Implied volatility, IV rank and IV percentile explained
What implied volatility is, how it is solved from an option price, how it differs from historical volatility, and how IV rank and IV percentile are computed.
- By
- Veloce team
- Published
- Reading time
- 8 min read
On this page
Implied volatility (IV) is the volatility that makes an option pricing model return the option’s market price. It is read backwards from the price, not measured from the index, so it tells you what the market is charging for movement rather than how much the index has moved. It is solved numerically: guess a volatility, price the option, compare, and adjust until the two agree. Historical volatility is the other side of the comparison: the standard deviation of the index’s past returns, annualised. IV rank places today’s IV between its one-year low and high, and IV percentile is the share of past days on which IV was lower than today. This article solves IV for a NIFTY weekly option step by step, builds a small volatility smile, computes historical volatility from a series of closes, and shows how IV rank and IV percentile can disagree.
The option prices, closes and IV history below are made up for the examples and labelled illustrative. Every IV is solved with Veloce’s own pricing code. Nothing here is a recommendation.
What implied volatility means
An option’s price depends on six inputs: the index level, the strike, the time to expiry, the interest rate, the dividend yield and volatility. The first five are known. Volatility, the size of future moves, is not. Put the market’s price into the model and solve for the one input you do not know, and the answer is the implied volatility.
IV is quoted as an annual figure. An IV of 12% on NIFTY means the option is priced as if NIFTY’s annualised standard deviation of returns will be 12% until expiry. Over shorter periods it scales with the square root of time:
- With 6.24 days to expiry (0.017085 years), a 12% IV implies a one-standard-deviation move of 26,000 × 0.12 × √0.017085 = 408 points by expiry.
- For one calendar day, it is 26,000 × 0.12 × √(1/365) = 163 points. Using 252 trading days instead gives 197 points. The convention matters, and the two figures are not interchangeable.
Under the model’s assumptions, NIFTY would end within one standard deviation about two times in three. Real index returns have fatter tails than that, so treat it as a yardstick, not a forecast.
How implied volatility is solved
There is no formula that gives IV directly from a price, so it is found by search. Here is the example, using these illustrative inputs: NIFTY at 26,000, the 26,000 call on the weekly contract expiring Tuesday, December 1, 2026, valued at 10:00 on Wednesday, November 25, with a rate of 6.5% and a dividend yield of 1.2%. The call trades at ₹180 (illustrative).
Newton’s method uses vega, the price change per unit of volatility, to step from a guess to the answer:
| Step | Volatility guess | Model price | Vega per 1.00 of volatility |
|---|---|---|---|
| 0 | 30.0000% | 418.32 | 1,354.28 |
| 1 | 12.4027% | 180.07 | 1,352.75 |
| 2 | 12.3976% | 180.00 | 1,352.75 |
The first step is 0.30 − (418.32 − 180) / 1,354.28 = 0.124027. Two steps later the model price matches ₹180 to the paisa, and the IV is 12.40%. For comparison, the same call priced at ₹174.62 gives 12.00%: about ₹5.40 of premium is 0.4 points of IV, which matches its vega of 13.53 per point.
Newton’s method can misbehave for far out-of-the-money options, where vega is tiny and a step can shoot off. Veloce’s solver keeps each Newton step inside a bracket between 0.01% and 1,000% volatility, which shrinks with every guess. If a step leaves the bracket, or the previous step failed to halve it, the solver bisects instead. It stops when the bracket is narrower than 10⁻¹² and always converges.
Prices that have no IV. Not every price can be explained by a volatility. A call can never be worth less than its discounted intrinsic value against the forward, or more than the discounted forward itself. Veloce’s solver checks these no-arbitrage limits first. In the example, a 25,500 call quoted at ₹480 is rejected: its lower limit is ₹522.97, so no volatility can produce ₹480. In live data this usually means a stale last-traded price, not a free lunch.
The volatility smile and skew
Each strike and expiry has its own IV. Solving a few illustrative quotes on the same weekly expiry gives:
| Option | Illustrative premium | Implied volatility |
|---|---|---|
| 25,600 put | ₹32.00 | 12.15% |
| 25,800 put | ₹78.00 | 12.33% |
| 26,000 put | ₹151.00 | 11.99% |
| 26,000 call | ₹175.00 | 12.03% |
| 26,200 call | ₹86.00 | 11.68% |
| 26,400 call | ₹34.00 | 11.31% |
The pattern, with out-of-the-money puts priced at higher IVs than out-of-the-money calls, is called skew. Index options usually show it because falls tend to be faster than rises and many participants buy puts as insurance. The practical consequence: “the IV” of NIFTY is a simplification. A strategy that sells the 25,600 put and one that sells the 26,400 call are selling different volatilities. When people quote a single IV, they usually mean the at-the-money option’s.
Historical volatility
Historical (or realised) volatility is measured from the index itself. Take daily closes, compute log returns ln(close today / close yesterday), take their sample standard deviation, and annualise by multiplying by the square root of the number of periods in a year.
With these 21 illustrative NIFTY closes, ending at 25,808.07:
25,500.00, 25,607.10, 25,440.65, 25,664.53, 25,584.97, 25,615.67, 25,346.71, 25,534.27, 25,626.20, 25,492.94, 25,735.12, 25,688.80, 25,845.50, 25,630.99, 25,700.19, 25,826.12, 25,722.82, 26,010.91, 25,826.23, 25,864.97, 25,808.07
the 20 daily log returns have a standard deviation of 0.6386%. Annualised:
- × √252 (trading days): 10.14%
- × √365 (calendar days): 12.20%
The difference is entirely the convention. Daily closes only exist on trading days, so √252 is the usual choice for historical volatility. Option models, including Veloce’s, usually measure time to expiry in calendar days, so their IV is per calendar year. Comparing a 10.14% HV with a 12% IV without adjusting for that is comparing different units. Either annualise both the same way, or compare expected moves in points over the same dates.
Why IV usually sits above HV. Across many studies of index options, implied volatility has on average been higher than the volatility later realised. Researchers call the gap the variance risk premium (Carr and Wu, 2009). Option sellers are paid for carrying the risk of large moves. “On average” hides long quiet stretches followed by sharp losses, so the premium is not something a backtest should assume; it is something a backtest measures, after charges.
India VIX
India VIX is NSE’s volatility index. It uses the CBOE method, adapted to NIFTY options, and is computed from the best bid and ask quotes of out-of-the-money NIFTY options in the near and next expiries, interpolated to a constant 30 days to expiry (NSE: India VIX). A reading of 12 means the options market is pricing about 12% annualised volatility over the next 30 days.
Three things to keep in mind:
- It is a 30-day measure. A weekly option with six days to go can have quite a different IV, especially before an event.
- It is not the at-the-money IV. It blends many strikes, including the skewed puts, so it can differ from the at-the-money IV of either expiry.
- It is a convenient history. For IV rank and IV percentile, India VIX gives a clean daily series without having to pick which option’s IV to track as contracts expire.
IV rank and IV percentile
Both place today’s IV in the context of the past year. They answer different questions.
IV rank is where today’s IV sits between the year’s low and high:
IV rank = (current IV − 52-week low) / (52-week high − 52-week low)
IV percentile is the share of days in the past year on which IV was below today’s:
IV percentile = days with IV below today / total days
An illustrative year of at-the-money IV: the low was 9.5%, the high was 24% during a single sharp sell-off, and today’s IV is 13%. IV was below 13% on 148 of the last 250 trading days.
| Measure | Calculation | Result |
|---|---|---|
| IV rank | (13 − 9.5) / (24 − 9.5) | 24.1% |
| IV percentile | 148 / 250 | 59.2% |
The two disagree because of the spike. The one day at 24% stretches the range, so today’s 13% looks low on IV rank. IV percentile counts days instead, and on most days of the year IV was lower than it is today. When a big event falls inside the look-back window, IV rank stays depressed until it rolls out a year later.
Which to use is a matter of what you want to know. IV percentile is less sensitive to outliers. IV rank reacts to how far IV is from its extremes. Neither is a trading signal by itself, and both depend on choices you should write down before testing: which IV (at-the-money, India VIX, a fixed delta), which window (252 days is common) and what time of day you sample.
Common mistakes with implied volatility
- Solving IV from a stale price. The last traded price of an illiquid strike can be minutes old. Use the mid of the bid and ask, or a candle that actually traded. Veloce’s closest-premium strike selection, for example, only looks at strikes whose last closed candle traded that day.
- Mixing day-count conventions. IV per calendar year against HV per trading year, as above.
- Treating weekly IV as stable. A weekly option’s IV can jump on expiry day and before events, because small premiums and few hours left make it very sensitive to the price. The expiry-day guide shows how little vega is left by then.
- Assuming high IV means sell. Premiums are higher when IV is high because moves are expected to be larger. A strategy that sells when IV percentile is high still needs to beat the moves that follow, after charges and slippage.
- Ignoring the model. IV is model-dependent. Black-Scholes on spot and Black-76 on the future give the same IV only if the forward is consistent. Black-Scholes and Black-76 for Indian options explains the difference.
Try it in Veloce
Veloce backtests option strategies on NIFTY and BANK NIFTY (NSE) and SENSEX (BSE) on one-minute data. Because it replays the recorded prices of the actual contracts, the IV the market charged on each day is already inside every fill; there is no volatility assumption to get wrong.
Veloce’s API solves implied volatility for a list of quotes in one call: POST /api/analytics/iv takes the index or futures price, the rate, the model and each option’s strike, expiry and price, and returns each IV or the reason it cannot be solved, such as a price outside the no-arbitrage limits. Coming soon: an option chain and analytics screen with IV by strike in the app. IV-based entry filters are not available as a signal today.
Create a free Veloce account to test the strategies that IV discussions are usually about, such as the short straddle and the iron condor. For the sensitivities that link IV to P&L, read option greeks explained; for how a test should be judged, see reports and parameter sweeps.
Frequently asked questions
- What is implied volatility in simple terms?
- It is the volatility you have to put into an option pricing model to get the option's market price out. A higher price for the same strike and expiry means a higher implied volatility. It is quoted as an annual percentage, such as 12%.
- How is implied volatility calculated?
- There is no formula that gives it directly, so it is solved numerically: guess a volatility, price the option, compare with the market price, and adjust. Newton's method converges in a few steps; Veloce's solver keeps Newton inside a shrinking bracket so it always converges.
- What is the difference between IV rank and IV percentile?
- IV rank places today's IV between the past year's low and high: (current - low) / (high - low). IV percentile is the share of past days on which IV was lower than today. A single spike drags IV rank down for a year but barely moves IV percentile.
- Is implied volatility the same as India VIX?
- No. India VIX is one number for the whole NIFTY options market: a 30-day expected volatility built from the bid and ask prices of out-of-the-money NIFTY options. Each option also has its own implied volatility, which varies by strike and expiry.
- Is high IV good for option sellers?
- Higher IV means higher premiums, but it usually comes with larger expected moves. Whether selling at high IV pays depends on whether the index then moves less than the IV implied, after charges. That is a question to test, not a rule.
- Why is the IV of out-of-the-money puts higher?
- Index options usually show skew: lower strikes trade at higher IVs, because traders pay more for protection against falls, which tend to be sharper than rises. In our illustrative chain the 25,600 put has an IV of 12.15% and the 26,400 call 11.31%.
- Can I backtest a strategy that enters only when IV is high?
- Not as a built-in signal yet. Veloce's entry signals work on the index's candles, such as moving averages, Supertrend and RSI. Veloce's API can solve IV for a list of option quotes, and an analytics screen for IV is coming soon.
Sources
- Black, F. and Scholes, M. (1973): The pricing of options and corporate liabilities, Journal of Political Economy 81(3)doi.org, accessed
- NSE: India VIX (computation methodology and white paper)nseindia.com, accessed
- Carr, P. and Wu, L. (2009): Variance risk premiums, Review of Financial Studies 22(3)doi.org, accessed
- NSE circular NSE/FAOP/68747: revision in expiry day of index and stock derivatives (June 25, 2025)nsearchives.nseindia.com, accessed