Does triple witching move stocks?
Myth. Since 1983, triple-witching Fridays moved a little less than other Fridays, not more: 0.65% against 0.73%. In the last 20 years: 0.70% against 0.80%.
Triple-witching Fridays bring bigger price moves than other Fridays.The other way: a bit calmer, in both halves, the last 20 years and most industries. Could be luck.
Stocks do worse on triple-witching Fridays than on other Fridays.Slightly lower on average, but the first half went the other way, and the likely range includes zero.
Stocks do worse in the week after triple witching than in the week after other Fridays.Softer in both halves and the last 20 years, but the likely range includes zero.
What do people believe about triple witching?
People believe triple witching brings wild price swings, a falling market on the day, and a weak week after.
Four times a year, on the third Friday of March, June, September and December, three kinds of contracts end on the same day. They are stock options, stock-index futures and stock-index options.
Traders call it triple witching, and the name does a lot of work. It sounds like a night when anything can happen.
The story goes back to the mid-1980s. A 1987 study found unusual swings in the last hour of trading on expiry days. A long-running trading almanac adds that the week after, especially after June, tends to be weak.
What exactly did we test?
We tested three exact claims, comparing every triple-witching day with every other Friday.
Why only Fridays? Fridays have habits of their own, so the fair match for a witching Friday is an ordinary one.
We used the whole US stock market from June 1983, the first triple witching, to August 2026. That's 173 witching days against 2,007 other Fridays.
Claim one: the day's move, up or down, is bigger. Claim two: the day's return is lower. Claim three: the next five trading days are weaker.
Returns for claims two and three are measured above what cash in Treasury bills earned. Gaps between two averages are in percentage points, or points for short.
Do witching days bring bigger moves?
No. The witches mostly stayed home: witching days moved a little less than other Fridays, not more.
The average move on a witching day, up or down, was 0.65%. On other Fridays it was 0.73%. That's a gap of -0.09 points, with a likely range of -0.18 to +0.02 points.
Chart data
| Period | Triple-witching days | Other Fridays |
|---|---|---|
| All years | 0.65% | 0.73% |
| First half 1983–2005 | 0.62% | 0.69% |
| Second half 2005–2026 | 0.67% | 0.78% |
| Last 20 years | 0.70% | 0.80% |
The last 20 years look the same, over 80 witching days. They moved 0.70% on average and other Fridays 0.80%, a gap of -0.10 points.
Witching days were calmer in both halves of the period too. Only 9 of the 49 industries we checked saw bigger moves on witching days.
So the claim points the wrong way. The likely range just crosses zero, so we can't even be sure witching days are calmer. We can say they haven't been wilder.
Do stocks fall on witching days?
Not reliably. Witching days came out about flat against cash (0.00%), while other Fridays averaged +0.04%.
The gap of -0.04 points has a likely range of -0.19 to +0.10 points, which includes zero. In the first half, witching days actually did slightly better. In the last 20 years they did a bit worse, by 0.12 points. Over the whole period, the day was weaker in 41 of 49 industries. Industries move together, though, so that isn't strong proof.
Is the week after weaker?
A bit, but not reliably enough to pass our test. This one is a close call.
Over the next five trading days, the market's return over cash averaged -0.08% after a witching day and +0.19% after other Fridays. That's a gap of -0.27 points.
Chart data
| Period | Week after triple witching | Week after other Fridays |
|---|---|---|
| All years | -0.08% | +0.19% |
| First half 1983–2005 | -0.09% | +0.16% |
| Second half 2005–2026 | -0.07% | +0.22% |
| Last 20 years | -0.01% | +0.22% |
The gap pointed the same way in both halves (-0.25 points and -0.29 points), in the last 20 years (-0.23 points) and in 47 of 49 industries. Industries all move with the market, though, so these aren't independent votes.
But the likely range of the gap runs from -0.62 to +0.09 points, which includes zero. One week's return swings a lot, so even a gap this size could be luck. Our verdict rule needs that range to stay clear of zero. It didn't, so luck can't be ruled out, and the claim stays a myth.
The almanac's version looks at June only. After June expiries the gap was -0.29 points, in line with the claim, but from about a quarter of the weeks.
And here's a twist. Before June 1983, when nothing expired together, the week after these same Fridays was soft too, a gap of -0.20 points. So any softness may not be about expiring contracts at all.
So is anything true about triple witching?
Possibly, but in a place our data can't reach. Older studies found the action in the last hour. We see one price per day, at the close, and the expiry rush happens at the open and near the close.
Trading volume isn't in our data either, so we can't test the heavy trading people talk about.
The busy last hour is also why the main index futures switched to settling at the opening price in June 1987. Since then, the gap in the size of moves was -0.11 points.
The before-1983 check tells the same story. Before June 1983, when nothing expired together, the same Fridays had a gap of +0.05 points. They were, if anything, a touch livelier with no witches at all.
Why does it feel true? The name is spooky, every quarter brings headlines about a record expiry, and one wild witching day is easy to remember.
What does this mean for you?
This is history, not a forecast. Over four decades, witching Fridays looked a lot like any other Friday at the close.
A myth doesn't mean you should do the opposite. Calm witching days aren't a signal either: the gaps are small and could be luck.
We also judged three claims here, which gives luck three chances to fool us. None of them passed.
What's in the small print?
Data: the US total stock market and one-month Treasury bills from Kenneth R. French's data library (Tuck School of Business at Dartmouth), daily from July 1926 to August 2026, and 49 US industry portfolios from the same library. The main tests start in June 1983; the before-1983 check uses 1926 to June 1983.
Method: a witching day is the third Friday of each quarter's last month. When the market is shut that Friday, we use the trading day before. Two moved to a Thursday: March 2008 (Good Friday) and June 2026 (Juneteenth).
Move size uses raw daily returns, before taxes. The day's return and the week after are measured above T-bills. Halves split the witching days by count. The likely ranges are 95% bootstrap intervals.
Test many beliefs and about 1 in 20 with nothing behind it will pass the usual 95% standard by luck alone. So the main ranges are also shown at 99.67%, which allows for testing 15 calendar beliefs.
At 99.67%, the move-size gap's range widens from -0.18 to +0.02 points to -0.23 to +0.08 points. The day's return goes from -0.19 to +0.10 points to -0.26 to +0.18 points, and the week after from -0.62 to +0.09 points to -0.77 to +0.28 points.
Against all other trading days, not just Fridays, witching days moved 0.65% against 0.74%. Before June 1983, the same Fridays' daily return gap was -0.14 points. Since June 1987, the week-after gap was -0.32 points.
What else did we test?
See all the calendar effects we tested, or how we test a myth.