Three measures, three comparisons
What each one holds today's number up against.
Implied volatility is the movement the market is charging for, backed out of an option's price. On its own, 35% means little: that is calm for some stocks and wild for others. So traders compare it with something, and the three common measures differ in what they compare it with.
| Measure | Compares today's IV with | Its blind spot |
|---|---|---|
| IV Rank | The highest and lowest IV of the past year | Two days decide it, so one spike can set the scale for a year |
| IV Percentile | Every day of the past year | Only relative to the stock's own history |
| IV/RV | How much the stock has actually been moving | Looks backward, and the same ratio can mean opposite things |
The first two compare options with their own past prices. The third compares them with the stock. That difference is why they're worth reading together.
IV Rank
Where today's IV sits between the past year's low and high, from 0 to 100:
IV Rank = (IV now − 52-week low) ÷ (52-week high − 52-week low) × 100
A stock whose IV ranged from 20% to 60% and sits at 30% today has an IV Rank of 25: a quarter of the way up its range.
What it's good for: it's simple, every platform shows it, and it puts stocks with very different volatility on one scale.
Where it misleads: it uses two days out of about 250. One crash week sets the high, and for the next year everything after it reads low, even when options are expensive by any other measure. On single stocks the usual culprit is earnings, below.
IV Percentile
The share of days in the past year when IV was lower than it is today. An IV Percentile of 80 means IV was lower on 80% of those days.
What it's good for: it uses every day, so one spike moves it by one day's worth. It answers "is this unusual for this stock" better than the rank does.
Where it misleads: it still only compares the stock with itself. A stock that was calm all year reads a high percentile at a level that would be ordinary anywhere else. And it needs a year of daily IV, which you can't work out from a chain on your own the way you can the rank.
When rank and percentile disagree
A made-up stock, to show how far apart they can land.
Say a stock's IV normally runs between 28% and 45%, but going into its last earnings report it reached 80%. Today it's at 38%, with the next report more than a month away.
| Range used | IV Rank at 38% | Reads as |
|---|---|---|
| The full year, 28% to 80% | about 19 | cheap |
| Without the earnings spike, 28% to 45% | about 59 | on the rich side |
The percentile sits much closer to the second reading, because the spike was a handful of days. When the two disagree, an old spike squashing the rank is the usual reason. The combinations:
| Rank | Percentile | What it usually means |
|---|---|---|
| High | High | Expensive against its own year by both measures. Worth asking why: an event is often coming. |
| Low | Higher | One spike set the top of the range. The percentile is the fairer reading. |
| Low | Low | Cheap against its own history. Selling pays little; buying costs relatively little. |
| High | Lower | Uncommon. IV sat higher than today for most of the year and one brief dip set the low. |
The IV/RV ratio
What the market charges for movement, against how much the stock is actually moving.
Realized volatility is how much the stock really moved, measured the same way implied volatility is quoted: the standard deviation of daily returns, annualized. Over the last 21 trading days, roughly a month, it lines up with a one-month option's implied volatility.
IV/RV = implied volatility ÷ realized volatility over the past 30 days
Above 1, options are charging for more movement than the stock has shown lately. Below 1, the stock has been moving more than the options charge for, which usually happens during a selloff.
What normal looks like on the index
Implied volatility has usually come in above what followed. From 1990 to mid-2024, the VIX averaged 19.59 while the S&P 500's realized volatility over the next 30 days averaged 15.50; the medians were 17.77 and 13.12. On the averages that's a ratio of 1.26, on the medians 1.35. So a ratio around 1.3 on the index isn't a bargain. It's normal.
A second study, on S&P 500 options since 1996, found implied above realized on 83.7% of days, by 3.82 volatility points on average. The other days include the ones that hurt sellers most: in the 1998 LTCM panic realized ran about 22 points above implied, and in the 2020 pandemic crash about 68.
That gap is what an option seller is paid for. The lesson on why there is a premium at all covers the reasoning.
The twist: forward against backward
Both studies compare implied volatility with the next 30 days, which nobody knows on the day they trade. The ratio a trader can actually work out divides by the past 30 days. Those aren't the same thing, and the difference shows up in two ways.
The same ratio can mean opposite things. Two made-up readings:
| IV | Realized | Ratio | Gap | |
|---|---|---|---|---|
| After a scare | 25 | 16 | 1.56 | 9 points |
| Dead calm | 13 | 8 | 1.63 | 5 points |
The calm market has the higher ratio and the cheaper options: a single ordinary move erases a five-point cushion. Read the gap in points beside the ratio, not the ratio alone.
A quiet month inflates it. Realized volatility can collapse in a few calm weeks while implied volatility stays where it was, so the ratio rises because nothing happened, not because options got richer.
The window matters too. Ten, twenty and thirty days give different numbers, and platforms rarely say which they use. Compare like with like.
Earnings, and why the index is cleaner
Single stocks report four times a year, and IV climbs for a couple of weeks into each report, because the nearest expiration includes the event. After the report it drops, often sharply. That distorts all three measures:
- IV Rank takes the earnings peak as the year's high, so between reports most stocks look cheaper than they are.
- IV Percentile is affected less, but the run-up lasts weeks, so it isn't earnings-free either.
- IV/RV jumps before a report, because implied prices the coming move and the past 30 days don't contain one. That's an event priced in, not options being rich.
Some data providers publish IV with the earnings effect taken out. An index fund has no single report date, which is one reason all three measures behave better on SPY than on any one stock.
What none of them tell you
- Direction. High IV is a price for movement either way.
- Whether the risk is worth the premium. A stock can carry high IV because something real is coming: a trial verdict, a guidance cut, a takeover rumour. Rich and dangerous often arrive together.
- Whether a rule built on them works. These describe today's price. Using one to decide when to trade is a separate claim, and it needs its own test.
For a buyer the same readings run the other way: low IV against its own history makes options cheaper to buy, which isn't the same as making the trade profitable.
What this site found when it tested a rule
The covered-call study here tested ten rules for when to write a call on the index, including the volatility risk premium (VIX less the past month's realized volatility, the gap in points above) and VIX against its own running median. Most rules that looked good on the whole window came apart when the sample was cut in half. The one that held in both halves, writing only when VIX was above its median, could not be told apart from luck.
The full results, with the method →
That's the reason this page describes the measures and stops there. They're a good way to see whether an option is expensive. Whether trading on that pays is a question for a backtest, with a control.
To see the implied volatility in one contract, type its price into the trade builder: it shows the volatility that price implies.
Sources
- Derek Horstmeyer, Alex Handley and Ana Claudia F. Manuel, How Well Does the Market Predict Volatility?, CFA Institute Enterprising Investor, 31 July 2024. VIX against the S&P 500's realized volatility over the following 30 days, from 1990. The ratios above are computed from its averages and medians.
- Brett Friedman, How Accurate Is Implied Volatility?, OptionMetrics, 2026. S&P 500 options from 1996, 7,655 trading days. Its table dates the pandemic miss 02/2000; the event is the 2020 crash.