Is This Volatility High Enough to Sell
Rank and percentile disagree. Both are useful, and a high reading is sometimes a warning.
Worth reading first: Is This IV High? Rank, Percentile and Skew
Course catalogue99 lessons
“Sell when volatility is high” is the most repeated instruction in premium selling, and the least specific. High relative to what? Measured how? And does a high reading actually predict anything?
Two different measures answer the first question differently, and they routinely disagree on the same day about the same underlying.
Rank and percentile on the same data
Change the shape of the year's volatility history and watch the two readings diverge. A single spike moves rank enormously and percentile barely at all.
52-week range: 20.6% to 62.3%
One spike in the year stretches the high, which drags IV rank down for every subsequent day. While IV percentile, which only counts days, barely notices. That is why the two figures can point in opposite directions on the same chain.
Two measures, two questions
IV rank asks where today sits between the year's low and high.
Rank = (IV − low) ÷ (high − low) × 100IV rank: position within the range.
IV percentile asks how much of the year was spent below today.
Percentile = days with IV below today ÷ total days × 100IV percentile: position within the distribution.
The difference is not academic. Rank depends on only two data points (the extremes) so a single panic day a year ago permanently depresses every subsequent reading. Percentile uses every observation and ignores the extremes entirely.
| History | Rank | Percentile |
|---|---|---|
| Ranged 15–35 all year | 50 | ~50 |
| Ranged 15–20, one spike to 80 | 15 | ~95 |
| Ranged 20–30, mostly near 28 | 50 | ~25 |
Today's IV is 25. Same underlying, same day, two very different signals.
In the middle row, rank says volatility is low and percentile says it is near the highest all year. Percentile is telling the truth; rank is being distorted by one day.
Thresholds people actually use
There is nothing magic in these numbers. They are conventions, and their value is consistency rather than precision.
| Reading | Typical stance |
|---|---|
| Below 30 | Premium thin; smaller size or stand aside |
| 30 to 50 | Ordinary conditions, normal size |
| 50 to 75 | Favourable; the usual target zone |
| Above 75 | Rich, but ask why before selling more |
A common framework for premium selling.
The important row is the last one, and it is the one most frameworks get backwards.
Why a very high reading is often a warning
Implied volatility is a forecast, not a mispricing. When it spikes, the market is anticipating movement, and it is frequently right.
A reading above 90 usually has a cause: an earnings print, a regulatory decision, a merger, a company in genuine trouble, or a broad market panic. Selling into that without identifying the cause means being paid handsomely to take a risk you have not examined.
The distinction that matters is market-wide versus name-specific. Broad stress lifts every chain and generally mean-reverts. That is the classic environment for selling premium. A single name at 95 while the market sits at 30 is idiosyncratic, and the market has usually identified something you have not.
What the reading does not tell you
Direction. High IV says a large move is expected, not which way.
Whether it is actually overpriced. The premise of selling premium is that implied volatility tends to exceed subsequent realised volatility. That is a long-run tendency, not a per-trade guarantee, and it fails exactly when the anticipated event materialises.
Whether the chain is tradeable. A high reading on an illiquid chain is not an opportunity, see the liquidity checklist.
Whether the underlying suits you. Volatility is one filter among several; selection criteria come first.
A practical sequence
Choose the underlying on its own merits. Check the chain is tradeable. Then read percentile and rank together, and only then look at the premium available.
Reversing that order (screening on volatility and then justifying the underlying) is how people end up wheeling companies they would never otherwise have considered.
What can go wrong
Using rank alone. One spike a year ago distorts it for twelve months.
Treating a high reading as free money. It is a forecast, and forecasts are often right.
Not distinguishing market-wide from name-specific. Only one of those tends to mean-revert benignly.
Screening on volatility first. It inverts the selection process.
Key takeaways
- Rank measures position within the year's range; percentile measures position within its distribution, they routinely disagree.
- Percentile is more robust because rank depends on only two data points and is distorted by a single outlier.
- Use percentile as the primary reading and rank as a sanity check; investigate when they diverge sharply.
- A reading above 90 is usually a warning, not an opportunity. Find the cause before selling into it.
- Broad market stress tends to mean-revert; a single name at extreme volatility is usually idiosyncratic.
Check your understanding
1. A stock traded IV 15–20 all year except one spike to 80. Today it is 25. What do rank and percentile say?
2. IV percentile reads 95 on a single stock while the broad market sits near 30. What should you do?
3. What does a high implied volatility reading tell you about direction?
Related lessons
Track what you have worked through. No account needed. View transcript
This browser only