Chapter 6 · Derivatives on NSE: Futures and Options

6.4 The Greeks, Implied Volatility and India VIX

How an option's price responds when the market moves, time passes or nerves change: delta, gamma, theta, vega and rho as derivatives, delta hedging and gamma risk near expiry, implied volatility solved by Newton's method, the volatility smile, and how to read India VIX as the market's expected volatility.

📍 The situation

Kabir, a software engineer in Hyderabad, bought one lot of the 30-day 22,000 call on a Nifty-style index at ₹406 — the option priced in §6.3. In our simulated market, the next morning the index is up 1% at 22,220 and his call is worth about ₹531: a 1% move in the index gave him a 31% gain. A few days later the index is back where it started, nothing seems to have happened, and yet the call is worth less than he paid. Then one quiet afternoon the index does not move at all, but a news channel reports that "India VIX jumped" and his call gains ₹100. On the option chain, next to each premium, he sees a column labelled IV.

Kabir's questions are the questions of this section. How much does an option's price change when the index moves by one point, when one day passes, or when volatility changes by one percentage point? How do dealers who sell options to people like Kabir protect themselves, and why is selling options in expiry week so dangerous? What is the implied volatility on the chain, and how is it computed? And how should one read India VIX? The tool is the oldest one in calculus: each sensitivity is a derivative of the Black–Scholes price (§6.3), and Taylor's formula combines them. The Greek letters traders use for these derivatives give this section its name.

⚠️ About the numbers on this page

All index levels, option premiums, volatilities, VIX readings, interest rates, dividend yields and lot sizes on this page are simulated or illustrative — they are not market data, forecasts or recommendations, and the people are hypothetical. Unless stated otherwise we use the index at 22,000, $r=7\%$, dividend yield $q=1.2\%$ (continuous, per year), and an illustrative lot size of 50 units. Contract specifications, margins and charges change; check current NSE circulars. India VIX is computed and published by NSE; its methodology is described on the exchange's website. Nothing here is investment advice.

The Greeks: Sensitivities as Derivatives

The Black–Scholes price $V=V(S,t,\sigma,r)$ is a smooth function of its inputs. Asking "how much does $V$ change when $S$ changes a little, with everything else fixed?" is asking for the partial derivative $\partial V/\partial S$. Each input has its own derivative and its own Greek letter.

📘 Definition 6.15 — The Greeks

For an option (or portfolio) with value $V$, let $T$ be the time remaining to expiry. Then

  • delta $\Delta=\dfrac{\partial V}{\partial S}$ — change in value per ₹1 (one point) rise in the underlying;
  • gamma $\Gamma=\dfrac{\partial^2 V}{\partial S^2}=\dfrac{\partial\Delta}{\partial S}$ — change in delta per one-point rise; the curvature of $V$ in $S$;
  • theta $\Theta=\dfrac{\partial V}{\partial t}=-\dfrac{\partial V}{\partial T}$ — change in value as calendar time passes (per year; traders quote per day, $\Theta/365$);
  • vega $\mathcal V=\dfrac{\partial V}{\partial\sigma}$ — change per unit change in volatility (traders quote per volatility point, i.e. per 1 percentage point: $\mathcal V/100$);
  • rho $\rho=\dfrac{\partial V}{\partial r}$ — change per unit change in the interest rate (quoted per 1 percentage point: $\rho/100$).

The Greeks of a portfolio are the sums of the Greeks of its parts, each multiplied by the number of units held (negative for a short position). The underlying itself has $\Delta=1$ and all other Greeks zero.

📐 Theorem 6.17 (Formula) — The Black–Scholes Greeks

Under the assumptions of the Black–Scholes formula (§6.3), with $d_1,d_2$ as there, $N$ the standard normal distribution function and $\varphi(x)=\tfrac1{\sqrt{2\pi}}e^{-x^2/2}$ its density:

  • $\Delta_C=e^{-qT}N(d_1)$, $\quad\Delta_P=-e^{-qT}N(-d_1)=\Delta_C-e^{-qT}$;
  • $\Gamma=\dfrac{e^{-qT}\varphi(d_1)}{S\sigma\sqrt T}$ (the same for the call and the put);
  • $\mathcal V=Se^{-qT}\varphi(d_1)\sqrt T$ (the same for the call and the put);
  • $\Theta_C=-\dfrac{Se^{-qT}\varphi(d_1)\sigma}{2\sqrt T}$ $-\,rKe^{-rT}N(d_2)$ $+\,qSe^{-qT}N(d_1)$;
  • $\Theta_P=-\dfrac{Se^{-qT}\varphi(d_1)\sigma}{2\sqrt T}$ $+\,rKe^{-rT}N(-d_2)$ $-\,qSe^{-qT}N(-d_1)$;
  • $\rho_C=KTe^{-rT}N(d_2)$, $\quad\rho_P=-KTe^{-rT}N(-d_2)$.

For a long call: $0\lt\Delta_C\lt1$; for a long put: $-1\lt\Delta_P\lt0$; for both, $\Gamma\gt0$ and $\mathcal V\gt0$, and $\Theta$ is usually negative.

Derivation of delta (the calculus link). The key fact is the identity $$Se^{-qT}\varphi(d_1)=Ke^{-rT}\varphi(d_2),$$ which follows by writing $d_1^2-d_2^2=(d_1-d_2)(d_1+d_2)$ $=\sigma\sqrt T\cdot\dfrac{2\log(S/K)+2(r-q)T}{\sigma\sqrt T}$, so that $\varphi(d_2)/\varphi(d_1)=e^{(d_1^2-d_2^2)/2}=\dfrac{S}{K}e^{(r-q)T}$. Now differentiate $C=Se^{-qT}N(d_1)-Ke^{-rT}N(d_2)$ with the product and chain rules, using $\partial d_1/\partial S=\partial d_2/\partial S=\dfrac1{S\sigma\sqrt T}$: $$\frac{\partial C}{\partial S}=e^{-qT}N(d_1)+\Big[Se^{-qT}\varphi(d_1)-Ke^{-rT}\varphi(d_2)\Big]\frac{1}{S\sigma\sqrt T}=e^{-qT}N(d_1),$$ because the bracket is zero. Differentiating once more gives $\Gamma=e^{-qT}\varphi(d_1)\,\partial d_1/\partial S$. Vega, theta and rho come out the same way, and the same identity makes the messy terms cancel. Put–call parity $C-P=Se^{-qT}-Ke^{-rT}$ explains the put formulas: the right side has delta $e^{-qT}$ and zero gamma and vega, so $\Delta_C-\Delta_P=e^{-qT}$ while gamma and vega are equal. ∎

Notice that delta is exactly the share holding $\Delta$ of the replicating portfolio in the binomial trees of §6.3, in the limit of tiny steps: the derivative of the option value with respect to the share price is the number of shares that copies the option.

✏️ Example 1 — The Greeks of Kabir's call and the matching put
For the 30-day options with strike 22,000 on the index at 22,000 ($\sigma=14\%$, $r=7\%$, $q=1.2\%$; from §6.3: $d_1=0.13884$, $d_2=0.09870$, $C=$ ₹405.80, $P=$ ₹301.27), compute delta, gamma, theta per day, vega per volatility point and rho per 1% for the call and the put. Interpret them for one lot of 50 units.
Ingredients $T=0.082192$, $\sqrt T=0.286691$, $e^{-qT}=0.999014$, $\varphi(d_1)=0.39512$, $N(d_1)=0.55521$, $N(d_2)=0.53931$.
Delta $\Delta_C=0.999014\times0.55521=0.5547$; $\ \Delta_P=0.5547-0.9990=-0.4444$.
Gamma, vega $$\Gamma=\frac{0.999014\times0.39512}{22{,}000\times0.14\times0.286691}=0.000447\ \text{per point};$$ $\mathcal V=22{,}000\times0.999014$ $\times0.39512\times0.286691$ $=2{,}489.6$ per unit of $\sigma$, i.e. ₹24.90 per volatility point.
Theta, rho $\Theta_C=-2{,}799.7$ a year $=$ −₹7.67 a day; $\Theta_P=-1{,}532.2$ a year $=-$₹4.20 a day. $\rho_C=$ ₹9.70 and $\rho_P=-$₹8.28 per 1 percentage point of $r$.
Per lot Multiply by 50. One lot of calls behaves like $0.5547\times50\approx27.7$ units of the index: a 100-point rise adds about ₹2,773. Each day costs about ₹384 in time decay; a 1-point rise in implied volatility adds about ₹1,245. Rho is small for a 30-day option — a 1-point change in rates moves the lot by only ₹485.
📐 Theorem 6.18 (Rule) — The Greek P&L approximation

Over a short time $\delta t$ (in years) in which the underlying moves by $\delta S$ and implied volatility by $\delta\sigma$, the change in an option's (or portfolio's) value is approximately $$\delta V\approx\Delta\,\delta S+\tfrac12\Gamma\,(\delta S)^2+\Theta\,\delta t+\mathcal V\,\delta\sigma$$ (add $\rho\,\delta r$ if rates move). This is Taylor's formula to second order in $S$ and first order in the other inputs; it is accurate for moderate moves and becomes less accurate for large moves, especially near expiry where gamma itself changes quickly.

✏️ Example 2 — Explaining Kabir's next day
Using the Greeks of Example 1, estimate the change in Kabir's call when (a) one day passes and the index rises 1% (220 points) with implied volatility unchanged; (b) the index falls 220 points over a day; (c) nothing moves except implied volatility, which rises by 1 point. Compare with exact Black–Scholes repricing.
(a) Delta: $0.5547\times220=122.03$; gamma: $\tfrac12\times0.000447\times220^2=10.82$; theta: $-7.67$. Total $\approx+$₹125.17. Exact (index 22,220, 29 days): $+$₹124.90 — the call goes from ₹405.80 to ₹530.69, a 30.8% gain on a 1% move.
(b) $-122.03+10.82-7.67=-$₹118.88; exact $-$₹118.44. The gamma term is positive both ways: an option buyer gains a little extra on any move — this is what the theta cost pays for.
(c) Vega: $+$₹24.90; exact $+$₹24.91. So the "VIX jump" afternoon (implied volatility up 4 points, from 14% to 18%) is worth about $4\times24.90\approx$ ₹100 per unit, ₹5,000 per lot, with no move in the index.
Leverage The call's percentage sensitivity is $\Delta\cdot S/C=0.5547\times22{,}000/405.80\approx30$: roughly 30 times the index's percentage move. This "option elasticity" is why options feel exciting — and why they lose money just as fast.

Delta Hedging, Gamma Risk and Theta Decay

When Kabir buys a call, someone sells it — often a dealer or market maker who does not want to bet on the market's direction. The dealer neutralises the direction risk by holding the replicating portfolio of §6.3: about $\Delta$ units of the underlying for every option sold. In practice an index option dealer hedges with index futures (§6.1), whose delta per unit is very close to 1 for short maturities.

📘 Definition 6.16 — Delta-neutral position and delta hedging

A position is delta-neutral if its total delta is zero, so that small moves in the underlying leave its value (to first order) unchanged. Delta hedging means holding $-\Delta$ units of the underlying (or futures) for each option held — buying $\Delta$ units per option sold — and rebalancing the hedge as $\Delta$ changes with the price and with time. Gamma measures how quickly the hedge goes out of date: a position with large $|\Gamma|$ needs frequent rebalancing.

✏️ Example 3 — A dealer hedges 20 lots of calls
A dealer has sold 20 lots (1,000 units) of the 30-day 22,000 call of Example 1. (a) How many lots of index futures should she buy to be delta-neutral? (b) The next day the index is at 22,300. Use gamma to estimate the new delta, compare with the exact value, and say how the hedge must change. (c) What was her overnight P&L with the hedge she held?
(a) Position delta: $-1{,}000\times0.5547=-554.7$ units. She needs $+554.7$ units: $554.7/50=11.09$ lots. She buys 11 lots (550 units), leaving a small delta of $-4.7$ units. (A futures unit has delta $e^{(r-q)T}\approx1.005$ against the index — close enough to 1 for this purpose.)
(b) Gamma estimate: $0.5547+0.000447\times300=0.6888$. Exact (22,300, 29 days): $0.6836$. The short calls now have delta $-683.6$ units, so she should hold about 684 units, i.e. buy about 134 units $\approx2.7$ more lots (in practice 3).
(c) Exact repricing: the call rose from ₹405.80 to ₹584.08, so the short calls lost $1{,}000\times178.28=$ ₹1,78,276, while the 550 futures units gained about $550\times300=$ ₹1,65,000 (taking the futures to move point for point with the index). Net: about $-$₹13,276. The Greeks explain it: gamma loss $\tfrac12\times0.000447\times300^2\times1{,}000=20{,}116$, minus one day's theta earned ($7.67\times1{,}000=7{,}670$), plus the unhedged residual delta ($-4.7\times300\approx-1{,}400$): about $-$₹13,800.
Lesson Delta hedging removes the first-order direction risk, not all risk. A short-option position is short gamma: every big move, up or down, costs money, and the daily theta income is the compensation. Whether the dealer makes or loses money over the month depends on how large the moves are compared with what the option price assumed.

That last sentence can be made exact. In the Black–Scholes model, the Greeks are tied together by one equation (it is the Black–Scholes partial differential equation, which we state without derivation; you can check it numerically from Theorem 6.17).

📐 Theorem 6.19 (Rule) — The theta–gamma trade-off

For any European option in the Black–Scholes model, $$\Theta+(r-q)S\Delta+\tfrac12\sigma^2S^2\Gamma=rV.$$ Consequently, for a delta-hedged long option over a short time $\delta t$, ignoring interest on the small cash balances, $$\text{P\&L}\approx\tfrac12\Gamma S^2\left[\Big(\frac{\delta S}{S}\Big)^2-\sigma^2\delta t\right],$$ where $\sigma$ is the implied volatility used to price and hedge. The hedged long option gains when the actual move is larger than $\sigma\sqrt{\delta t}$ in percentage terms and loses (theta) when it is smaller; the short side has the opposite P&L. The break-even move is about $\pm S\sigma\sqrt{\delta t}$; for one day, $\pm S\sigma/\sqrt{365}$.

Why. In Theorem 6.18 with $\delta\sigma=0$, a hedged position has delta zero, so $\delta V_{\text{hedged}}\approx\tfrac12\Gamma(\delta S)^2+\Theta\,\delta t$. When $r$ and $q$ are small the first equation gives $\Theta\approx-\tfrac12\sigma^2S^2\Gamma$; substituting gives the bracket. Squared moves are compared with the variance the option price "assumed": this is why an option is, at heart, a contract on volatility. ∎

📐 Theorem 6.20 (Rule) — At-the-money options near expiry

For an option struck at the forward price, $K=Se^{(r-q)T}$ (for short expiries and small $r,q$ this is essentially at the money), with $\sigma\sqrt T$ small:

  1. $C\approx P\approx\dfrac{1}{\sqrt{2\pi}}S\sigma\sqrt T\approx0.4\,S\sigma\sqrt T$, so a straddle (call + put) costs about $0.8\,S\sigma\sqrt T$;
  2. $\Gamma\approx\dfrac{0.4}{S\sigma\sqrt T}$, which grows like $1/\sqrt T$ as expiry approaches;
  3. theta per year $\Theta\approx-\dfrac{0.2\,S\sigma}{\sqrt T}$: time decay accelerates as $T\to0$, and half of the value is lost in the last quarter of the time;
  4. $\mathcal V\approx0.4\,S\sqrt T$, which shrinks near expiry.

Why. With $K$ at the forward, $d_1=\tfrac12\sigma\sqrt T$ and $d_2=-\tfrac12\sigma\sqrt T$, so $C=Se^{-qT}\big[N(\tfrac12\sigma\sqrt T)-N(-\tfrac12\sigma\sqrt T)\big]$. For small $x$, $N(x)-N(-x)\approx2x\varphi(0)=2x/\sqrt{2\pi}$, giving (1) since $1/\sqrt{2\pi}=0.3989$. Gamma and vega follow from Theorem 6.17 with $\varphi(d_1)\approx\varphi(0)$, and $\Theta=-\partial C/\partial T\approx-0.4S\sigma\cdot\tfrac{1}{2\sqrt T}$ from (1). Since value $\propto\sqrt T$, a quarter of the time left means $\sqrt{1/4}=\tfrac12$ of the value left. ∎

✏️ Example 4 — Gamma and theta in expiry week
For the at-the-money 22,000 call ($\sigma=14\%$, usual $r$, $q$), compute value, delta, gamma, theta per day and vega per point with 30 days, 7 days and 1 day to expiry. What happens to a dealer's hedge in the last days?
Table
days leftvaluedeltagammatheta/dayvega/pt
30405.800.5550.00045−7.6724.90
7182.530.5270.00093−13.9312.12
166.070.5100.00247−33.924.59
Check the rule $0.4/(S\sigma\sqrt T)$ gives gammas $0.00045$, $0.00094$, $0.00248$ — almost exact. From 30 days to 1 day, $\sqrt{30}\approx5.5$: gamma multiplies by 5.5, theta per day by 4.4, while vega falls to a fifth.
Gamma risk With 1 day left, a 100-point move changes delta by $0.247$ — a quarter of a unit per option. A dealer short 20 lots would see her hedge need about 5 lots of futures more or less after an ordinary 100-point move. Near the strike in the final hours, delta swings between almost 0 and almost 1: this is the gamma risk near expiry that makes short options in expiry week dangerous.
Theta With one day left the option is losing value at ₹33.92 a day at the start of the day, and if the index stays at the strike all of its ₹66.07 is gone by the end of it. Sellers collect that decay, but on a single sharp move the gamma loss can wipe out many days of theta.
Interactive 1 — Greeks Explorer: How Delta, Gamma, Theta and Vega Change Towards Expiry

A 22,000-strike option on the simulated index ($r=7\%$, $q=1.2\%$). The blue curve is the chosen quantity against the index level with the days to expiry you set; the faint curves show the same quantity with 30, 7 and 1 day(s) left. Readouts are at the strike. Press Play to count down from 60 days to 1 and watch delta turn into a step, gamma into a spike and theta into a cliff. (Simulated, illustrative inputs.)

3014%
value—
delta—
gamma (per point)—
theta (₹/day)—
vega (₹/vol pt)—
✏️ Example 5 — Theta decay is not linear
To isolate the effect of time, take $r=q=0$ and an at-the-money option on the index at 22,000 with $\sigma=14\%$. Find its value and its theta per day with 20, 10, 5, 4 and 1 day(s) left, compare with $0.4S\sigma\sqrt T$, and find what fraction of the 4-day value disappears in the last day.
Values
daysBlack–Scholes$0.4S\sigma\sqrt T$theta/day
20287.61288.39−7.19
10203.38203.92−10.17
5143.81144.19−14.38
4128.63128.97−16.08
164.3264.49−32.16
Rule check The approximation is within 0.3%. Halving the time from 20 to 10 days removes only 29% of the value ($1-1/\sqrt2$); from 4 days to 1 day (a quarter of the time) the value halves.
Last day Of ₹128.63 with 4 days left, ₹64.32 remains with 1 day left and all of it goes on the last day — the final day takes half the 4-day value, as much as the three days before it together.
Shape Theta per day doubles each time the remaining time is divided by four ($1/\sqrt T$). A buyer of a short-dated option is fighting an accelerating clock; a seller is collecting it — while carrying the growing gamma of Example 4.
Interactive 2 — Delta-Hedging Simulator: What Does a Hedged Option Seller Actually Earn?

A dealer sells one lot (50 units) of a 30-day 22,000 call at the Black–Scholes price with the implied volatility you set, and delta-hedges with the index at the chosen frequency ($r=7\%$, no dividends). The index then follows a simulated GBM path whose realised volatility you also set. Blue: hedged P&L (marked to market); red: unhedged P&L (premium kept in the bank). The bottom readouts summarise 400 simulated paths. Try realised = implied, then realised above and below implied, then change the rebalancing frequency. (Simulated data, not a backtest.)

14%14%
day · index—
premium received—
units held (delta)—
hedged P&L—
unhedged P&L—
400 paths: hedged mean ± SD—
400 paths: unhedged SD—

Implied Volatility and the Smile

Five of the six Black–Scholes inputs are known. The sixth, volatility, is about the future. Traders therefore turn the formula around: instead of feeding in a volatility to get a price, they take the market price and ask which volatility would produce it. That number is the IV column on an option chain.

📘 Definition 6.17 — Implied volatility

The implied volatility (IV) of an option with market price $V_{\text{mkt}}$ is the value $\sigma_{\text{imp}}\gt0$ for which the Black–Scholes price equals the market price, all other inputs ($S,K,T,r,q$) being fixed: $\ V_{\text{BS}}(\sigma_{\text{imp}})=V_{\text{mkt}}$. It is quoted as an annual percentage, like $\sigma$.

📐 Theorem 6.21 — Implied volatility exists and is unique

For a European call, if the market price satisfies the no-arbitrage bounds $$\max\big(Se^{-qT}-Ke^{-rT},0\big)\lt C_{\text{mkt}}\lt Se^{-qT},$$ then there is exactly one $\sigma_{\text{imp}}\gt0$ with $C_{\text{BS}}(\sigma_{\text{imp}})=C_{\text{mkt}}$. The same holds for a put with bounds $\max(Ke^{-rT}-Se^{-qT},0)\lt P_{\text{mkt}}\lt Ke^{-rT}$, and by put–call parity a call and a put with the same strike and expiry have the same implied volatility. A price outside the bounds has no implied volatility — it signals an arbitrage or a data error.

Proof. As a function of $\sigma$ on $(0,\infty)$, $C_{\text{BS}}$ is continuous, and its derivative is vega, $\mathcal V=Se^{-qT}\varphi(d_1)\sqrt T\gt0$, so it is strictly increasing — it takes each value at most once. As $\sigma\to0^+$ the share becomes riskless and $C_{\text{BS}}\to\max(Se^{-qT}-Ke^{-rT},0)$; as $\sigma\to\infty$, $d_1\to\infty$ and $d_2\to-\infty$, so $C_{\text{BS}}\to Se^{-qT}$. By the intermediate value theorem every price strictly between these limits is attained. For the put, $P_{\text{BS}}(\sigma)=C_{\text{BS}}(\sigma)-Se^{-qT}+Ke^{-rT}$, so a put price consistent with parity gives the same equation for $\sigma$. ∎

🧭 How to compute implied volatility (Newton's method)
  1. Check the bounds of Theorem 6.21. If the price is outside them, stop: there is no IV.
  2. Start from a sensible guess: near the money use $\sigma_0\approx\dfrac{V_{\text{mkt}}}{0.4\,S\sqrt T}$ (Theorem 6.20), or simply 20%.
  3. Iterate Newton's method on $f(\sigma)=V_{\text{BS}}(\sigma)-V_{\text{mkt}}$, whose derivative is vega: $$\sigma_{k+1}=\sigma_k-\frac{V_{\text{BS}}(\sigma_k)-V_{\text{mkt}}}{\mathcal V(\sigma_k)}.$$
  4. Stop when the price error is below, say, ₹0.01. Near the money this takes 2–4 steps; convergence is quadratic (the number of correct digits roughly doubles each step).
  5. Safeguard: far from the money vega is tiny and Newton can overshoot. Then use bisection: keep an interval $[\sigma_{\text{lo}},\sigma_{\text{hi}}]$ with prices on either side of $V_{\text{mkt}}$ and halve it until it is narrow enough. Spreadsheet users can use Goal Seek.
✏️ Example 6 — Backing out the IV of a quoted call
On a nervous day the 30-day 22,000 call (index at 22,000, $r=7\%$, $q=1.2\%$) trades at ₹450 instead of ₹405.80. Find its implied volatility with Newton's method, starting from $\sigma_0=14\%$.
Bounds $Se^{-qT}-Ke^{-rT}=104.52\lt450\lt21{,}978$ ✓, so a unique IV exists.
Iterate
$k$$\sigma_k$$C_{\text{BS}}(\sigma_k)$errorvega
014.0000%405.797−44.2032,489.6
115.7755%450.034+0.0342,493.2
215.7741%450.0000.0002,493.2
Step 1: $\sigma_1=0.14+44.203/2{,}489.6=0.157755$.
Answer $\sigma_{\text{imp}}\approx15.77\%$. Two Newton steps suffice because near the money the price is almost linear in $\sigma$ (vega hardly changes).
Interpret The market is pricing the next 30 days as if annualised volatility will be about 15.8% rather than 14%. The extra ₹44.20 is almost exactly $1.77\times24.9$ — vega times the change in volatility points.

The smile. If the Black–Scholes model were exactly right, every option on the same index and expiry would have the same implied volatility — the volatility of the index. In real markets they do not. Plot implied volatility against strike and the result is usually a curve: higher at low strikes, lower near and above the money for equity indices (a downward skew or "smirk"), or rising on both sides (a smile), as is common in currency options.

📘 Definition 6.18 — Volatility smile and skew

The volatility smile is the graph of implied volatility against strike (or moneyness, e.g. $\log(K/F)$) for options with the same expiry. A skew is a smile that slopes, typically downward for equity index options: out-of-the-money puts (low strikes) have higher implied volatility than out-of-the-money calls (high strikes). The smile shows where the constant-volatility assumption of Black–Scholes fails.

Why does the skew exist? Three explanations are usually given, and they reinforce one another. (1) Real return distributions have fat tails and sharp falls are more common than sharp rises (§3.1), so deep out-of-the-money puts pay off more often than a lognormal model says. (2) Investors buy index puts as insurance for their portfolios, and dealers who sell that insurance charge for the crash risk they take on. (3) Volatility tends to rise when prices fall, so low-strike scenarios are also high-volatility scenarios. The smile is a description of market prices, not a law; its shape changes from day to day and from expiry to expiry.

✏️ Example 7 — Reading a smile off a simulated option chain
In our simulated market (index 22,000, 30 days, usual $r$, $q$), the out-of-the-money and at-the-money options trade at: 21,000 put ₹80.00; 21,500 put ₹155.65; 22,000 call ₹411.30; 22,500 call ₹176.65; 23,000 call ₹56.95. Compute the implied volatility of each and describe the shape. How expensive is the 21,000 put compared with its price at 14%?
IVs Solving each price for $\sigma$ (Newton or bisection; check the bounds first):
strikeoptionpriceIV
21,000put80.0017.10%
21,500put155.6515.48%
22,000call411.3014.22%
22,500call176.6513.30%
23,000call56.9512.69%
Shape Implied volatility falls steadily from 17.1% at the low strike to 12.7% at the high strike: a downward skew, steepest on the put side — the typical shape for an equity index.
Price effect At a flat 14% the 21,000 put would cost ₹41.02; at its 17.1% IV it costs ₹80 — almost twice as much. Crash insurance is priced well above what a constant-volatility model suggests, and the 23,000 call (₹56.95 versus ₹76.58 at 14%) is cheaper.
Use Because IV varies by strike, traders compare options by their IVs rather than their rupee prices, and risk systems use the whole smile, not one number. Which IV should "the" volatility of the index be? That is exactly what India VIX tries to answer.

India VIX: Reading the Market's Expected Volatility

One number summarises the whole smile: a volatility index. NSE publishes India VIX, computed during the trading day from the prices of Nifty 50 options. News reports call it a "fear gauge". Mathematically it is something more precise.

📘 Definition 6.19 — India VIX

India VIX is a volatility index computed by NSE from the order-book prices of out-of-the-money Nifty 50 call and put options across a range of strikes for the near and next expiries, combined so as to measure the market's expected (risk-neutral) volatility of the Nifty over the next 30 calendar days. It is expressed as an annualised percentage: a reading of 15 means an expected annualised volatility of about 15%. (NSE's methodology is adapted from the one used for the CBOE's VIX index; see the exchange's documentation for details.)

Two features matter for reading it. VIX uses many strikes, weighted so that the result estimates the expected variance over 30 days rather than the IV of any single option — so it is usually close to, but slightly above, the at-the-money IV when the skew is downward. And it is a risk-neutral expectation: it includes the premium that option buyers pay for insurance, so it is not a pure forecast of realised volatility.

📐 Theorem 6.22 (Rule) — Turning a VIX reading into an expected move

If India VIX reads $v$ (in percent), then under the lognormal model the implied one-standard-deviation move of the index over $h$ calendar days is about $$\pm\,v\%\times\sqrt{\frac{h}{365}}.$$ In particular, over the 30 days VIX measures it is $v\sqrt{30/365}\approx0.287\,v$ percent; a typical calendar day is $v/\sqrt{365}\approx v/19.1$ percent; and if all the variance is assigned to trading days (252 a year by convention), a typical trading day is $v/\sqrt{252}\approx v/15.9$ percent. The range $S\,e^{\pm v\sqrt{h/365}/100}$ contains the index at the horizon with risk-neutral probability of roughly 68% — less in practice when tails are fat.

✏️ Example 8 — What does "India VIX at 15" mean for the next month?
Suppose India VIX reads 15 with the index at 22,000. (a) Find the implied 1-SD move over 30 days, and the corresponding range. (b) Find the typical daily move (calendar-day and trading-day versions). (c) Estimate the price of a 7-day at-the-money straddle at 15% volatility and compare with its expected absolute move.
(a) $15\%\times\sqrt{30/365}=4.30\%$, i.e. about 946 points. Range: $22{,}000e^{\pm0.0430}$, from about 21,074 to 22,967. Roughly a 2-in-3 chance (under the model) of ending the month inside it, and a 1-in-3 chance outside — in either direction.
(b) Calendar day: $15/19.1\approx0.79\%$ (about 173 points). Trading day: $15/15.87\approx0.94\%$ (about 208 points).
(c) Theorem 6.20: $0.8\times22{,}000\times0.15\times\sqrt{7/365}\approx$ ₹365.6 per unit (the exact Black–Scholes straddle is ₹364.9). This is also the expected absolute move $E|S_T-S|\approx\sqrt{2/\pi}\,S\sigma\sqrt T=$ 364.6 points: an at-the-money straddle costs about what the market expects the index to move.
Caution VIX says how much movement the option market is pricing, not which way. A VIX of 15 is consistent with the index rising 946 points, falling 946 points, or staying flat.

How VIX behaves (descriptively). Volatility indices around the world tend to rise sharply when markets fall and drift lower in calm, rising markets, so they are negatively correlated with index returns. They also tend to revert towards a long-run level rather than trend forever — high readings rarely persist for long, and very low readings do not last either. And implied volatility has, on average across many markets and periods, tended to exceed the volatility that was subsequently realised: the difference is the insurance premium earned by option sellers, paid for by occasional heavy losses. None of these tendencies is a trading rule, and none says what the index will do next.

⚠️ Common mistakes and myths
  1. "Delta is the probability that my option expires in the money." Only roughly. That (risk-neutral) probability is $N(d_2)$; delta is $e^{-qT}N(d_1)$, slightly larger for a call. For Kabir's call: delta 0.555, $N(d_2)=0.539$.
  2. "Option selling is steady income because theta is on my side." Theta is the price of being short gamma (Theorem 6.19). A short straddle earns small amounts on quiet days and can lose many days' income on one sharp move — see the Market case. An option seller's profit is capped at the premium; the loss is not.
  3. "Hedge once and you are safe." Delta changes with the price and with time; near expiry it can swing from 0 to 1 within a few points (Example 4). An unrebalanced hedge is barely a hedge (Interactive 2).
  4. "Theta decays evenly." It accelerates like $1/\sqrt T$: half the at-the-money value goes in the last quarter of the time (Example 5).
  5. Vega units. Vega from the formula is per unit of $\sigma$ (per 100 volatility points). Divide by 100 for "per vol point". Likewise theta from the formula is per year: divide by 365 for per day.
  6. "Implied volatility is a forecast that will come true." It is the volatility that makes the model match the price, including a risk premium. Realised volatility can be far above or below it.
  7. "Low VIX means the market will rise; high VIX means it will crash." VIX measures the size of expected moves, not their direction. A low VIX means options are cheap, not that the market is safe.
  8. "One IV for all strikes." The smile means each strike has its own IV (Example 7). Using the at-the-money IV to price a deep out-of-the-money put underprices it.
  9. "A ₹2 weekly option is a cheap lottery ticket." Its price reflects a tiny probability of paying off. Buying many of them is a reliable way to lose the premiums; selling many is a way to earn small amounts until one large move.
📈 Market case — The straddle seller's expiry week

Five days before expiry, Meenal (hypothetical) sells one lot of the 22,000 at-the-money straddle — a call and a put — on the simulated index at 22,000, at an implied volatility of 14% (usual $r$, $q$; lot 50). She wants to understand exactly what she has sold.

  1. Premium. Call ₹152.63 + put ₹135.17 = ₹287.80 per unit, ₹14,390 per lot. (Theorem 6.20's quick estimate $0.8S\sigma\sqrt T=$ ₹288.4.) Break-even at expiry: 21,712.20 and 22,287.80 — a move of 1.31% either way.
  2. Greeks of her short position (per unit): delta $-0.045$ (almost neutral), gamma $-0.00221$ per point, theta $+$₹28.81 a day, vega $-$₹20.51 per vol point.
  3. Break-even daily move (Theorem 6.19): $S\sigma/\sqrt{365}=22{,}000\times0.14/19.105\approx161$ points (0.73%). Smaller daily moves favour her; larger ones hurt.
  4. Day 1 scenarios (exact repricing with 4 days left, unhedged): index flat, $+$₹1,521 per lot; a 0.5% rise, $+$₹563; a 1% rise, $-$₹1,763; a 1.5% rise, $-$₹5,220; a 2% rise, $-$₹9,514 (falls of the same sizes give $+$₹997, $-$₹955, $-$₹4,136 and $-$₹8,256 — slightly better, because her delta is slightly negative). If instead the index stays flat but implied volatility jumps to 18%, $-$₹2,150.
  5. If nothing ever moves. The straddle's value with 5, 4, 3, 2, 1 day(s) left is 287.80, 257.39, 222.88, 181.95, 128.65 and then 0: she earns ₹30, ₹35, ₹41, ₹53 and ₹129 per unit on successive days — accelerating theta.
  6. A single 3% move by expiry (to 22,660) costs her the payoff of 660 points against 287.80 received: $-$₹372.20 per unit, $-$₹18,610 per lot — more than the entire maximum profit of ₹14,390.

What she learns. Selling the straddle is a bet that the index will move less than the 0.73% a day the option market is pricing. The maximum profit is capped at the premium; the loss grows with the square of the move until expiry, then linearly — without limit on the call side. Margins required by the exchange for such positions are large for this reason (check the current rules). Delta hedging, as in Interactive 2, converts the bet into a cleaner one on realised versus implied volatility, but gamma near expiry makes that hedge expensive to maintain. The mathematics does not say whether to sell; it says precisely what is being sold.

Exercises

All inputs are illustrative. Unless stated otherwise: rates continuously compounded per year, theta per calendar day ($/365$), vega per volatility point ($/100$), index at 22,000 with $r=7\%$, $q=1.2\%$, and lot size 50.

I
Greeks by formula. A share of Krishna Cement Ltd (hypothetical, no dividends) is at ₹1,000; consider 3-month options with strike ₹1,000, $r=7\%$, $\sigma=25\%$.
  • 1 Compute $d_1$, $d_2$, the call delta and the put delta.
  • 2 Compute gamma. Use it to estimate the call delta after a ₹20 rise, and compare with the exact value.
  • 3 Compute vega per volatility point and estimate the change in the call price if implied volatility rises from 25% to 27%.
  • 4 Compute theta per day for the call and the put. Why is the put's theta smaller in size?
  • 5 Compute rho per 1% for the call and the put.
  • 6 Use Theorem 6.18 to estimate the change in the call price when, over one day, the share rises ₹15 and implied volatility falls 1 point. Compare with exact repricing.
  • 7 Verify $\Delta_C-\Delta_P=1$ here and explain, using put–call parity, why the call and put have the same gamma and vega.
1 $\sigma\sqrt T=0.125$; $d_1=(0.07+0.03125)\times0.25/0.125=0.2025$, $d_2=0.0775$; $\Delta_C=N(0.2025)=0.5802$, $\Delta_P=-0.4198$.
2 $\Gamma=\varphi(0.2025)/(1000\times0.125)$ $=0.39085/125=0.003127$. New delta $\approx0.5802+20\times0.003127=0.6428$; exact $0.6409$.
3 $\mathcal V=1000\times0.39085\times0.5=195.42$, i.e. ₹1.954 per point; $+2$ points $\Rightarrow\approx+$₹3.91 (exact $+$₹3.91). The call is worth ₹58.56 at 25%.
4 $\Theta_C=-134.23$ a year $=-$₹0.368 a day; $\Theta_P=-65.44$ a year $=-$₹0.179 a day. Both lose the same volatility time value ($-\tfrac12S\varphi\sigma/\sqrt T$), but the put gains from time passing through interest: the strike it delivers comes closer, worth $+rKe^{-rT}N(-d_2)$, while the call pays the strike and loses $rKe^{-rT}N(d_2)$.
5 $\rho_C=1000\times0.25\times e^{-0.0175}$ $\times N(0.0775)=130.42$ → ₹1.304 per 1%; $\rho_P=-115.24$ → $-$₹1.152 per 1%.
6 $0.5802\times15+\tfrac12\times0.003127\times225$ $-\,0.368-1.954$ $=8.703+0.352-0.368-1.954\approx$ ₹6.73; exact $+$₹6.77.
7 $0.5802-(-0.4198)=1$ ✓ ($q=0$). Parity: $C-P=S-Ke^{-rT}$. The right side is linear in $S$ with slope 1 and does not involve $\sigma$, so $\Delta_C-\Delta_P=1$, $\Gamma_C-\Gamma_P=0$ and $\mathcal V_C-\mathcal V_P=0$.
II
Hedging, gamma and theta (index options; 30-day ATM call delta 0.5547 as in Example 1).
  • 1 A trader is short 10 lots of the 30-day 22,000 call. How many futures lots make the position delta-neutral?
  • 2 A portfolio is long 5 lots of calls with delta 0.60 and short 8 lots of puts with delta −0.30. Find its delta in units and the futures trade that neutralises it.
  • 3 A dealer is short 10 lots of the 7-day 22,000 call (gamma 0.000933 per point) and delta-hedged. Estimate her loss from an immediate 300-point move in either direction.
  • 4 At 14% implied volatility, what daily move is the break-even for a delta-hedged at-the-money option?
  • 5 Using Theorem 6.20, by what factor is the at-the-money gamma with 1 day left larger than with 25 days left? Check with the formula.
  • 6 With $r=q=0$ and $\sigma=14\%$, find the value of the at-the-money call with 16 and with 4 days left. What fraction of the 16-day value is lost in the last 4 days?
  • 7 Explain the trade-off a dealer faces when choosing how often to rebalance a delta hedge.
1 Delta $=-500\times0.5547=-277.3$ units; buy $277.3/50=5.55$ lots — 5 or 6 lots in practice, with a small residual delta.
2 Calls: $+250\times0.60=+150$; short puts: $-400\times(-0.30)=+120$. Total $+270$ units $=5.4$ lots: sell 5 lots of futures (residual $+20$ units).
3 $\tfrac12\times0.000933\times300^2\times500\approx$ ₹20,990 (Theorem 6.18; exact repricing gives about ₹19,790 for a rise and ₹20,560 for a fall, because gamma falls away from the strike). Up or down, the hedged short loses.
4 $S\sigma/\sqrt{365}=22{,}000\times0.14/19.105\approx161$ points, about 0.73%. (With $r$ and $q$ included the exact theta-gamma balance is a little different, about 173 points for the 7-day option, but the rule of thumb is what traders use.)
5 $\Gamma\propto1/\sqrt T$: $\sqrt{25}=5$. Formula: $0.002474/0.000491=5.04$.
6 16 days: ₹257.25; 4 days: ₹128.63. Exactly half the value ($\sqrt{4/16}=\tfrac12$) disappears in the last quarter of the time.
7 Frequent rebalancing keeps the position close to delta-neutral, so the hedged P&L tracks the realised-versus-implied variance closely, but every trade costs brokerage, taxes and bid–ask spread. Infrequent rebalancing saves costs but leaves large unhedged moves (and the P&L becomes very noisy, especially when gamma is high near expiry). Interactive 2 shows the spread of outcomes shrinking roughly like $1/\sqrt{\text{number of rebalances}}$.
III
Implied volatility and the smile (30-day options on the index unless stated).
  • 1 The 22,000 call trades at ₹480. Find its implied volatility by Newton's method from $\sigma_0=14\%$ (vega at 14% is 2,489.6).
  • 2 Compute the quick first guess $\sigma_0=C/(0.4S\sqrt T)$ for part 1. Why is it too high? Improve it by first subtracting half of $Se^{-qT}-Ke^{-rT}=104.52$ from the price.
  • 3 Explain why Newton's method for implied volatility cannot find two different answers for the same price.
  • 4 A data feed shows the 22,000 call at ₹100. What is its implied volatility?
  • 5 The 22,000 call has an IV of 15%. Without computing, what is the IV of the 22,000 put, if put–call parity holds? Why?
  • 6 In Example 7's smile, the 21,000 put has IV 17.1% while the at-the-money IV is about 14%. Compute the 21,000 put's price at 18% and at 14%, and explain why a risk manager should not value it at 14%.
Check: an implied-volatility solver (Newton's method with a bisection safeguard)
1 $\sigma_1=0.14+(480-405.797)/2{,}489.6$ $=0.169805$; $C_{\text{BS}}(\sigma_1)=480.089$, vega 2,495.0; $\sigma_2=0.169805-0.089/2{,}495.0=0.169770$. IV $\approx16.98\%$.
2 $480/(0.4\times22{,}000\times0.286691)=19.03\%$ — too high, because the at-the-money-spot call is partly in the money relative to the forward (worth extra because $r\gt q$). Corrected: $(480-52.26)/2{,}522.9=16.95\%$, within 0.03 points of the answer.
3 The Black–Scholes price is strictly increasing in $\sigma$ (vega $\gt0$), so each price corresponds to at most one $\sigma$ (Theorem 6.21). Newton may converge slowly or overshoot when vega is small, but it cannot find a second root because there is none.
4 None: the lower bound is $Se^{-qT}-Ke^{-rT}=104.52\gt100$. Either the quote is stale or wrong, or there is an arbitrage (buy the call, sell the index forward). The solver above reports this.
5 Also 15%. Parity $C-P=Se^{-qT}-Ke^{-rT}$ does not involve $\sigma$, so if $C=C_{\text{BS}}(15\%)$ then $P=P_{\text{BS}}(15\%)$.
6 At 18%: ₹93.17; at 14%: ₹41.02. Valuing at 14% would understate the cost of replacing (or the risk of being short) the put by more than half; the market charges for crash risk at low strikes, so positions must be marked at their own strike's IV.
IV
India VIX and decisions.
  • 1 India VIX reads 12. Find the implied 1-SD move of the index over the next 30 days, in percent and in points.
  • 2 India VIX reads 20. Find the typical calendar-day and trading-day moves, in percent and in points.
  • 3 Overnight, at-the-money implied volatility rises from 14% to 18% while the index is unchanged. Estimate the change in value of one lot of the 30-day at-the-money call, and compare with exact repricing.
  • 4 With VIX at 15 and a lognormal model with zero log drift, estimate the probability that the index moves by more than 5% (either way) in 30 days. Would fat tails make the true probability larger or smaller?
  • 5 Over a month, the index's realised volatility turned out to be 11% while the at-the-money call had been bought at an IV of 15% and delta-hedged daily. Who made money — the hedged buyer or the hedged seller? Estimate the size per unit with Theorem 6.19 ($\Gamma=0.000417$ per point for the 30-day option at 15%).
  • 6 A friend says "VIX is at a five-year low, so the market can only go up." Give two reasons why this does not follow.
1 $12\times\sqrt{30/365}=3.44\%$, about 757 points.
2 Calendar day: $20/\sqrt{365}=1.05\%$ (230 points); trading day: $20/\sqrt{252}=1.26\%$ (277 points).
3 Vega ₹24.90 per point $\times4\approx$ ₹99.6 per unit, ₹4,980 per lot; exact $+$₹99.73 per unit.
4 $\sigma\sqrt T=0.15\times0.2867=0.0430$; $P(|\log(S_T/S)|\gt\log1.05)$ $\approx2\big(1-N(0.05/0.0430)\big)=2\big(1-N(1.16)\big)\approx24.5\%$. (Using $\log1.05$ and $\log0.95$ separately gives $12.8\%+11.7\%$, the same 24.5%.) Fat tails make large moves more likely than the normal model says, so the true probability of a very large move is likely higher.
5 Realised (11%) below implied (15%): the hedged seller gains. Size: $\tfrac12\Gamma S^2(\sigma_R^2-\sigma_I^2)T$ $=\tfrac12\times0.000417\times22{,}000^2$ $\times(0.0121-0.0225)\times\tfrac{30}{365}$ $\approx-$₹86 per unit for the buyer, so about $+$₹86 per unit ($+$₹4,300 per lot) for the seller, before costs. (A cruder vega estimate, $4\times$ ₹25, gives a similar size.)
6 (i) VIX measures the expected size of moves, not their direction; a low VIX is consistent with a fall. (ii) Low volatility tends not to persist — volatility indices tend to revert towards a long-run level, and calm periods can end abruptly. Low VIX means option insurance is cheap, nothing more.
🎯 Practice corner NISM-styleCase studyInterview

E1. NISM-style The delta of a deep in-the-money call option on a non-dividend-paying stock is closest to (A) 0 (B) 0.5 (C) 1 (D) −1.

Idea Deep in the money, $d_1$ is large and $N(d_1)\to1$: the call moves almost one-for-one with the stock. (B) is at the money; (A) is deep out of the money; (D) is a deep in-the-money put.
Answer (C).

E2. NISM-style The sensitivity of an option's premium to a change in the volatility of the underlying is measured by (A) delta (B) gamma (C) vega (D) theta.

Idea Vega $=\partial V/\partial\sigma$. Delta and gamma concern the underlying's price, theta the passage of time.
Answer (C).

E3. NISM-style For an at-the-money option, as expiry approaches (other things equal), (A) gamma falls and theta becomes less negative (B) gamma rises and theta becomes more negative (C) gamma and theta are unchanged (D) vega rises.

Idea Theorem 6.20: $\Gamma\propto1/\sqrt T$ and $\Theta\propto-1/\sqrt T$ both grow in size, while vega $\propto\sqrt T$ shrinks (Example 4).
Answer (B).

E4. Case study A dealer is short 20 lots (1,000 units) of the 7-day 22,000 call ($\sigma=14\%$; per unit: delta 0.5266, gamma 0.000933, theta −₹13.93 a day, vega ₹12.12 per point). (a) Give the Greeks of her position. (b) How many futures lots hedge it? (c) After hedging, the index rises 2% (440 points) by the next day. Estimate her P&L with the Greeks; the exact figure is a loss of ₹71,069 — why is the Taylor estimate worse here? (d) What daily move would leave her hedged position roughly flat?

(a) Delta $-526.6$ units; gamma $-0.933$ per point; theta $+$₹13,930 a day; vega $-$₹12,124 per volatility point.
(b) $526.6/50=10.5$ lots: buy 10 or 11 lots of futures.
(c) Gamma loss $\tfrac12\times0.933\times440^2\approx$ ₹90,314, theta gain ₹13,930: net about $-$₹76,384 (ignoring the small residual delta). The exact loss ₹71,069 is smaller because for a 440-point move gamma does not stay at its at-the-money peak — the option's curvature falls off away from the strike, and Taylor's second-order formula overstates large moves near expiry.
(d) About $S\sigma/\sqrt{365}\approx161$ points (0.73%) a day (Theorem 6.19).

E5. Case study The 30-day 22,000 call trades at ₹430 while India VIX reads 15.5. (a) Find the call's implied volatility. (b) Find the put price implied by parity and the cost of the at-the-money straddle; compare with $0.8S\sigma\sqrt T$. (c) Give the 1-SD 30-day range implied by the option's IV and by VIX. (d) Suggest why VIX might be a little above the at-the-money IV.

(a) Newton from 14%: $\sigma_1=0.14+(430-405.80)/2{,}489.6$ $=0.14972$; one more step gives IV $\approx14.97\%$.
(b) $P=430-104.52=$ ₹325.48; straddle ₹755.48. Estimate: $0.8\times22{,}000\times0.1497\times0.28669\approx$ ₹755.4 — very close.
(c) From the IV: $22{,}000\times0.1497\times0.28669\approx944$ points (4.29%). From VIX: $22{,}000\times0.155\times0.28669\approx978$ points (4.44%).
(d) VIX averages over many strikes, including out-of-the-money puts whose IVs are higher in a downward skew (Example 7), and it blends two expiries to a 30-day horizon; both can lift it slightly above the at-the-money IV.

E6. Interview Without a calculator: a stock is at ₹100 with 20% volatility, interest rates are zero. Roughly what is a one-year at-the-money call worth? What is the corresponding straddle worth, and what does that number mean?

Rule $C\approx0.4S\sigma\sqrt T=0.4\times100\times0.2\times1=$ ₹8 (exact Black–Scholes: ₹7.97).
Straddle With $r=0$ the put is also ≈ ₹8, so the straddle ≈ ₹16 — about the expected absolute move, $\sqrt{2/\pi}\,S\sigma\sqrt T\approx0.8\times20=16$. A strong answer explains where 0.4 comes from: $1/\sqrt{2\pi}\approx0.399$.
🧑‍🏫 For the teacher

Suggested plan — 2 lectures × 60 min.

  1. Greeks and hedging. Open with Kabir's three days (📍) and ask the class to guess which input caused each change. Define the Greeks as partial derivatives (Definition 6.15), derive delta on the board with the identity $Se^{-qT}\varphi(d_1)=Ke^{-rT}\varphi(d_2)$, and do Examples 1–2 (the Taylor P&L). Project Interactive 1 with delta and gamma, then press Play. Example 3 for delta hedging; state the theta–gamma trade-off (Theorem 6.19) and the √T rules (Theorem 6.20) with Example 4.
  2. Volatility. Project Interactive 2 three times: realised = implied (hedged P&L ≈ 0), realised above implied (seller loses), and "never rebalance" (P&L spreads out). Then implied volatility: Theorem 6.21's monotonicity argument and Newton's method (Example 6, the mini solver). The smile from the simulated chain (Example 7), India VIX and Example 8. Finish with the 📈 Market case and Practice corner E4.

Class activity. Give each group a different day count (30, 15, 7, 3, 1) and have them compute the at-the-money call's gamma and theta with the formula; plot all groups' results on the board against $1/\sqrt T$ to discover Theorem 6.20.

Misconceptions to watch for. Delta = probability; theta as "free income"; vega and theta units (per year vs per day, per unit vs per point); "IV is a forecast"; "VIX predicts direction"; one IV for every strike; believing a hedge set once stays a hedge.

Suggested homework. I(1–4, 6), II(1–4, 6), III(1, 2, 4, 6), IV(1, 3, 4, 5); Practice corner E4, E5.

🛠️ Do it yourselfSpreadsheet + Python mini-project · about 60–90 minutes

Project. Build an option "Greeks and IV" sheet, then use a real option chain to draw today's volatility smile and compare it with India VIX.

Why it matters. The option chain shows prices and IVs, but not where they come from. Computing them yourself shows how sensitive an option is, how fast it decays, and how the market prices crash risk — before any money is at stake.

What to do.

  1. Inputs: spot (B1), strike (B2), days (B3), $r$ (B4), $q$ (B5), $\sigma$ (B6); $T$ in B7 =B3/365. Compute d1 and d2 as in §6.3's sheet, and phi1 =EXP(-d1^2/2)/SQRT(2*PI()).
  2. Greeks: call delta =EXP(-B5*B7)*NORM.S.DIST(d1,TRUE); gamma =EXP(-B5*B7)*phi1/(B1*B6*SQRT(B7)); vega per point =B1*EXP(-B5*B7)*phi1*SQRT(B7)/100; call theta per day =(-B1*EXP(-B5*B7)*phi1*B6/(2*SQRT(B7))-B4*B2*EXP(-B4*B7)*NORM.S.DIST(d2,TRUE)+B5*B1*EXP(-B5*B7)*NORM.S.DIST(d1,TRUE))/365.
  3. Check each Greek numerically: reprice with spot $+1$ and $-1$ and compute (C(S+1)-C(S-1))/2 for delta, C(S+1)-2C(S)+C(S-1) for gamma; with days $-1$ for theta; with $\sigma\pm0.005$ for vega. They should agree with the formulas to several digits — derivatives as limits of difference quotients.
  4. IV solver. Put a market price in B9 and a trial $\sigma$ in B6; use Goal Seek ("set the price cell to B9 by changing B6"), or in Python: from scipy.optimize import brentq; iv = brentq(lambda s: bs(S,K,T,r,q,s) - price, 1e-4, 3).
  5. Your smile. From the publicly available NSE option chain for the index, copy the mid-prices (average of bid and ask) of out-of-the-money puts and calls for one expiry, at 8–12 strikes. Compute each IV, plot IV against strike, and note the same day's India VIX from NSE's website.

Hand in: the Greeks table for three expiries with the numerical checks, your smile chart with the at-the-money IV and VIX marked, and three sentences describing its shape.

Going further. Repeat the smile for a near and a far expiry and compare the slopes; or simulate delta hedging in Python as in Interactive 2 and measure how the spread of hedged P&L depends on the number of rebalances.

Instructor note: the difference quotients should match the analytic Greeks to about four significant figures for steps of 1 point and 0.5 vol point (gamma is the most sensitive; use $h\ge1$ point to avoid rounding noise). Real chains contain stale or illiquid quotes far from the money — students should check the bounds of Theorem 6.21 and drop strikes with no IV or very wide bid–ask spreads. Expect the smile to slope down for the index; do not promise a particular level of IV or VIX.

🧪 Self-check quiz

0/5 correct

Q1. Other things equal, an option's gamma is largest when the option is

Q2. Implied volatility is

Q3. India VIX reads 16. The implied one-standard-deviation move of the index over the next 30 days is about

Q4. A long call or long put position typically has

Q5. An option bought at an implied volatility of 15% is delta-hedged frequently until expiry. It makes money (before costs) if