Two measures, two questions
IV rank = (current IV − 52w low) ÷ (52w high − 52w low) × 100
IV percentile = days IV was lower ÷ trading days in the window × 100
Rank asks: where does today sit between the highest and lowest readings of the year? It uses two numbers — the extremes — and ignores everything in between.
Percentile asks: on how many days was volatility lower than it is now? It uses the whole series and ignores how extreme the extremes were.
They are routinely quoted as if interchangeable. They are not, and the case where they diverge is common.
Where they disagree, and why it matters
Suppose volatility spiked to 90% for three days last October and has spent the rest of the year between 18% and 30%. Today it is 28%.
- IV rank = (28 − 18) ÷ (90 − 18) = 14%. Looks cheap.
- IV percentile ≈ 85%. Looks expensive.
Both are correct. Rank is anchored to a single event that no longer describes the market; percentile says volatility is higher than it has been on most days this year. For deciding whether to sell premium, percentile is the more useful answer here — and rank is what most platforms display.
The general rule: one spike sets the range permanently and drags rank towards zero for a year. Whenever rank looks surprisingly low on a market that feels active, check when the high was set.
This calculator computes rank from the three numbers you have and percentile only if you supply the day count, because percentile genuinely requires the full series. Estimating it from the extremes would be inventing a number, which is worse than leaving the row empty.
How it is used
| Rank | Conventional reading |
|---|---|
| Above 70 | Premium expensive — favours selling |
| 40–70 | Middle of its own range |
| 20–40 | Below average |
| Under 20 | Premium cheap — favours buying |
The logic is mean reversion: volatility is one of the few market quantities that reliably returns towards its own average, so an unusually high reading is more likely to fall than to keep rising. That is the basis for selling iron condors and credit spreads into high rank and buying long options into low rank.
Two honest limits. Volatility is mean-reverting until a regime changes, and a rank of 95 immediately before a crisis is not a sell signal, it is an early reading of a move that is still starting. And rank is relative to one instrument's own history, so a rank of 80 on a quiet utility may still be a lower absolute volatility than a rank of 20 on a small-cap.