusting a delta hedge?

Because delta is not constant — gamma makes it change every time the underlying moves, so a hedge that was exact this morning is wrong by the afternoon. Each adjustment costs a spread, which is why you rebalance in steps rather than continuously.

Can retail traders in India actually delta-hedge?

Yes, using NIFTY or BANKNIFTY futures, which are liquid and cheap enough to place the hedge at small impact cost. The harder part is not placing the hedge but deciding how often to adjust it, because frequent hedging eats the position in spreads.

Is it cheaper to hedge with stocks or futures?

Futures, decisively, for an index option. Replicating the index with its fifty constituents means paying the spread and transaction tax fifty times and carrying tracking error, whereas one futures contract does the job in a single deep market.

Does hedging remove all the risk from selling an option?

No. It removes the directional part but leaves you short gamma and short vega, which is where the real danger in a short-option book lives. The move that hurts a delta-neutral short-option position is exactly the large, fast one people assume the hedge protects against.

Why do people say delta hedging is expensive?

Because to keep the hedge accurate you have to trade frequently, and every trade pays a bid-ask spread and transaction tax. The more precisely you hedge, the larger that running bill, so accuracy and cost pull in opposite directions and you can never make both small.

Sources & references

Last reviewed 10 July 2026. Educational content only — not investment advice.