Volatility Metrics Intermediate The forward-looking figure against the backward-looking one One backward, one forward

HV vs IV

The comparison everyone makes, and almost everyone makes wrong.

Quick answer: HV vs IV is the comparison between historical volatility — how much the underlying actually moved in the past — and implied volatility — how much option prices say it will move in the future, and the gap between them is the market's price for uncertainty rather than a mispricing to be harvested.

In simple words

There are two volatility numbers on every underlying and they are asking different questions. Historical volatility (HV, also called realised volatility) measures how much the price actually moved over some past window — you compute it directly from the price history. Implied volatility (IV) is baked into option prices and represents how much the market thinks the price will move from now on. Put them side by side and it is tempting to say "IV is 14% but HV is only 11%, so options are overpriced by 3 points". That sentence is the single most common error in options, and it is usually wrong for a reason that has nothing to do with the market: the two numbers being compared are measuring different periods.

To compare HV and IV honestly you have to line them up on the same tenor. A 30-day implied volatility is a statement about the next 30 days, so the only fair comparison is against 30 days of realised volatility — and, more subtly, against the 30 days of realised volatility that arrive after you buy the option, not the 30 days that already happened. That second point is the uncomfortable one: the honest version of "are these options expensive?" requires a realised-volatility number for a period that has not occurred yet, which means nobody actually has it. Everything traders do — comparing IV to trailing HV, to a forecast, to India VIX — is an approximation of a comparison that is, strictly, impossible to make in advance.

Not to be confused with: The volatility risk premium, which is what the HV-vs-IV gap is called once you compare implied volatility against the realised volatility that subsequently arrives, averaged over many periods. HV vs IV is the general comparison; the volatility risk premium is the specific, forward-looking, averaged version of it — and only the latter is a coherent measure of whether option sellers are being paid.

The picture

Two lines that usually run parallel, until they cross

Trailing realised volatility and at-the-money implied volatility on the same underlying over time.

In calm periods implied volatility sits a couple of points above trailing realised volatility — the everyday premium for uncertainty. The lines cross precisely when a shock hits and realised volatility leaps above implied: that inversion is the moment short-volatility positions lose money, and it is visible here as the realised line punching through the implied one.

Professional explanation

Like tenors, or the comparison is meaningless

The first rule of comparing HV and IV is that they must cover the same length of time. A 30-day implied volatility summarises the market's view of the next 30 days; comparing it against a 10-day realised volatility is comparing a month's expectation against ten days' history, and any difference you find is mostly an artefact of the mismatched windows, not a signal about the option market. This error is everywhere because platforms default to convenient windows — a 10-day or 20-day HV against a front-month IV — and traders read the difference as edge. If the implied volatility you are looking at is for a 30-day option, pair it with a 30-day realised volatility. If you want to judge a weekly option, use a realised volatility measured over roughly a week. Match the horizons first; everything else is downstream of that.

The right realised volatility is the one that has not happened yet

Even with matched tenors, there is a deeper trap. Trailing realised volatility measures a period that has already finished, but the implied volatility you are comparing it against is a price for a period that has not begun. The economically correct comparison for "is this option expensive?" is implied volatility today against the realised volatility that will unfold over the option's life — a number that does not exist yet and will only be known at expiry. Trailing HV is used as a proxy for it, on the assumption that volatility is persistent, which it usually is. But that assumption fails exactly when it matters: on the eve of a shock, trailing HV is low and forward realised volatility is about to be enormous. So the honest version of the question is unanswerable in advance, and the trader who forgets this is comparing today's price against yesterday's weather.

The normal state, and the inversion that costs money

In ordinary conditions implied volatility sits a little above trailing realised volatility — typically a couple of points on a liquid index. That gap is not a mispricing; it is the price of insurance, the compensation option sellers require for underwriting moves that have not happened. Because it is usually positive, selling options usually collects it, which is exactly the seductive pattern that gets people hurt. The gap inverts — realised volatility punches above implied — precisely when a shock arrives, and that inversion is the moment a short-volatility book takes its losses. The everyday parallel gap lulls, the sudden crossing bites. Any honest treatment of HV vs IV has to say plainly that the comfortable regime and the catastrophic regime are the same trade in two different weathers, and you do not get to choose which one you are in when you put the position on.

Realised volatility is not one number either

"The" historical volatility is itself a choice of estimator, and different estimators disagree. The standard close-to-close measure uses only closing prices and throws away everything that happened intraday, so it understates volatility on days that swung violently and closed flat. Range-based estimators like Parkinson use the day's high and low and see that intraday movement, typically reading higher and with less noise. The window length, the annualisation convention (252 trading days on this site), and whether you subtract the mean all move the number. So before declaring that IV exceeds HV by 3 points, you should know which HV — a close-to-close 20-day figure and a Parkinson 30-day figure on the same stock can differ by more than the premium you are trying to measure, which means the premium can be an artefact of estimator choice rather than a fact about the market.

The gap, and why its sign matters

Implied minus realised volatility over time, shaded above and below zero.

The gap is positive most of the time — a small, steady premium — and occasionally slams deeply negative when realised volatility overwhelms what was implied. The area above zero is frequent and shallow; the area below is rare and deep. That asymmetry, not the average, is what the comparison is really about.

Formula

HV vs IV — the like-for-like comparison

spread = IV_T − HV_T

The comparison only means something when both figures cover the same tenor T. IV_T is the implied volatility of an option with T days to expiry; HV_T is realised volatility measured over a comparable horizon and annualised on 252 trading days. A positive spread means options price more movement than recently occurred — usually the everyday premium, not a mispricing.

spreadThe HV-vs-IV difference in volatility points. Usually small and positive; turns negative when realised volatility overtakes implied, which is when short-vol positions lose.
IV_TImplied volatility of an option with tenor T (e.g. the 30-day at-the-money IV), annualised. Forward-looking — a price for the period ahead.
HV_TRealised (historical) volatility over a horizon comparable to T, annualised on 252 trading days. Backward-looking — a measurement of a period that has finished.

Close-to-close realised volatility (the backward figure)

HV = stdev(ln(P_t / P_{t−1})) × √252 × 100

Realised volatility is the annualised standard deviation of daily log returns over the window. P_t is the closing price on day t; the log ratio ln(P_t / P_{t−1}) is that day's return; stdev is the sample standard deviation; √252 annualises it, and ×100 expresses it in percent. Range-based estimators such as Parkinson use the day's high and low and generally read higher.

How to compare HV and IV correctly

Pick the option tenor you actually care about — say the 30-day at-the-money IV.
Measure realised volatility over a comparable horizon. For a 30-day IV, use roughly 21 trading days of returns, not 10, so the windows line up.
Annualise both on the same convention — 252 trading days on this site — so the two numbers are in the same units.
Decide which realised estimator you mean. Close-to-close is standard; a range-based estimator like Parkinson reads higher because it sees intraday movement. State which one, because the choice can exceed the premium you are measuring.
Compute the spread IV_T − HV_T, but read it as a price for uncertainty, not a mispricing. Remember trailing HV is a proxy for the forward realised volatility that actually matters and does not exist yet.
Watch for inversion. When realised volatility climbs above implied, the everyday premium has flipped negative — that is the regime in which short-volatility positions take losses.
Never conclude "options are expensive, sell them" from a positive spread alone. The spread is usually positive because insurance is sold above its expected cost, not because the market is wrong.

Practical example

NIFTY worked example

NIFTY has spent the past month grinding quietly: its 21-day close-to-close realised volatility works out to about 11%. The 30-day at-the-money implied volatility reads 13%. Matched on tenor, the spread is 13 − 11 = 2 volatility points, implied over realised. Read correctly, that is not "options are 2 points overpriced" — it is the everyday premium the market charges to insure the next month, sitting right where it usually does. Now interpret what a 13% IV prices: a one-standard-deviation move over 30 days of 24,000 × 0.13 × √(30/365) ≈ 894 points. If NIFTY over the coming month actually realises 11% again, the option seller keeps roughly that 2-point premium; if a shock lands and NIFTY realises 20%, realised volatility inverts far above the 13% that was implied, and the seller's small steady premium is wiped out many times over by the one month it went wrong. The 2-point gap was never the whole story — the distribution of outcomes around it was.

BANKNIFTY worked example

BANKNIFTY teaches the tenor trap directly. Suppose its 10-day realised volatility reads 14% after a couple of sharp sessions, while its 30-day at-the-money implied volatility reads 17%. A careless trader compares 17% against 14% and declares BANKNIFTY options cheap-ish, only 3 points of premium. But the 10-day HV is measuring a short, recent, jumpy window, and the 30-day IV is pricing a full month that will average out those jumps. Re-measure realised volatility over 21 trading days to match the tenor and it might come out at 15.5%, shrinking the honest spread to 1.5 points. The 3-point figure was an illusion created by comparing a jumpy ten days against a smoother month. On BANKNIFTY, whose sharp bank-driven moves make short-window HV especially noisy, matching tenors is not a nicety — it is the difference between a real comparison and a mirage.

Lot sizes used above (NIFTY 75, BANKNIFTY 30) are those in force at the time of writing; NSE revises them periodically. Figures exclude brokerage, STT, exchange charges, stamp duty and GST. Examples are teaching scenarios built on round numbers — they are not historical quotes, not backtests and not trade calls.

Advantages & limitations

What it is good for

It grounds an option's price in something checkable. Implied volatility alone is abstract; setting it against realised volatility gives a reference point drawn from the underlying's actual behaviour.
Properly tenor-matched, it is the cleanest single diagnostic of whether options are pricing more or less movement than the underlying has recently delivered.
It exposes regime changes early: when realised volatility starts climbing toward and through implied, the comparison flags that the calm premium is inverting before most other signals do.
It disciplines the trader into thinking about horizons and estimators, which are exactly the details that separate a real edge from a measurement artefact.
It is the conceptual foundation of the volatility risk premium, so understanding it correctly is the prerequisite for every serious volatility-selling or dispersion strategy.

Where it breaks down

The economically correct comparison needs the realised volatility that arrives after the option is bought — a number that does not exist until expiry — so every real HV-vs-IV comparison is an approximation using a proxy.
It breaks the instant the tenors are mismatched: a 30-day IV against a 10-day HV compares two different periods, and the difference is an artefact, not a signal.
Realised volatility is estimator-dependent. Close-to-close and range-based figures on the same stock can differ by more than the premium being measured, so the sign of the spread can flip with the choice of estimator.
Trailing HV assumes volatility is persistent, which fails exactly on the eve of a shock — precisely when the comparison would matter most, it is most wrong.
The comparison says nothing about direction or about the shape of the tail, so two underlyings with the same HV-vs-IV spread can carry very different crash risk.
A positive spread is the normal state, not an opportunity, so reading every positive spread as "sell options" mistakes the price of insurance for a mispricing.

Common mistakes

Comparing a 30-day implied volatility against a 10-day realised volatility and concluding options are expensive. You have compared two different periods and blamed the gap on the option market. Match the tenors first.
Treating a positive IV-minus-HV spread as free edge to sell. The spread is usually positive because insurance is sold above its expected cost, and selling it collects a small premium against a rare large loss — a rule of thumb some traders live by and some are ruined by.
Using trailing realised volatility as if it were the realised volatility over the option's life. On the eve of a shock, trailing HV is low and forward realised is about to be huge, which is exactly the setup that traps premium sellers.
Mixing estimators without noticing — comparing a close-to-close HV against an implied volatility, then switching to a Parkinson HV another day, and reading the change as a market move when it is only an estimator change.
Ignoring the inversion. When realised volatility climbs above implied, the comfortable premium has flipped, and a trader who does not watch for that crossing holds a short-vol position straight into its worst regime.
Comparing raw HV-vs-IV spreads across two different underlyings and treating them as equivalent. A 2-point spread on NIFTY and a 2-point spread on a single stock carry different tail risk and different liquidity.
Annualising the two figures on different conventions — one on 252 days, one on 365 — so the numbers are not even in the same units before they are compared.

Professional usage

Volatility arbitrage desks live in the HV-vs-IV comparison, but they do it forward, not backward: they build a forecast of the realised volatility that will unfold over an option's life, compare it against the implied volatility they can trade at, and take the option accordingly while delta-hedging continuously so the residual profit and loss depends on the difference between their forecast and what actually realises, rather than on the direction of the underlying. The trailing HV a retail screen shows is, to them, just one input into that forecast, and a lagging one. The whole craft is in estimating the number that does not exist yet — forward realised volatility — better than the option market has priced it.

Market makers and risk teams watch the spread between implied and trailing realised volatility as a gauge of how richly the book is being paid to warehouse risk, and as an early-warning line for regime change. When realised volatility starts closing the gap to implied and then crosses it, the desk knows its short-gamma exposure is about to become expensive to hedge, and it adjusts quotes and hedging frequency before the inversion is complete. The comparison is less a trade signal than a continuous read on which volatility weather the book is currently standing in.

Key takeaways

HV vs IV compares realised volatility (what the underlying did) against implied volatility (what options price for the future); the gap is the price of uncertainty, not automatically a mispricing.
The comparison is only valid on matched tenors — a 30-day IV against a 30-day realised volatility, never against a 10-day one.
The economically correct realised figure is the one that arrives after the option is bought, which does not exist yet, so every real comparison uses trailing HV as an imperfect proxy.
Implied volatility normally sits a couple of points above trailing realised; it inverts when a shock hits, and that inversion is when short-volatility positions lose money.
Realised volatility is itself estimator-dependent, so the sign of the spread can change with the choice of close-to-close versus range-based measurement.

HV vs IV looks like the simplest comparison in options — two volatility numbers, subtract one from the other — and it is quietly one of the easiest to get wrong. Match the tenors or you are comparing two different periods; remember that the realised volatility that actually matters has not happened yet; and never mistake the usual small positive gap for a mispricing, because it is the price of insurance and insurance is sold above its expected cost for a reason. The comfortable months when implied sits calmly above realised and the terrible month when realised punches through it are the same position in different weather, and the whole discipline is in respecting that you do not choose the weather.

Frequently asked questions

What is the difference between HV and IV?

Historical volatility (HV) measures how much the underlying actually moved over a past window, computed from price history. Implied volatility (IV) is extracted from option prices and represents how much the market expects it to move in the future. HV is a measurement of the past; IV is a price for the future.

Does IV higher than HV mean options are overpriced?

Not by itself. Implied volatility normally sits a couple of points above trailing realised volatility, and that gap is the price of insurance, not a mispricing. Calling it overpriced assumes you know the realised volatility over the option's life, which has not happened yet.

Why must I compare HV and IV on the same tenor?

Because they must cover the same length of time to be comparable. A 30-day IV prices the next month; a 10-day HV measures ten past days. Any difference between mismatched windows is an artefact of the horizons, not a signal about the option market.

What tenor of HV should I compare to a 30-day IV?

Roughly a month of realised volatility — about 21 trading days — annualised on the same convention. Using a 10-day or 5-day HV against a 30-day IV compares a jumpy short window against a smoother month and produces a misleading spread.

Which realised volatility should I use, close-to-close or Parkinson?

Either, as long as you know which and stay consistent. Close-to-close uses only closing prices and understates days that swung and closed flat; Parkinson uses the day's high and low and reads higher. The two can differ by more than the premium you are measuring.

What does it mean when HV rises above IV?

It means the underlying is now realising more volatility than options had priced — the normal premium has inverted. This is the regime in which short-volatility positions lose money, because the seller collected less than the movement that arrived.

Why is IV usually higher than HV?

Because implied volatility contains a risk premium: option sellers are underwriting movement that has not happened, and insurance is sold above its expected cost. The gap is compensation for bearing that risk, not evidence the market is wrong.

Is the HV-vs-IV gap the same as the volatility risk premium?

Closely related but not identical. The volatility risk premium is implied volatility minus the realised volatility that subsequently arrives, averaged over many periods and measured forward. HV vs IV is the general comparison, often done with trailing HV, which is a backward-looking proxy.

Can I use the HV-vs-IV gap to decide when to sell options?

Some traders use a positive gap as a rule of thumb to sell premium, but it is dangerous. The gap is usually positive because sellers are paid to bear crash risk, and the losses when it inverts are large and rare. A positive spread is the normal price of insurance, not an edge.

Why is trailing HV a poor proxy for future realised volatility?

Because it assumes volatility is persistent, which is usually true but fails exactly on the eve of a shock. Just before a crash, trailing HV is low while forward realised volatility is about to be enormous — the moment the proxy is most wrong is the moment it matters most.

How do I annualise realised volatility?

Take the standard deviation of daily log returns over the window and multiply by the square root of 252, the number of NSE trading days in a year, then express as a percent. Implied volatility is quoted on the same annualised basis, so both are comparable.

Does a small HV-vs-IV spread mean options are fairly priced?

It means options are pricing roughly what the underlying recently realised, which is a reasonable starting point but not a verdict. The realised volatility over the option's life may differ sharply from the trailing figure, and that difference decides the outcome.

Why do close-to-close and range-based HV disagree?

Because close-to-close uses only closing prices and misses everything that happened intraday, so a day that swung wildly and closed flat registers as calm. Range-based estimators like Parkinson use the high and low, capturing that movement, and generally read higher with less noise.

Can the sign of the HV-vs-IV spread change with the estimator?

Yes. If close-to-close HV reads 11% and Parkinson HV reads 14% on the same stock, an implied volatility of 13% is above one and below the other. The apparent richness or cheapness of options can flip purely on which realised estimator you chose.

Does HV vs IV tell me anything about direction?

No. Both are measures of magnitude and say nothing about which way the underlying moves. A wide HV-vs-IV spread is equally consistent with an impending rally or an impending selloff.

Why does the gap invert during a crisis?

Because a shock makes the underlying realise far more volatility than was priced, so realised volatility leaps above implied. Implied volatility also rises, but realised overshoots it, and the seller who collected the earlier small premium now faces a move many times larger than they were paid for.

Is India VIX a form of implied volatility for this comparison?

Yes. India VIX is a model-free 30-day implied volatility summary of the NIFTY option chain, so it is a reasonable IV side of the comparison — paired against a roughly 30-day NIFTY realised volatility for a like-for-like read.

How often does IV exceed HV?

On liquid index options, implied volatility exceeds subsequently realised volatility most of the time, by a small amount. But the exceptions are large and arrive during shocks, so the base rate is not an edge — it is the premium paid for insurance against exactly those exceptions.

Should I compare HV-vs-IV spreads across different stocks?

Only with care. A 2-point spread on NIFTY and a 2-point spread on a single stock carry different tail risk, liquidity, and event exposure. The raw spread does not normalise for how dangerous each underlying's rare moves are.

What is the honest version of 'are these options expensive'?

It is implied volatility today against the realised volatility that will unfold over the option's life — a number nobody has yet, known only at expiry. Every practical answer substitutes a forecast or trailing HV for that missing number, which is why the question is genuinely hard.

Why do professionals forecast realised volatility instead of using trailing HV?

Because the trade depends on the realised volatility over the option's future life, and trailing HV is a lagging proxy for it. A volatility arbitrage desk builds a forward forecast, compares it to tradeable implied volatility, and hedges so the profit depends on the difference between forecast and outcome.

Voice search & related questions

Natural-language questions people ask about hv vs iv.

Are options expensive if IV is higher than HV?

Not necessarily. Implied volatility normally runs a couple of points above trailing realised volatility, and that gap is just the price of insurance for the month ahead, not a mispricing. To call options truly expensive you would need to know how much the underlying moves over the option's life, which nobody knows yet.

Why does my HV look so different from the IV on my screen?

Most likely because they cover different tenors. If your screen shows a 10-day HV next to a 30-day IV, you are comparing ten jumpy days against a whole smoother month. Line them up on the same horizon and the gap usually shrinks to something sensible.

When does selling options against a high IV-vs-HV gap go wrong?

When the underlying suddenly realises more than was priced and realised volatility punches above implied. That inversion is exactly when a short-volatility position takes its losses, and it tends to arrive without warning after long comfortable stretches of collecting the premium. Short options carry theoretically unlimited loss.

Which is the real volatility, HV or IV?

Both are real, they just answer different questions. HV is a fact about what already happened; IV is a price for what might. Neither is more correct — the interesting number, the realised volatility over the option's future life, is the one that has not happened yet.

How do I know if my HV and IV are comparable?

Check three things: same tenor, same annualisation convention, and a stated realised estimator. If the HV window matches the option's days to expiry, both are annualised on 252 days, and you know whether it is close-to-close or Parkinson, the comparison is fair. Miss any of those and the gap can be an illusion.

Does a shrinking gap between HV and IV mean anything?

It can be an early sign that realised volatility is rising toward implied — the calm premium is starting to compress. If realised keeps climbing and crosses implied, the comfortable regime has flipped into the one where short-vol books hurt. It is worth watching as a regime signal rather than a trade trigger.

Why can't I just sell options whenever HV is below IV?

Because HV being below IV is the normal state, not a special opportunity — it is the market charging its usual premium for uncertainty. Selling into it collects a small, frequent gain against a rare, large loss, and the rare loss is not optional. It is the price the premium was paying for all along. None of this is investment advice.

Sources & references

Last reviewed 10 July 2026. Educational content only — not investment advice.