Sunday 1 Feb 2026. In the Budget session the Nifty fell 1.96%. On Friday 30 Jan India VIX had closed at 13.63, which prices an ordinary trading day at about 0.86 percent. The Budget session moved 2.3 times that. Then came Tuesday 3 Feb, which rose 2.55 percent, a bigger move than the event itself. Anyone hedging "for the Budget" had two questions: what did the market charge for the event, and was the hedge pointed at the right day? My data set has no chain from 30 Jan, so I cannot show the price. This lesson shows the arithmetic on chains I do have, and the week as it unfolded. Dated NSE data, nothing live, nothing a view on the next event.
3By the end you can
- Read the variance of an event day from two expiries and say what base volatility it rests on.
- Compare the Budget week of 2026 with the move an ordinary day was priced for.
- Price a put and a put spread before an event and read the crush in a scenario grid.
The idea, plainly
What is event variance?
Variance adds up over days. An expiry that holds a known event day carries the variance of its ordinary days plus the variance of the event day, which is larger than an ordinary day's. Subtract the ordinary days and what remains is the market's price for the event. Its square root is the one day move the price allows for. The Budget, an RBI policy day, election counting and a results day all work this way.
Total variance = IV^2 * T Two expiries, T1 below T2: span = IV2^2*T2-IV1^2*T1 event day = span -base^2*(T2-T1-d) d: 1 day in years base: quiet day vol
How do I read it from two expiries?
Take the at the money implied volatility of a near expiry and a later one that holds the event. Work out the variance of the span between them. Take away the variance of the quiet days, at a base volatility you have to choose. What is left belongs to the event day. The forward volatility tool on this page does the sum for any dated pair.
A term structure is a path with forward volatilities between the points
Read 3 more paragraphs, 211 words
Which expiries to pair matters. The nearest expiry is distorted by expiry day gamma, so I avoid a near leg with only a day or two left. A weekly and a monthly are priced a little differently, so a mixed pair can show a bump that is only that difference. Pairs of the same kind, with a few days between, are cleaner.
The sum has one catch. The chart shows a rise between expiries, and a rise is exactly what an event lump looks like. It is also what an ordinary upward sloping term structure looks like. The term structure cannot tell the two apart. Only a calendar can: a dated event that falls inside the span. I read event variance only when I can name the date.
India makes the calendar easy to read and hard to price. Budget day is announced, RBI policy dates are published well ahead, results dates are filed with the exchange, and counting day is fixed. The Budget of 1 Feb 2026 fell on a Sunday and the exchange held a special trading session. Each is a name on a calendar, which is what the sum needs. What none of them gives is the size, which is why the sum is only half the work.
How fragile is the answer?
Take the 16 Jun and 30 Jun 2026 expiries of the 10 Jun 2026 chain. No dated event sits between them, and the pair is a weekly against a monthly, so this tests the sum and reads no event. Priced at 14.61% and 15.03% for 6 and 20 days, the span carries more variance than quiet days at the near volatility would. If the quiet days run at 14.61%, the leftover day carries 2.17 times an ordinary day's variance. If they run at 15.03%, it carries 1.33 times. If at 15.53%, only 0.42.
A nearly flat curve leaves little for an event, and the answer rests on one guess
The maths: event day as a multiple of an ordinary day
Span variance: w = IV2^2*T2-IV1^2*T1 event = w-b^2*(T2-T1-d) ratio = event / (b^2 * d) ratio = 1: an ordinary day Move = S * sqrt(event) b = forward vol: ratio = 1 b below it: ratio above 1
If the quiet days run at exactly the forward volatility of the span, nothing is left over and the ratio is one. A lower base leaves more for the event, and a higher one leaves less. The ratio is a multiple of variance. The move is its square root.
Read one more paragraph, 61 words
Same chain, three guesses, three answers from more than double to under half of an ordinary day. The event variance is the residue after a guess, and the residue swings when the guess moves a few tenths of a vol point. That is why I show the sensitivity, as the tool does, and why I treat the reading as a range.
Worked example: the Budget week and a hedge before it
What did the Budget week do?
The ordinary day at VIX 13.63 is a move of 0.86 percent. Here is the week, five sessions, with moves of 0.30, 0.39, 1.96, 1.06 and 2.55 percent in size. The Budget session was 2.3 times an ordinary day, and the next two sessions were both larger than an ordinary day too.
The biggest move of the Budget week came two sessions after the Budget
Read one more paragraph, 71 words
An event hedge is a bet on a date. A put that expires on the day of the event covers the event. A put that expires a week later covers the event and everything after. The week showed that the move can come later. I do not read this as a case for hedging after the fact. I read it as a reason to know exactly which sessions a hedge covers.
What does a hedge before an event cost?
Take one lot of 65 units on the 10 Jun 2026 chain, with the 16 Jun weekly. A put struck at 22,700 costs 27.60 a unit, Rs 1,794 a lot, which is 0.12% of the lot's value. The index has to fall 2.34% for it to pay at expiry. An ordinary day at that date's India VIX is 0.98 percent, so the break even is about 2.4 ordinary days.
After the event, the put loses on time and on crush even when the index is flat
| Index move | 3 days on, rupees a lot | ||
|---|---|---|---|
| IV −3 | IV same | IV +3 | |
| −4% | +24,599 | +25,189 | +26,001 |
| −2% | +2,976 | +4,538 | +6,113 |
| 0% | −1,665 | −1,355 | −829 |
| +2% | −1,794 | −1,787 | −1,752 |
| +4% | −1,794 | −1,794 | −1,793 |
Green is profit and red is loss; deeper colour means a larger amount. The outlined cell is the starting point.
Read 2 more paragraphs, 147 words
The grid shows the part people forget. With the index unmoved, the put loses on time, and it loses more when implied volatility falls three points once the event is out: compare the left column with the middle one on the unmoved row. A fall of two percent pays, and pays less if volatility drops with it. The price of waiting for an event is paid whether or not it moves.
Costs add to this. A buyer pays the premium up front, STT does not apply to a put that expires worthless, and exercise on an in the money index option is taxed on its intrinsic value, so a put that pays in a fall has a charge at expiry too. The cost of a trade lesson adds the lines up. None of it changes the shape of the grid, and all of it is on top.
The term this section turns on: Crush
Are there cheaper shapes?
A put spread sells a lower strike against the put. On the same chain, the 22,700 put bought against the 22,400 put sold costs Rs 1,216 a lot, and the most it can pay is Rs 18,284 a lot. Cheaper, and capped. Selling a lower strike gives away the part of the move beyond it, which is the part a hedge buys for a real shock. The collar and put spread lesson sets the trade offs out.
Read 3 more paragraphs, 207 words
Two other tools carry no option premium. Hold less beforehand, which has no volatility exposure at all, or sell a future against the holding, which has margin and costs of its own but no crush. Neither pays like an option, and neither loses to crush. Which suits an event depends on whether it is the size of the move or the timing of it that worries the holder, and no number in this file decides that.
One more line on the week. The straddle logic of the IV crush lesson says a long straddle gains only if the move beats its price. The same holds for a hedge: a put that cost the event's price pays only for the part of the move beyond what was paid for. A hedge is insurance, so the premium is a cost every time the event passes quietly. Whether that cost is worth paying is a question about the holder's exposure, not about the chain.
Elections, RBI policy days and results days follow the same arithmetic. The files here have no pre event chain for them, so I draw no number. The method carries over: name the date, find the two expiries around it, choose the base, and show the range.
The common mistake
The claim: The term structure tells you how big the event move will be.
It tells you what is left after a guess. On the 10 Jun 2026 chain the same prices give 2.17 times an ordinary day, 1.33 times or 0.42 times, depending on the quiet day volatility chosen. Even when the reading is sound, the Budget week of 2026 shows a move can land on a different day than the event, and a bigger one. Treat the implied event move as a range with a stated base, and know which sessions the hedge covers.
Key takeaways
- Event variance is the variance of a span less its quiet days.
- It is a residue: the answer swings with the quiet day guess.
- Only a dated event inside the span makes a rise in the curve an event.
- The biggest move of the Budget week came two sessions after the Budget.
- A put bought before an event loses to crush even when the index is flat.
- A hedge covers sessions: know which ones it covers.
Two terms to hold on to from 17.19
Before you trade. Derivatives trading carries a high risk of loss. SEBI's study "Profitability of Individual Traders in the Equity Derivatives Segment (FY25 to FY26)" (August 2026) found that 87.7% of individual traders in equity F&O made net losses in FY26, with aggregate net losses of Rs 91,685 crore. Read the SEBI study