Implied Volatility, and How to Tell Whether a Reading Is High

The Question Everyone Actually Asks

There are three honest comparisons, and they are not equally available.

Against its own recent history. This is the comparison behind the figures usually published as IV Rank and IV Percentile. Both need roughly a year of daily closing readings. Rank asks where today sits between the year's lowest and highest; percentile asks what share of days closed below today.

Against today's own range. Far weaker, but available immediately: where the current reading sits between this session's own high and low. The catch is that the range mechanically widens as the day goes on, so the same position means different things at 10am and at 3pm. This workspace shows it with that caveat attached rather than as a headline.

Against what actually happened afterwards. The comparison that decides whether options were expensive, and the only one that cannot be made in advance.

A note on why this site does not currently print an IV Rank. Rank depends on exactly two numbers — the window's highest and lowest readings. With a short window you have probably not observed a real spike, so the highest is too low, so the rank reads too high. Worse, it does not settle down as you add data: one volatile week entering the window can move a rank from 90 to 15 without today's reading changing at all. A rank computed from a few weeks is a different statistic wearing a familiar name, and publishing it under that name would be the misleading option, not the helpful one.

Why the Call and the Put Show Different Numbers

Open an exchange option chain and the call and the put at the same strike will usually show different implied volatilities — sometimes by a fraction of a point, sometimes by a great deal. This has been read for years as a directional signal. It is not one.

So when two figures disagree, one of the inputs is wrong — and in practice it is almost always the same input: the centre the calculation is anchored to.

The size of the disagreement is a measurement of how far off that centre is. On a weekly expiry, one volatility point of gap corresponds to only a handful of points of error in the underlying's forward price. That makes the gap a very sensitive diagnostic, and a useless signal — it is telling you about the calculation, not about the market's opinion.

The second cause is liquidity. At any strike, one of the two legs is in the money and the other is out of it. The in-the-money leg is mostly intrinsic value, trades in a wider spread, and its price barely moves when volatility changes — so inverting its price for a volatility is numerically unstable, and a stale quote on it produces a large and entirely fictional gap.

This site therefore solves one volatility per strike, anchored to the forward, and shows the call-versus-put difference separately as a data-quality reading with the forward error it implies. A stale quote manufacturing an apparent signal is a documented failure mode, not a hypothetical one.

The Centre Is the Forward, Not Spot

"At the money" sounds like it means "at the current price". For option pricing it does not. The centre of the distribution an option is priced against is the forward — where the market prices the underlying at expiry — and that sits above spot by the cost of carrying a position until then.

The gap is not a rounding detail. On a monthly expiry it is comfortably more than one strike interval on a major index, and it grows with time to expiry. Reading "at-the-money implied volatility" at the strike nearest spot therefore reads the volatility curve at a point systematically off its centre, and picks up a slice of the curve's slope along with it.

The forward can be recovered from the option prices themselves, using the same arbitrage relationship as the previous section, without needing a futures quote. That is what this site does.

Turning a Percentage Into Points

Implied volatility is quoted as an annual figure. Almost nobody holds an option for a year, which is why the raw number is so hard to act on.

The conversion is one line:

Expected move = underlying level × implied volatility × √(days to expiry ÷ 365)

The square root is the part that surprises people. Four times the time does not give four times the expected move — it gives twice. Volatility scales with the square root of time, which is also why a week-out option is not worth a quarter of a month-out one.

This workspace prints the result directly, so the reading can be read as points rather than as a percentage. What that figure means: it is a one-standard-deviation move, so roughly two sessions in three finish inside a band that wide, and about one in three does not. The arithmetic says nothing whatsoever about which side of the range the move lands on.

Two cautions on the figure. It is symmetric, while the distribution it approximates is not — the approximation is good near the money and degrades in the tails. And it describes the move to expiry, not the move today.

What an Intraday Reading Actually Shows

Watching implied volatility through a session shows something a closing figure cannot: whether an option's price moved because the underlying moved, or because the market repriced movement itself.

That distinction is the one most often missed. A premium going from 174 to 207 could be the underlying moving in your favour, or volatility rising, or both — and if it then falls back, it matters a great deal whether the underlying reversed or volatility simply drained away. The premium alone cannot tell you. The volatility reading alongside it can.

Two patterns are worth knowing before reading a session chart.

Around known events. Volatility tends to be bid up ahead of a scheduled event and to fall sharply once the event has passed and the uncertainty is resolved — regardless of which way the underlying went. An option held through such an event can lose value even when the direction was right, which is the single most common source of "the stock moved my way and I still lost money".

Across weekends and holidays. Time to expiry here is counted in calendar days, matching the convention used by the exchange's own volatility index and by brokers generally. That keeps figures comparable — but it means non-trading days are counted at full weight, so readings tend to sag into a long weekend and recover after it, without the market having done anything. It is a convention artefact, and worth recognising rather than trading.

Why Two Platforms Show Different Numbers for One Option

The same option, at the same moment, will show different implied volatilities on different services. None of them is necessarily wrong. Implied volatility is not observed — it is solved for, and the answer depends on the inputs chosen.

The interest rate. Every implied volatility needs one, and there is no single correct choice. Exchanges and vendors use different rates, and a different rate on the same option price gives a different volatility. This site uses 6.5%. An exchange publishing its own figure at a materially different rate will print a different number for the same contract, and neither is an error.

The centre. Spot or the forward, as covered above. This alone can shift an at-the-money reading by a meaningful fraction of a point, and shifts which strike gets called "at the money" at all.

The price used. Last traded price, or the midpoint between bid and ask. Last traded price can be minutes stale on an illiquid strike; the midpoint reflects the current book but can be wide.

Whether the two legs are averaged. Some platforms publish separate call and put figures; some publish one. Averaging two separately-solved figures is worse than it looks — where the centre is off, the errors on the two legs do not cancel away from the money, and the average traces out a curve that looks like a real volatility skew and is entirely an artefact of the averaging.

The practical consequence: comparing an implied volatility figure across platforms is not meaningful unless both state their method. Comparing a figure against itself over time, from one consistent method, is meaningful — which is the comparison this workspace is built for.

Where the Reading Stops Being Worth Much

Four situations degrade the figure, and this site marks each rather than publishing a number that looks the same as a good one.

A settled series. Once a contract has expired, its prices no longer describe anything live, and the arithmetic on them degenerates — time to expiry heads to zero and the solved volatility runs away. Readings on a settled series are refused outright rather than shown.

The last hours before expiry. Sensitivity to volatility collapses as expiry approaches, which means the smallest possible price change corresponds to a larger and larger change in the implied figure. Late on an expiry day, a single tick can move the reading by a substantial fraction of a point. The number is still computable; it is simply no longer precise, and it should not be read as though it were.

Illiquid strikes. A volatility can only be solved where something actually traded, at a price that responds to volatility. Deep in-the-money options are mostly intrinsic value and barely respond, so their prices do not identify a volatility at all. This site declines to return a figure in that case rather than returning an arbitrary one.

Gaps in the session record. A minute with no snapshot is a minute with no reading. Such minutes are left out of the line rather than filled in, and the count is reported alongside it.

The general principle is the one the rest of this site follows: a figure that could not be measured is named as such, never rendered as a zero or a bare dash. A blank reads as "nothing is happening" when the truth is "this could not be computed", and those are very different statements.