IV Percentile measures the share of days over a trailing window on which implied volatility closed below today's reading. If IV Percentile is 80, then IV was lower than it is now on 80% of the past year's trading days.
It answers the same question as IV Rank — are these options expensive for this stock? — but it counts days instead of measuring distance, and that difference matters more than it sounds.
The formula
IV Percentile = (days with IV below today) ÷ (total days in window) × 100
No arithmetic on the extremes. Just a count.
Why it beats IV Rank on messy data
IV Rank looks only at the highest and lowest readings in the window. That makes it hostage to outliers.
Take a stock that trades between 25% and 35% IV all year, except for one March afternoon when a takeover rumour spiked it to 120%. It sits at 34% today — near the top of its normal range, options genuinely expensive.
- IV Rank: (34 − 25) ÷ (120 − 25) × 100 = 9. Looks dirt cheap.
- IV Percentile: IV closed below 34% on ~92% of days. Reads 92. Correctly expensive.
The rank is not wrong arithmetically. It is answering a question you did not ask — how far is today from the extremes — when what you wanted to know was how unusual is today.
When they disagree
Divergence between the two is information, not noise. It almost always means the window contains a spike that is distorting the range.
The practical rule: when IV Rank and IV Percentile disagree sharply, trust the percentile. The rank is being dragged by a day that has nothing to do with current conditions.
When they agree, you have a clean read and can stop thinking about it.
The same trap applies
A high IV Percentile still is not a signal. It tells you options are historically expensive; it does not tell you why. Check the calendar before selling premium into it — an earnings print or a pending catalyst explains a high reading and makes the premium fair rather than generous.
Both measures are context. Neither is a trade.