Delta: Directional Bias & Portfolio Hedging

Delta is the rate of change of the option premium with respect to the underlying. Everything else — time, volatility, interest rates — is held still while you ask that one question. If NIFTY moves one point and nothing else changes, delta says how many rupees of premium that is worth. A delta of 0.52 means 52 paise per index point.

Plotted against the underlying, delta traces an S-curve rather than a straight line. Far below the strike a call barely responds, so delta sits near zero. Far above the strike the call is behaving almost exactly like the index itself, so delta approaches 1. In between there is a steep middle section around the strike, where a small move in the index produces a large change in how responsive the option is. That steep section is where all the interesting behaviour lives, and its steepness is gamma, which module 12 takes apart.

Two properties follow from the shape. First, delta is bounded: a single call cannot have a delta above 1, and a single put cannot go below −1. An option can never respond more than one-for-one with the thing it is written on. Second, delta is not a constant. The number your terminal shows is correct for the current spot, current time and current volatility assumption, and it will be a different number tomorrow even if the index closes unchanged.

A call gains when the underlying rises, so its delta is positive: between 0 and +1. A put gains when the underlying falls, so its delta is negative: between −1 and 0. The sign is not a convention chosen for tidiness. It is doing arithmetic. A NIFTY put with a delta of −0.40 in a session where the index drops 100 points gains ₹40 of premium, because −0.40 multiplied by −100 is +40. The two negatives are what turn a falling market into a rising premium.

Both of those statements are from the holder’s point of view — the person who bought the option. Selling flips every sign. A short call carries negative delta, because the seller loses when the index rises. A short put carries positive delta, because the seller gains when the index rises. This is the single most common source of confusion in a multi-leg position, and the fix is mechanical: write down whether you are long or short each leg before you add anything up.

The at-the-money case deserves one note of precision, because rounded numbers hide it. An at-the-money call does not sit at exactly 0.50. It sits slightly above, often 0.52 or 0.53, and the at-the-money put slightly above −0.50 in magnitude terms is not the mirror image. The reason is the model’s built-in drift and the discounting of the strike. It is a small detail, but it is why a straddle built at the money is not perfectly delta-flat at inception.

Reading one row of an option chain tells you very little. Reading the delta column down the chain tells you almost everything about how the different strikes will behave. Suppose NIFTY is at 24,500 with twelve sessions left in the front expiry. The deep in-the-money call moves nearly point-for-point with the index. The far out-of-the-money call barely notices a 100-point rally at all.

The table below prices the same 100-point move against five different strikes. Every one of them is a bullish position. They are not remotely the same bullish position. The 23,500 call captures ₹6,900 on one lot. The 25,500 call captures ₹450 on the same move. If you buy the far strike because it is cheap, you have not bought a cheaper version of the same bet — you have bought a different bet, one that needs a much larger move to do anything.

This is what module 5 on moneyness is describing in structural terms and what delta puts a number on. A trader who says "I am bullish on NIFTY" has said nothing about which of these five positions they want. The choice between them is a choice about how much of the move you need to capture against how much premium you are willing to put at risk, and the delta column is where that trade-off is visible.

It is close to it, and the distinction matters enough to spell out. In the Black-Scholes framework, a call’s delta is N(d₁) and the model’s probability of the option finishing in the money is N(d₂). Those are two different quantities from the same equation, and N(d₁) is always the larger of the two for a call. So reading delta as the probability of finishing in the money systematically overstates it, by a little on a near-dated at-the-money strike and by more on a long-dated or high-volatility one.

There is a second, deeper gap. Even N(d₂) is a risk-neutral probability, which as module 7 explained is a bookkeeping construct designed to remove arbitrage from the pricing, not a forecast about the real world. It contains no view on whether NIFTY is likely to rise. Calling it "the chance of expiring in the money" is a useful shorthand that is wrong in two separate ways stacked on top of each other.

What survives all that is the ranking. A 0.30-delta strike genuinely is less likely to finish in the money than a 0.50-delta strike, and far more likely than a 0.05-delta strike. Using delta to sort strikes by how demanding they are is sound. Using it to compute an expected value, or to tell yourself a 0.30-delta short option "wins seven times out of ten", is not — and that second habit is how a book of small credits gets built on a number that was never a forecast.

Individual deltas are only useful once they are combined. A position delta is the sum of every leg, with each leg converted into the same unit — index points of exposure — and signed according to whether you are long or short it. The arithmetic is simple and the discipline is everything: get one sign wrong and the total is not merely inaccurate, it points the wrong way.

Work through a two-leg example. You are long 4 lots of the NIFTY 24,500 call at delta 0.52. Multiply: 0.52 × 75 × 4 gives a position delta of +156. That means the book behaves, for small moves, like being long 156 units of NIFTY. Add a short of 2 lots of the 24,800 call at delta 0.32, and that leg contributes −0.32 × 75 × 2, or −48. Net position delta is +108.

Now the number is actionable. A 50-point rally in NIFTY is worth roughly 108 × 50, or about ₹5,400, to that book. A 50-point fall costs about the same. That single figure replaces staring at four separate premiums and guessing. Every risk system on a professional desk reduces a book of hundreds of positions to exactly this number, refreshed continuously.

A market maker does not want a directional view. They are being paid to provide two-sided quotes and collect the spread, and every option they take on brings unwanted directional exposure with it. So they compute their aggregate position delta and cancel it with the cheapest, most liquid instrument available, which for an index book is the index future. That process is delta hedging, and the target state is delta neutrality.

Take the +156 position from the previous section. One NIFTY futures lot carries a delta of 75 — a futures contract moves point-for-point with the index by definition, so its delta per unit is exactly 1. Shorting two lots contributes −150 and takes the book to a net delta of +6. That residual is small enough to ignore. A 100-point move in either direction is now worth about ₹600 rather than ₹15,600.

The catch is that delta neutrality does not stay. As NIFTY moves, the option deltas change while the futures delta does not, so the book drifts back into directional exposure and has to be re-hedged. The rate of that drift is gamma. This is why a desk hedges continuously rather than once, why gamma near expiry is genuinely dangerous, and why a fund with a physical equity portfolio uses the same arithmetic in reverse — that application is the whole of module 31.

Three things change delta while you sleep. The obvious one is spot: as the underlying moves toward or away from the strike, the option slides along the S-curve. The second is time. As expiry approaches, the S-curve steepens — in-the-money options push harder toward 1, out-of-the-money options collapse harder toward 0, and the transition between them compresses into an ever-narrower band around the strike. An option that was 0.35 delta with fifteen sessions left may be 0.20 with three left, at the very same index level.

The third is volatility, and it works in the direction most people do not expect. When the volatility assumption rises, every strike looks more reachable, so out-of-the-money deltas rise and deep in-the-money deltas fall — the whole curve flattens toward the middle. When volatility collapses after an event, the curve sharpens and out-of-the-money deltas drop away. Module 11 on vega is about the premium consequence of that shift; this is its consequence for your directional exposure.

The practical upshot is that a delta reading is a snapshot with a short shelf life. It is entirely correct for the conditions in front of you and entirely capable of being wrong by the next session. Traders who set a position once against a delta number and never look again are not managing a fixed exposure. They are holding an exposure that changes without asking them.