Wednesday 10 Jun 2026. India VIX closed at 15.63. I took the settlement prices of the 30 Jun and 28 Jul Nifty expiries from NSE's end of day file, ran the recipe in NSE's own India VIX methodology note, and got 15.50. Same recipe, different prices, −0.13 points apart. India VIX is not a model and not an at the money vol. It is a sum over a strip of options. This module rebuilds that sum step by step, so you can see where the number comes from, and why a rebuild never matches to the second decimal. Every figure comes from NSE end of day files, dated, so nothing here is live and nothing is a view on the next move.
3By the end you can
- Rebuild a 30 day volatility index from two dated Nifty expiries, step by step, and compare it with the published India VIX.
- Say which strikes carry the most weight in the sum, and why far out of the money options still count.
- List the differences between a rebuild on end of day prices and NSE's own calculation.
The idea, plainly
What is India VIX, in one line?
It is NSE's volatility index. Its methodology note says it uses the best bid and ask quotes of the near and next month Nifty options, follows a method first built for the US index options market with amendments for the Nifty order book, and reads as the expected volatility over the next 30 calendar days, as a yearly percentage. The India VIX module covers what it means. This one covers how it is made. The index does not pick the at the money option. It adds up the out of the money options on both sides of the forward.
Why does a strip of options give a variance?
The fair price of a contract that pays the realised variance of the index over a period can be built, in theory, from out of the money options of every strike, each weighted by one over its strike squared. That is the result in the volatility swaps paper listed in the sources. India VIX takes that price, scales it to a year and reads the root. Two things follow. Far out of the money options count for more than their small prices suggest, because the weight grows as the strike falls. And no model of the index is used, only prices.
Six steps from a chain to the index
What are the steps?
- The forward. NSE takes the latest traded price of the Nifty future of that expiry. This rebuild takes the parity forward, found at the strike where the call and put settled closest. For the 30 Jun expiry that was 23,249.60, against 23,240.10 for the future.
- K0. The strike just below the forward: 23,200 here.
- The strip. Puts below K0, calls above it, and at K0 the average of the put and the call. NSE uses the middle of the best bid and ask, and fills a strike whose spread is over thirty percent of the middle with a cubic spline. The rebuild uses the settlement price, skips a strike that settled at zero, and stops each side after two strikes in a row that did not trade.
- Each strike's share. The gap between strikes over the strike squared, times the option price, times a small interest factor. The factor is e raised to the rate times the time, and for the near expiry here it is 1.0028.
- The expiry's variance. Add the shares, double, divide by time, take off a small term for the forward not sitting on K0.
- The blend. Near and next expiry are weighted so that the result sits at thirty days, then scaled to a year.
- The root. Take the square root and multiply by one hundred.
One more rule on the strip itself. NSE's note says a spline is built only when at least three usable quotes exist. If they do not, no volatility is computed for that expiry on that tick, and the previous tick's value for it is carried forward. I take the method from that published note. NSE may have refined details since, which is one more reason a rebuild differs.
What are the steps?, in 7 steps
The maths: the three formulas
For one expiry, with strikes K, the gap between neighbouring strikes dK, the option price Q, the rate R and time T in years:
s^2 = (2 / T) x sum of
[dK / K^2 x e^(RT) x Q]
less (1 / T) x (F / K0 - 1)^2
The gap dK is half the distance between the strikes either side, and the whole step at the two ends of the strip. The blend takes the near and next variances, with N the days to each expiry:
w1 = (N2 - 30) / (N2 - N1)
w2 = (30 - N1) / (N2 - N1)
VIX = 100 x sqrt(
[T1 s1^2 w1 + T2 s2^2 w2]
x 365 / 30 )
NSE counts time in minutes, to the close at 3:30 pm on expiry day, uses an NSE MIBOR tenor for the rate, and rolls to the next and far expiry when three trading days are left in the near one. This rebuild counts calendar days, uses the T-bill rate of the date, and stays on the two monthlies that bracket 30 days.
Worked example: 10 Jun 2026
What does each step give?
The two expiries are the 30 Jun monthly with 20 days left and the 28 Jul monthly with 48. Thirty days lies between them, so the blend interpolates. Read the table down each column and then across.
| Step | Near | Next |
|---|---|---|
| Days to expiry | 20 | 48 |
| Forward F | 23,249.60 | 23,348.84 |
| K0 | 23,200 | 23,300 |
| Sum of dK/K^2 x Q | 0.000707 | 0.001483 |
| Variance s^2 | 0.02581 | 0.02267 |
| Volatility s | 16.06% | 15.06% |
| Weight in blend | 0.6429 | 0.3571 |
Read one more paragraph, 92 words
The near variance, 0.02581, is above the next, 0.02267: the shorter expiry carries more variance per year, as the curve of at the money vols also showed. Blend the two variances with those weights, scale to a year, and the 30 day variance is 0.02401. Its root, times one hundred, is 15.50. NSE published 15.63. The two expiries' volatilities, 16.06% and 15.06%, are both above the 15.03 percent at the money vol of the near expiry, because the strip includes the left side of the smile.
Where does the number come from?
Split the near expiry's sum by how far each strike is from K0. The four groups sum to 0.000397, 0.000203, 0.000070 and 0.000038, from the centre outward. The centre dominates, because option prices are largest there. The wings count less each but cover many strikes, and their weight per rupee of premium is higher.
More than half the sum sits within two percent of the money
Read one more paragraph, 64 words
Strikes within two percent of K0 carry 56.1 percent of the sum, those two to five percent away 28.7 and those five to ten percent away 9.8. The strikes more than ten percent away, which most traders never look at, still carry 5.4 percent. A rise in the price of those far puts would lift India VIX with no change at the money.
How close does the rebuild land?
On 7 Apr 2025 the near expiry had 17 days and the next 52. Their variances were 0.07064 and 0.03724, and the rebuild gave 22.17 percent against 22.79 published. On 11 Jun 2025 the expiries had 15 and 50 days, variances 0.02000 and 0.01970, and the rebuild gave 14.07 against 13.67. The rebuild was under the published index on two dates and over it on one.
The rebuild landed within a point of the published index each time
The gap has no fixed sign because the differences pull in different directions.
- Prices. NSE uses the middle of the best bid and ask through the day. The rebuild uses one settlement price per strike.
- The forward. NSE takes the future. The rebuild takes parity.
- Gaps in the strip. NSE fills them with a cubic spline. The rebuild stops each side after two strikes that did not trade.
- Time and rate. Minutes to 3:30 pm and a MIBOR tenor, against calendar days and a T-bill.
- The roll. With three trading days left in the near expiry, the note has NSE move to the next and far expiry. The three dates here are all well clear of that.
- The note itself. Its worked example uses index levels from long ago, and I read the method from it, not the numbers.
Is it the at the money implied volatility?
No. On 10 Jun 2026 the at the money vol of the near expiry was 15.03 percent and India VIX 15.63. The chart sets India VIX beside a constant 30 day at the money measure from the same files.
The at the money measure runs under India VIX for most of the year
Read one more paragraph, 55 words
Over the 300 sessions from 1 Jul 2025 to 15 Sep 2026 in the series files, a longer span than the chart, the at the money measure averaged 12.42 and India VIX 13.49. The file's own note says why: the at the money measure reads one strike, and India VIX reads the whole smile.
What the index does not do
It does not say which way Nifty goes. The strip holds puts and calls, and a variance has no sign. It does not read one strike, so a quiet at the money option with dear wings still lifts it. And it does not forecast: it is a price for the next thirty days of variance, with the charge for bearing the risk inside it. When India VIX jumps and the at the money vol barely moves, the wings are the place to look, and a rebuild like this one shows how much of the move came from them.
The common mistake
The claim: India VIX is the implied volatility of the at the money Nifty option.
It is not. On 10 Jun 2026 the at the money vol of the 30 Jun expiry was 15.03 percent against 15.63 for the index, and the strip behind the index carried 5.4 percent of its sum in strikes more than ten percent from K0. The at the money option is the centre of the strip, not the whole of it. I read India VIX as a price of variance over the next 30 days, and I do not read it as one option's vol.
Key takeaways
- India VIX adds up out of the money options on both sides of the forward.
- Each strike counts for its price over the square of its strike.
- Two expiries are blended to a constant 30 days, then rooted.
- The rebuild landed within a point of the published index on three dates.
- More than half the sum sat within two percent of the at the money strike.
- India VIX is not the at the money implied volatility.
The term this section turns on: India VIX
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