Probability and Statistics

IV Rank and IV Percentile

Q: Which metric should I use if I can only watch one?

IV Percentile, for statistical robustness: it does not hinge on two extreme observations but on the full distribution of the year. That said, in liquid names without recent traumatic episodes the two give very similar readings, and IV Rank is faster to interpret. The practical rule is to look at both and be suspicious whenever they differ by more than 20 points — that gap almost always reveals an isolated spike inside the window.

Q: What IV Rank is high enough to sell premium?

The most widely used reference is IV Rank ≥ 50 for opening credit structures, with ≥ 30 as an acceptable minimum in quiet markets where nothing reaches 50. Below 25 the premium collected rarely compensates the risk taken and it makes more sense to invert the structure. It is not a mechanical rule: an IV Rank of 60 with earnings two weeks away is a worse trade than an IV Rank of 35 with a clear calendar.

Q: Why does my platform show a different IV Rank than another?

Three common reasons. First, the window: 52 weeks is the convention, but some platforms use 30, 90 or 180 days. Second, which IV is taken as "the" IV of the asset: some use the interpolated 30-day IV, others the ATM of the nearest expiration, others a chain-weighted average. Third, whether closing or intraday prices are used. None is wrong; what matters is not comparing one platform’s figure against a threshold you learned on another.

Q: How does IV Rank relate to the VIX?

The VIX is a level of implied volatility — that of the S&P 500 over 30 days — while IV Rank is a relative position. They combine: a "VIX Rank" or VIX percentile applies exactly the same normalisation to the index and gives you the market-wide regime thermometer. A VIX of 18 can be high or low IV Rank depending on what kind of year it has been; that nuance is precisely what the absolute level does not tell you.

ES: IV Rank e IV Percentile PT: IV Rank e IV Percentil

The two metrics that turn a loose implied volatility reading into a tradeable signal: where today’s IV sits relative to its own past year, and what structure that implies.

Why raw IV tells you nothing

An implied volatility of 45% is neither high nor low: it is a number without context. For a regulated utility, 45% would be a historic panic; for a small-cap biotech in clinical trials, it would be unusual calm. The absolute level of IV blends two different things: the structural volatility of the underlying — how much that asset normally moves — and the current regime — whether it is more agitated or calmer today than it usually is. Only the second is tradeable, because only the second tends to revert. IV Rank and IV Percentile exist precisely to separate them: both normalise current IV against the asset’s own recent history, so that 45% in one name and 18% in another become comparable on the same 0-to-100 scale.

IV Rank: position within the annual range

IV Rank answers a geometric question: within the range IV has travelled over the last 52 weeks, where does it sit today? The formula is (current IV − 52-week low) / (52-week high − 52-week low) × 100. If a stock’s IV has oscillated between 20% and 60% over the past year and today trades at 40%, IV Rank is exactly 50: dead centre of its range. An IV Rank of 0 means IV is at its annual low; 100 means it is at its high. The virtue of IV Rank is that it takes three numbers and reads unambiguously. Its flaw is that it depends entirely on two extreme observations: the year’s high and low.

IV Percentile: frequency, not geometry

IV Percentile answers a different and statistically more robust question: what share of trading days over the past year closed with IV below today’s? If across 252 sessions IV was lower than today’s on 189 of them, IV Percentile is 75. It does not look at the range; it looks at the whole distribution. The difference becomes decisive when there has been an isolated spike. Imagine a stock whose IV spent the entire year between 18% and 24% except for a single panic day when it touched 90%. With IV at 23% today — near its usual ceiling — IV Rank comes out around 7, falsely suggesting premium is being given away. IV Percentile, by contrast, would return a high and correct reading, because almost every session of the year closed below 23%.

How each level translates into a structure decision

The practical value of these metrics is that they place the decision before the trade. With high IV Rank (≥50) premium is expensive relative to its own history: this favours selling structures with negative vega — credit spreads, iron condors, short strangles, cash-secured puts, covered calls — which profit if volatility compresses back toward its mean. With low IV Rank (≤25) premium is cheap: this favours buying structures with positive vega — debit spreads, calendars, diagonals, long straddles — which profit if volatility expands. The middle band (25–50) is no-man’s land: there the directional thesis should carry the weight of the trade rather than volatility, and low-vega structures such as moderately wide verticals are usually preferred.

The limits worth knowing before you trade them

Neither IV Rank nor IV Percentile is an entry signal on its own, and mistaking them for one is the most expensive error made with them. First, IV can be high for good reason: if there is a known catalyst ahead — earnings, a court ruling, an FDA decision — expensive premium is not an inefficiency, it is the fair price of real risk. Selling volatility there is not harvesting a risk premium, it is accepting a binary bet. Second, both metrics are backward-looking: they describe the last year and nothing more. A stock entering a genuinely new regime — a takeover bid, a sector crisis — will break its historical range and leave IV Rank pinned at 100 for weeks without that meaning volatility is about to compress. Third, the 52-week window is a convention, not a law: some platforms use 30, 90 or 180 days, and the same asset can show very different figures depending on the provider.

How it is used on the desk, day to day

The working use is as a pre-filter, not a trigger. The typical flow starts by discarding every candidate whose IV Rank is not in the zone of interest — selling premium only above 30–50, buying it only below 25 — and only then applies the criteria that actually select the trade: chain liquidity, bid-ask width, absence of events before expiration, and a thesis on the underlying. Many traders also keep an eye on the index IV Rank as a context thermometer: when the broad market sits at low IV Rank, premium-selling opportunities are scarce everywhere at once, and forcing them is the fastest route to collecting premium that does not compensate the risk taken.

IV Rank vs IV Percentile: when they diverge and why

Both normalise IV onto a 0-to-100 scale, but they answer different questions.

Aspect	IV Rank	IV Percentile
Question it answers	Where does IV sit within the year’s high-low range?	What share of sessions closed below today’s IV?
Data used	Three points: current IV, 52-week high and low	The full series of closes for the year (~252 observations)
Sensitivity to isolated spikes	Very high: one extreme day distorts the whole year	Low: one extreme day is one observation out of 252
Ease of calculation	Immediate, can be done mentally	Requires the full historical series
When to prefer it	Clean IV ranges with no recent traumatic episodes	After a crash, violent earnings or any one-off spike

Frequently Asked Questions

Which metric should I use if I can only watch one?

IV Percentile, for statistical robustness: it does not hinge on two extreme observations but on the full distribution of the year. That said, in liquid names without recent traumatic episodes the two give very similar readings, and IV Rank is faster to interpret. The practical rule is to look at both and be suspicious whenever they differ by more than 20 points — that gap almost always reveals an isolated spike inside the window.

What IV Rank is high enough to sell premium?

The most widely used reference is IV Rank ≥ 50 for opening credit structures, with ≥ 30 as an acceptable minimum in quiet markets where nothing reaches 50. Below 25 the premium collected rarely compensates the risk taken and it makes more sense to invert the structure. It is not a mechanical rule: an IV Rank of 60 with earnings two weeks away is a worse trade than an IV Rank of 35 with a clear calendar.

Does IV Rank work the same for indices and single stocks?

It works, but reads differently. Indices have far more stable and mean-reverting IV because they average hundreds of names whose idiosyncratic moves cancel out, so their IV Rank rotates between extremes more regularly and is more reliable as a regime signal. A single stock can stay pinned at IV Rank 100 for months if it has entered a special situation — a takeover, an existential lawsuit — and there the metric loses predictive value entirely.

Why does my platform show a different IV Rank than another?

Three common reasons. First, the window: 52 weeks is the convention, but some platforms use 30, 90 or 180 days. Second, which IV is taken as "the" IV of the asset: some use the interpolated 30-day IV, others the ATM of the nearest expiration, others a chain-weighted average. Third, whether closing or intraday prices are used. None is wrong; what matters is not comparing one platform’s figure against a threshold you learned on another.

Does an IV Rank of 100 guarantee volatility will fall?

No. It guarantees only that IV is at its highest level of the past 52 weeks, and highs exist to be exceeded. Volatility mean-reverts on average and over time, not in every individual episode: during a genuine crisis IV can rise from an IV Rank of 100 and keep rising for weeks. That is why selling premium at extremes demands defined risk; anyone selling naked at IV Rank 100 is betting this spike resembles the historical average of spikes, and sometimes it does not.

How does IV Rank relate to the VIX?

The VIX is a level of implied volatility — that of the S&P 500 over 30 days — while IV Rank is a relative position. They combine: a "VIX Rank" or VIX percentile applies exactly the same normalisation to the index and gives you the market-wide regime thermometer. A VIX of 18 can be high or low IV Rank depending on what kind of year it has been; that nuance is precisely what the absolute level does not tell you.