IV rank places today's implied volatility between its 52-week low and high on a 0 to 100 scale; IV percentile counts the share of the past year's days on which IV was lower. Both ask whether premium is high or low for this particular stock, and after a single volatility spike they can disagree sharply, because the spike stretches the rank's range for a year.
They answer two different questions about the same year of implied volatility readings.
A worked example shows why they part company. A stock's IV is 32 today; its 52-week low was 20 and a single spike day reached 80. IV rank is (32 - 20) / (80 - 20) = 20, which reads cheap. If IV sat below 32 on 190 of the last 252 trading days, IV percentile is about 75, which reads rich. The spike stretched the denominator of the rank for a full year, while the percentile treats every day the same, so one outlier barely moves it.
| IV rank | IV percentile | |
|---|---|---|
| Question it answers | Where is IV between the year's high and low? | On what share of days was IV lower? |
| Formula | (IV - low) / (high - low) | Days below today / days in the window |
| Sensitive to one spike | Yes, for a year | Barely |
| Moves quickly | Yes | More slowly |
The percentile method. The app keeps a store of each day's at-the-money implied volatility and ranks today against those actual observations; the figure it labels IV rank is computed that way, because a single spike day otherwise distorts the reading for a year. When only a realized-volatility proxy is available, the number is labeled differently so a proxy is never read as a measured figure.
Option premium is the price of volatility. The same 30-delta credit spread pays meaningfully more when IV is in the top decile of its range than in the bottom decile, while the width, and therefore the worst case, stays the same. Premium-selling disciplines therefore often include an IV rank floor among their entry criteria, and premium-buying disciplines an IV rank ceiling: paying up for options when volatility is already rich is a headwind that has to be overcome by the directional move.
IV rank is self-referential. It compares a stock only to its own history, which produces two classic failure modes:
Because of the first failure mode, a rank number is most trustworthy when cross-checked against a measure of the premium itself, such as the net volatility risk premium, which compares implied volatility to the stock's realized movement. When the two disagree, the premium measure is the one that describes what a seller actually collects.
No. IV percentile counts the share of days in the past year with implied volatility below today's reading, while the min-max form of IV rank scales today's IV between the 52-week low and high. The two can disagree sharply after a single extreme spike, because the min-max form keeps comparing every later reading against that one outlier. Many platforms label the percentile form as IV rank, so the label alone does not settle which formula is behind a number.
There is no universal threshold. Premium-selling rule sets commonly use floors somewhere between 30 and 50, and premium-buying rule sets often use ceilings, but any cutoff is a filter the trader chooses and owns. A reading in the top decile of a stock's own year is high by definition, though the failure modes described above explain why a high rank does not always mean rich premium.
The arithmetic is identical, but index and broad-ETF volatility mean-reverts more predictably than single-stock volatility, which can jump on company news and stay elevated. Single-stock rank readings around earnings are the least stable, because the scheduled event inflates and then deflates implied volatility on a known date.
More in Volatility: VIX Term Structure: Contango, Backwardation, and Sizing · Reading Volatility Skew: What the Smile Says About Risk