A 20% one-day jump in the VIX happened on 133 sessions from January 1990 to November 6, 2026, grouped into 74 separate episodes. The S&P 500 fell on every single one of those days, by a median 2.45%. The next session it closed higher 63.2% of the time, against 53.7% for an average day, and three months later it was up 73.3% of the time with a median gain of 4.54%.
The latest surge came on November 5, 2026, when the VIX closed at 28.18, up 69.7% from 16.61, while the S&P 500 fell 4.36%. That is the third-largest one-day VIX jump in Cboe’s data. A surge here means the VIX close was at least 20% above the previous close. Every number was calculated from the daily data listed under Sources.
How often the VIX jumps 20% in a day
Much more often than it used to. Since 2010 the rate has run about four times the 1990s rate.
| Years | Sessions | Surge days | Per year | Per 1,000 sessions | Separate episodes |
|---|---|---|---|---|---|
| 1990-1999 | 2523 | 14 | 1.40 | 5.5 | 11 |
| 2000-2009 | 2515 | 22 | 2.20 | 8.7 | 14 |
| 2010-2019 | 2516 | 57 | 5.70 | 22.7 | 27 |
| 2020-2026 | 1722 | 40 | 5.70 | 23.2 | 22 |
The 2020-2026 row counts 2026 as a full year, so the per-1,000-sessions column is the fairer comparison. An episode starts with a surge that had no other surge in the prior 21 sessions. Seven calendar years had none at all: 1992, 1999, 2000, 2003, 2004, 2012 and 2023. The busiest was 2018, with 12.
Part of the rise is arithmetic. The test is a percentage, so a VIX sitting at 12 needs only a 2.4-point move to qualify, while a VIX at 40 needs 8 points. Long calm stretches with a low VIX, like much of the 2010s, make 20% days easier to hit.
The biggest one-day VIX jumps
| Date | VIX prior close | VIX close | VIX change % | S&P 500 that day % | S&P 500 next 21 sessions % |
|---|---|---|---|---|---|
| 2018-02-05 | 17.31 | 37.32 | 115.60 | -4.10 | 2.94 |
| 2024-12-18 | 15.87 | 27.62 | 74.00 | -2.95 | 3.65 |
| 2026-11-05 | 16.61 | 28.18 | 69.70 | -4.36 | n/a |
| 2024-08-05 | 23.39 | 38.57 | 64.90 | -3.00 | 6.43 |
| 2007-02-27 | 11.15 | 18.31 | 64.20 | -3.47 | 1.30 |
| 2021-01-27 | 23.02 | 37.21 | 61.60 | -2.57 | 1.61 |
| 2021-11-26 | 18.58 | 28.62 | 54.00 | -2.27 | 4.17 |
| 1991-11-15 | 13.96 | 21.18 | 51.70 | -3.66 | 0.03 |
| 1990-07-23 | 15.63 | 23.68 | 51.50 | -1.74 | -9.41 |
| 2025-04-04 | 30.02 | 45.31 | 50.90 | -5.97 | 10.50 |
Seven of the ten biggest jumps came after 2017, and most started from a VIX under 20. Of the nine with a full month of data, only July 1990 saw the S&P 500 lower 21 sessions later, down 9.41%.
What the S&P 500 did after a 20% VIX surge
On average a bit better than normal, with the clearest difference on the next day and at three months. Returns run from the close of the surge day:
| Set | n | 1d avg % | 1d median % | 1d % up | 5d avg % | 5d median % | 5d % up | 21d avg % | 21d median % | 21d % up | 63d avg % | 63d median % | 63d % up |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| All sessions (baseline) | 9276 | 0.04 | 0.06 | 53.65 | 0.19 | 0.32 | 57.64 | 0.81 | 1.23 | 63.91 | 2.44 | 3.16 | 70.26 |
| Every surge day | 133 | 0.41 | 0.29 | 63.16 | 0.36 | 0.63 | 62.12 | 0.54 | 1.58 | 63.36 | 2.98 | 4.54 | 73.28 |
| First surge in 21 sessions | 74 | 0.36 | 0.37 | 66.22 | 0.49 | 0.64 | 60.81 | 0.82 | 1.34 | 65.75 | 2.82 | 5.00 | 75.34 |
The 21-day average after a surge (0.54%) was actually below the all-session average (0.81%), even though the median was higher. A few deep follow-through declines pull the average down.
By era
The 2010 post on this URL looked at SPY’s next five sessions after a 20% surge, using data from 1990 on, and saw a strong next-day bounce. The bounce tendency held up overall, but era by era it varies:
| Set | n | 1d avg % | 1d median % | 1d % up | 5d avg % | 5d median % | 5d % up | 21d avg % | 21d median % | 21d % up | 63d avg % | 63d median % | 63d % up |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1990-1999 first surges | 11 | 0.53 | 0.42 | 63.64 | 0.82 | 0.73 | 72.73 | 1.34 | 0.59 | 81.82 | 3.28 | 4.51 | 81.82 |
| 1990-1999 all sessions | 2523 | 0.06 | 0.05 | 53.59 | 0.29 | 0.36 | 58.98 | 1.27 | 1.26 | 65.40 | 3.84 | 3.39 | 77.84 |
| 2000-2009 first surges | 14 | 0.88 | 0.34 | 78.57 | 1.60 | 1.02 | 64.29 | -0.53 | 0.60 | 50.00 | -0.11 | 4.76 | 71.43 |
| 2000-2009 all sessions | 2515 | -0.00 | 0.05 | 52.25 | -0.01 | 0.11 | 52.21 | -0.06 | 0.63 | 56.62 | -0.16 | 0.96 | 54.51 |
| 2010-2019 first surges | 27 | 0.03 | 0.15 | 55.56 | -0.58 | -0.34 | 48.15 | 0.39 | 0.36 | 59.26 | 2.70 | 4.08 | 74.07 |
| 2010-2019 all sessions | 2516 | 0.05 | 0.06 | 54.85 | 0.23 | 0.39 | 60.69 | 0.96 | 1.40 | 67.25 | 2.68 | 3.39 | 74.88 |
| 2020-2026 first surges | 22 | 0.37 | 0.88 | 72.73 | 0.93 | 1.50 | 68.18 | 2.00 | 2.64 | 76.19 | 4.70 | 6.13 | 76.19 |
| 2020-2026 all sessions | 1722 | 0.06 | 0.09 | 54.04 | 0.28 | 0.46 | 59.17 | 1.19 | 1.74 | 67.55 | 3.87 | 4.75 | 75.59 |
In the 2010s the pattern weakened: the next day was close to a coin flip (55.6% up) and the following week averaged a 0.58% loss. Since 2020 the bounce has been stronger again. Each era has only 11 to 27 episodes, so a couple of events can swing a row.
Does the starting VIX level matter?
| Prior VIX close | n | 5d avg % | 5d median % | 5d % up | 21d avg % | 21d median % | 21d % up |
|---|---|---|---|---|---|---|---|
| 0-15 | 35 | -0.05 | 0.37 | 65.71 | 1.24 | 1.68 | 74.29 |
| 15-20 | 42 | 0.58 | 0.27 | 58.54 | -0.11 | 1.36 | 62.50 |
| 20-30 | 36 | 0.68 | 1.25 | 63.89 | 0.00 | 0.67 | 52.78 |
| 30+ | 20 | 0.04 | 1.93 | 60.00 | 1.60 | 3.43 | 65.00 |
Surges from a calm base (VIX under 15) were followed by a higher S&P 500 a month later 74.3% of the time. Surges from 20 to 30 did worst at 21 days. With 20 to 42 cases per row, read this as a description of the past, not a rule.
What the VIX itself did next
It eased quickly but rarely gave the whole jump back right away.
| Later | n | Avg VIX change % | Median VIX change % | VIX lower than on surge day | VIX back at or below pre-surge close |
|---|---|---|---|---|---|
| 1 sessions | 133 | -2.81 | -5.73 | 70.7% | 0.8% |
| 5 sessions | 132 | -7.51 | -12.32 | 75.8% | 24.2% |
| 21 sessions | 131 | -12.24 | -17.55 | 74.0% | 38.2% |
| 63 sessions | 131 | -17.91 | -25.21 | 80.2% | 55.0% |
The day after a surge, the VIX was lower 70.7% of the time, by a median 5.7%. But it was back to its pre-surge close only once in 133 cases on that next day, and only about half the time three months on. On November 6, 2026, the VIX closed at 25.64, down from 28.18 but well above the 16.61 of November 4.
VIX surges come in clusters
One 20% day roughly doubles the chance of another soon after.
| Window | Surges with a full window | Another surge followed within window | Any session: surge within window |
|---|---|---|---|
| 5 sessions | 132 | 20.5% | 6.5% |
| 10 sessions | 132 | 34.8% | 11.8% |
| 21 sessions | 131 | 44.3% | 21.2% |
| 63 sessions | 131 | 71.8% | 46.4% |
After a surge, another followed within 21 sessions 44.3% of the time; for a random session the figure is 21.2%. Still, 42 of the 74 episodes were a single surge. The longest run was the early-2020 pandemic selloff: 8 surges in an episode that began on January 27, 2020.
How to read these numbers
These are S&P 500 price-index returns (no dividends, no costs), measured close to close, and the windows overlap when surges cluster. The VIX is a calculated index of expected 30-day volatility from S&P 500 options; you can’t trade it directly, and VIX futures and ETPs behave differently. The statistics describe what followed past surges and don’t predict the next one.
For a price-only version of the same fear signal, see Williams’ VIX Fix. For the calm end of the scale, see the level-of-bullishness streaks, and for how the VIX compares with the volatility that follows, the volatility risk premium series. More on the Market Stats page.
