VIX at 17.31, 90th Percentile on the 2026-10-20 Close: When Calm Is Relative
A middling VIX on the October 20 close ranked high against its recent past. Here is why level and rank are different questions.
Published 2026-10-23 · Data as of 2026-10-20 · Market & data intelligence · Educational, not advice.
On the 2026-10-20 close, the VIX settled at 17.31 (FRED). That is a modest level, yet it ranked in the 90th percentile of the prior 60 sessions by Delta Arc's calculation. The same day, the term structure held in contango, the yield curve sloped upward and the Baa credit spread printed 1.44 percentage points (FRED). The lesson: a level and a rank answer different questions, and a careful read uses both.
The paradox in the October 20 close
As of the 2026-10-20 close, the VIX settled at 17.31, according to FRED. On its own, that reads as an unremarkable number. It was not a panic print, and it was not a sleepy one either.
For the same session, Delta Arc's calculation on the FRED series ranked that close in the 90th percentile of the prior 60 sessions. Put the two facts side by side and you get the interesting part of the tape: a middling absolute level that was high compared with its own recent history.
The gap between absolute and relative is the concept worth taking from this snapshot. These are end-of-day settles, which FRED publishes with a lag, and this post was written after that close. Nothing here describes the market as you read it.
Absolute level versus percentile rank
A VIX reading is a level of annualized implied volatility, backed out of S&P 500 option prices. It tells you how much movement option buyers were paying to protect against over the next 30 days. That is an absolute measure. A given number means the same thing whenever it prints.
A percentile rank answers a different question: compared with the last 60 sessions, where does this close fall? A 90th percentile reading means the close was higher than most of the window it was measured against. It says nothing about whether the level is high by long-run standards.
That is why both numbers matter. On 2026-10-20, the absolute reading of 17.31 (FRED) pointed to calm. The relative reading, the 90th percentile (Delta Arc's calculation), pointed out that most of the preceding window had been calmer still. Neither cancels the other. Read together, they describe a close that was elevated against a quieter recent base. It was not a market in disorder.
What the term structure added
The VIX measures 30-day implied volatility. Its three-month counterpart, VIX3M, settled at 18.99 on the same date (FRED). Because the longer-dated measure was above the shorter one, Delta Arc's classification labeled the term structure as contango, the normal and calm state.
Contango means option buyers that day priced more uncertainty further out than in the near term. Backwardation is the inverted case: near-term implied volatility above the longer-dated reading, which is the shape of a market pricing stress as immediate. The 2026-10-20 close did not show that shape. The 30-day reading sat high within its recent window, but it stayed below the three-month measure.
The rates and credit side of the same session
Volatility is one dial. The bond market is another, and on 2026-10-20 it printed a normal shape. On that close, the 2s10s spread was +0.33 and the 10-year minus 3-month spread read +0.86, both per FRED. Delta Arc's classification of the curve state that day: positive, a normal slope.
Underneath those spreads, on the same 2026-10-20 close, the 10-year Treasury yield was 5.12 percent and the 3-month bill 4.26 percent (FRED).
A positive curve is the textbook arrangement: lenders ask for more yield to lock money up for longer. Inversion, where short rates exceed long rates, is the shape that draws headlines. On 2026-10-20, neither measure showed it.
Credit spreads complete the picture. The Baa corporate spread measures the extra yield investors demand to hold medium-grade corporate debt instead of comparable Treasuries. It closed at 1.44 percentage points on 2026-10-20 (FRED). Spreads widen when lenders grow nervous about default and narrow when they relax. A single print does not tell you which way they were moving. It tells you the level the corporate bond market settled on that day.
How to read a mixed dial
Snapshots like the 2026-10-20 close invite one of two errors. The first is to anchor on the absolute VIX and call it quiet. The second is to anchor on the 90th percentile and call it alarming. Each throws away half the information.
The disciplined read keeps every dial in view. On that date, volatility was elevated against its recent window but in contango. The yield curve was positively sloped on both common measures. On 2026-10-20, credit printed a spread of 1.44 percentage points. That is a description, not a forecast. It tells you what configuration the market was in, not where it went next.
The snapshot also shows why the length of a percentile window matters. A 60-session lookback is short by design. Because it reacts quickly to recent change, it is good at flagging when a market has moved off its base and poor at judging whether a level is high over the long run. Pair a short window with an absolute level and you get both views.
A simple checklist for any single-day snapshot:
- Level: what did the series actually settle at?
- Rank: where did that settle fall within its recent window?
- Shape: did the term structure or the curve confirm the level or contradict it?
- Credit: did the corporate bond market agree with the equity options market?
What members see on top
The public numbers above are the raw tape. Delta Arc members get the next layer: the base-rate history of configurations like this one, where the VIX was elevated against its 60-session window while the term structure stayed in contango and the curve stayed positive. Members also get our read on how much weight that history deserves. We do not publish those odds here.
The next edition of The Dial will compare later closes against this snapshot. The question worth watching is whether a 90th percentile reading in contango was a step toward something, or a single-session wobble that the window absorbed. Come back to see what the tape printed.
Sources
The readings in this post are end-of-day closes from these public series. Percentiles and term-structure labels are Delta Arc calculations on them.
References for this topic
- FRED, Federal Reserve Bank of St. Louisfred.stlouisfed.org
- CBOE Volatility Index (VIXCLS)fred.stlouisfed.org
- 10-Year minus 2-Year Treasury spread (T10Y2Y)fred.stlouisfed.org
- Moody's Baa yield relative to the 10-Year Treasury (BAA10Y)fred.stlouisfed.org
- Nominal broad U.S. dollar index (DTWEXBGS)fred.stlouisfed.org
This is the free read. Delta Arc members get the base-rate odds and the specific read built on top of it. See the plans or get on the early-access list.