VIX Explained: What a High VIX Means, From 35 Spikes Since 1990

TL;DR: the answer box

• The VIX is the options market's guess at the next 30 days of S&P 500 movement, written as a yearly percentage. Divide by the square root of 12 for one month: a VIX of 20 means about 5.8%, up or down.
• High, by Chase's line, is common. Chase calls 20 to 25 high, but 37.0% of days since 1990 closed above 20. The median close is 17.59. Only 6.9% closed above 31 and 2.2% above 40.
• Spikes are common and short. 35 spikes from the first close above 31 to the first close below 24. Median length 22 trading days. 25 of 35 were over within two months.
• The spike comes after the fall. When the VIX first closed above 31, SPY was already a median 11.8% below its 52-week high, in 32 of 32 spikes.
• The next month was strong, with pain along the way. From the next open, SPY averaged +2.83% over a month against +0.82% for all days. Only 0.5% of random-date samples matched it. The average worst point inside that month was -5.33%.
• The next year was ordinary. Twelve months on, the spikes averaged +8.08% against +10.06% for all days, and 74.3% of random-date samples did at least as well.

What is the VIX index?

The VIX is one number. It tells you how much the options market expects the S&P 500 to move over the next 30 days. Cboe calls it "a leading measure of market expectations of near-term volatility" and has published it since 1993 (Cboe, VIX Index). Volatility here just means how big the price swings are, up or down.

People call it the "fear gauge". The nickname comes from Robert Whaley's 2000 paper (Whaley, "The Investor Fear Gauge", Journal of Portfolio Management).

Here is the easiest way to read it: four zones.

The lines at 24 and 31 are the two this study uses for "calm" and "high". The line at 15 is where Chase's guide puts "low".

Think of the VIX as a smoke detector. It tells you the kitchen is smoky right now. It does not tell you whether the house burns down, and it goes off loudest after the toast has already burnt. Every chart below tests that picture.

How to read the VIX number

The VIX is a yearly number. To turn it into a one-month move, divide it by the square root of 12, about 3.46.

So a VIX of 20 means a move of about 5.8% over the next month. A VIX of 40 means about 11.5%. That is one standard deviation: a "typical" big move, not the worst case.

How I measured it

The data. Cboe's official daily VIX history, 1990-01-02 to 2026-08-27: 9,260 daily closes. SPY daily from 1993-02-02. E-mini S&P 500 futures (ES) from 2007-01-03, adjusted so percentage returns are real.

What counts as a spike. A spike starts on the first daily VIX close above 31. It ends on the first later close below 24.

The timing. The VIX closes at 4:15 pm ET, after the stock market closes at 4:00 pm. So you only see a spike after stocks have closed. Every S&P 500 number here is decided at the VIX close and measured from the S&P 500's next open. That is Execution rule v2's timing: decide at the close, act at the next open.

The yardsticks. Two of them. First, the same measurement from every SPY day, so "after a spike" is always compared with "any day". Second, a random-date check: pick as many random SPY days as there were spikes, 10,000 times, and count how often random dates did as well.

One thing you cannot do. You cannot buy the VIX itself. It is an index. This study measures the S&P 500 and uses the VIX as a reading. Nothing here trades.

What is considered a high VIX?

Above 31 is a top 6.9% day. Here is how the days split across the four zones.

Chase's guide calls a VIX "above 20 to 25" high and "below 15" low (Chase, What is the VIX?). On the data, 37.0% of all days since 1990 closed above 20. More than one day in three. That is not rare.

The quiet side is just as common. 32.0% of days closed below 15, and the median close was 17.59.

The extremes are real, though. The highest daily close was 82.69 on 2020-03-16. The highest reading during a day was 89.53 on 2008-10-24. The lowest close was 9.14 on 2017-11-03.

How often does the VIX spike?

Often. 35 spikes since 1990, starting in 21 different years. And at every one SPY's history covers, SPY was already below its 52-week closing high, by 2.6% to 26.9%.

Not every year gets one.

21 of 36 full years from 1990 to 2025 had at least one close above 31. So "almost every year" overstates it. The calm runs are long: five years in a row from 1992 to 1996, and again in 2004 to 2006 and 2012 to 2014.

How many spikes you count depends on the ruler. Move the start and end lines, or read monthly bars instead of daily closes, and the count moves a lot.

Same 31 and 24 lines: 35 spikes on daily closes, 11 on month-end closes. When someone quotes you a spike count, ask which ruler.

How long does a high VIX last?

Weeks, usually. The median spike lasted 22 trading days, 1.1 months.

16 were over in under a month and 9 more within two. The one long exception ran from 2008-09-15 to 2009-07-21: 213 trading days, 10.2 months. That one also held the deepest SPY drop of any spike, -45.13%.

I entered the markets in 1998, at the dot-com peak. That October the VIX closed at 45.74, and that spike lasted 3.5 months. Most do not last that long. Some last much longer. Plan for both.

Where was the S&P 500 when the VIX crossed 31?

Already down. Every time.

When the VIX first closed above 31, SPY was a median 11.8% below its 52-week high. The range ran from 2.6% to 26.9% down. ES futures showed the same median, 11.8%.

That is the smoke detector going off after the toast has burnt. By the time the alarm sounds, the fall has started. How much further it goes, the alarm does not say.

What happens to the S&P 500 after a VIX spike?

The next month was strong. The next year was ordinary.

A month after a spike, SPY averaged +2.83%. Any day averaged +0.82%. But look at the right-hand column. The average worst point inside that month was -5.33%. Over 12 months it was -15.27%. The gains came with real dips.

The odds of being up barely moved. A month after a spike, SPY was up in 21 of 32 spikes, 65.6%. Any day: 64.1%. A bigger average, nearly the same coin flip.

Was it luck? For the next month, probably not. For the next year, it looks like luck.

At three and six months, random sets matched the mean 6.1% and 8.2% of the time. At 12 months, 74.3% of random sets did at least as well, and 97.6% were up as often.

The 12-month numbers carry a warning. The three worst all started between November 2007 and March 2008: -40.92%, -38.12% and -34.93%. Three early alarms, one bear market. A spike on its own did not tell you whether it was a scare or the first leg of a crash.

ES futures, from 2007, agree. A month after a spike, ES averaged +2.24% against +0.85% for all days. Twelve months after, +6.73% against +10.32%.

From spike to calm: what the S&P 500 did

Now measure each spike from the open after it started to the open after the VIX closed back below 24. SPY ended higher in 30 of 32, 93.8%. Mean +5.20%, median +4.76%.

Before you read that as "buy the spike", read the dots. A drawdown is the biggest drop from a high point. Inside a spike, it was under 5% in 18 spikes, 5% to 15% in 11, 15% to 25% in 1, and 25% or more in 2: -45.13% from 2008-09-15 and -28.74% from 2020-02-27.

And the endpoint is the first close below 24. Nobody can date that in advance. The two that ended lower were 2008-09-15 to 2009-07-21 (-18.98%) and 2022-04-26 to 2022-06-08 (-1.89%).

Here is one spike, read exactly the way the study reads all of them.

The VIX closed at 45.31 on 2025-04-04, with SPY already 17.6% below its 52-week high. SPY opened at $489.19 on 2025-04-07. The VIX peaked at 52.33 the next day. It closed at 22.68 on 2025-05-02, 19 trading days after the spike. From open to open: +15.00%, with a worst drop of -6.33% along the way.

What does a low VIX mean?

Calm, and not much more. Here is every SPY day since 1993, sorted by the VIX close the evening before.

A low VIX meant fewer dips. Below 15, a 10% dip within a month followed on 1.5% of days. At 40 and up it was 35.1%. That is the VIX doing its job. It measures risk, and the risk showed up.

A low VIX was not a warning about the year ahead. Below 15, the next 12 months averaged +11.17% and were up on 86.4% of days.

The +33.05% at 40 and up is not a gift. Those 208 days come from just 33 stretches across 10 years. A few crises supply all of it, and each crisis counts dozens of times. The weakest 12-month row was 20 to 25, at +4.26%.

Does the VIX predict how much the market will move?

It usually overshoots.

On 83.2% of SPY days, the VIX was higher than the swings the S&P 500 actually delivered over the next 21 trading days. The median VIX was 17.67. The median that followed was 13.39.

The gap was widest when the VIX was high: a median 8.31 points with the VIX at 30 and up, against 3.87 below 20. So on SPY since 1993, measured against the next 21 trading days, the VIX was an upper-end estimate of next month, not a fair one.

Can you trade the VIX?

Not the index. You cannot buy the VIX itself, and the futures and funds that track it carry their own costs and quirks. I tested short VIX futures rules separately in this VIX trading strategy study.

What the index did show, on SPY, is that VIX levels grouped next-month dip risk. Whether that helps your own system is a test to run on it. A volatility filter is the obvious one. If you want a structured way to build and test one, that testing workflow is what the Algo Trading Masterclass teaches.

The honest limits

• Small samples. 32 spikes on SPY and 22 on ES. They cluster in a few years, and the 12-month windows of late 2007 and 2008 cover the same bear market.
• One headline rule. Other rules give 10 to 50 spikes.
• No S&P 500 result for 1990 and 1991. SPY starts in 1993, so 3 of the 35 spikes are counted but not measured.
• Measurement, not trading. No costs, no dividends. This does not test a strategy that buys spikes or filters trades with them.
• Overlap in the VIX-level table. Neighbouring days share most of their window, so day counts overstate the evidence.

What this means for you

1. Check how rare a reading is before you call it high. Above 31 is a top 6.9% day. Above 20 happened on 37.0% of days. A VIX of 20 is above the median, not an alarm.
2. Expect the spike after the drop. The S&P 500 was already a median 11.8% off its high when the VIX crossed 31. Selling at that point meant selling after the fall, every time in this sample.
3. Expect it to pass, and plan for the one that drags on. 25 of 35 spikes were over within two months. One lasted 10.2 months and saw SPY fall 45.13% along the way. Ask whether your system could sit through that one.
4. Read the VIX as next month's risk. It sorted 10% dips well, from 1.5% of days below 15 to 35.1% at 40 and up. Pair any return you see with the dip it came with.

The VIX is a smoke detector. Respect the alarm. Don't sell the house.

Reference tables

For readers who want the detail behind the charts.

VIX daily close percentiles, 1990 to 2026

SPY after a VIX spike, from the next open

Mean dip: the average of each spike's lowest low inside the window, below the entry open. The last spike, 2026-03-27, has a 1-month result only.

How many SPY days sit behind each VIX level

Stretches: unbroken runs of days in the band. Years: calendar years that added days. Fewer stretches means thinner evidence.

If you want the next study like this before anyone else, it goes out in the Overfit newsletter. StatOasis.com/Overfit

Methodology: StatOasis research engine 1.21.0. Cboe official daily VIX history 1990 to 2026, loaded through the engine's data layer. Spikes start on the first daily close above 31 and end on the first close below 24. S&P 500 outcomes measured on SPY (1993 to 2026) and ES futures (2007 to 2026) from the next open after the VIX close, Execution rule v2's timing, against every SPY day and a seeded random-date control. Measurement only: no trades, no costs, no dividends. The VIX index itself cannot be traded. Past behaviour of an index is not a promise about its future. Research, not investment advice.