IV rank and IV percentile are not the same thing
Two ways of placing a volatility reading in its own history, why they disagree, and what a single past extreme does to each.
IV rank is where the current reading sits between the lowest and highest readings of a lookback window. IV percentile is the share of sessions in that window whose reading was below the current one. They answer different questions and frequently disagree.
The two definitions
IV rank is (current − minimum) ÷ (maximum − minimum) over the window. It describes a position between two extremes. IV percentile counts how many observations in the window fell below the current reading and divides by the number of observations. It describes a position in a distribution.
On a well-behaved history the two are close. They separate as soon as the window contains an outlier.
What an outlier does
Suppose a name spends a year with 30-day implied volatility near 20%, apart from one week when it reached 200%. For the rest of that year, IV rank sits near zero: 20% is close to the bottom of a range whose top is 200%, whatever else happens.
IV percentile is unaffected by how far the spike reached. It only counts sessions, so a reading of 21% still ranks above the many sessions at 20%. Neither number is wrong. They are answers to different questions, which is why this site reports both.
When neither is reported
A percentile computed from sixty sessions is not a weaker version of one computed from three hundred. It is a different and more misleading quantity, and there is no way to caveat it on a page that a reader will remember.
Both statistics are therefore withheld until the window holds at least 180 sessions. Early in a symbol’s history the fields read as unavailable rather than as a number.