Triple Witching: What 135 Quarterly Expirations Did to SPY
Triple witching is the quarterly day when stock index futures, stock index options and stock options all expire together: the third Friday of March, June, September and December. Volume swells as positions are closed or rolled, and SPY's record since 1993 shows what that did and did not do to prices.
Triple witching is the day each quarter when three kinds of contract expire together: futures on stock indexes, options on stock indexes, and options on individual stocks and ETFs. It falls on the third Friday of March, June, September and December, and the name comes from the three expirations landing at once.
The same day is sometimes called quadruple witching. Writers count the contract types differently; the date is the same either way. What matters for a trader is what the day does to trading, and that can be measured.
How it works
The calendar is written into the contracts. CME’s specifications for the E-mini S&P 500 list quarterly contracts in March, June, September and December, and state that trading “terminates at 9:30 a.m. ET on the 3rd Friday of the contract month.” Cboe’s specifications for SPX index options give the expiration date as “the third Friday of the expiration month.”
Settlement happens at the open, not the close. Cboe says the SPX settlement value is calculated from “the opening sales price in the primary market of each component security on the expiration date.” So the prices that settle those contracts are set in the first minutes of that Friday’s session.
Holders have three choices before then. A trader with an expiring futures contract can close it, let it settle in cash, or roll it into the next quarter by selling the old contract and buying the new one. Option holders face the same choice, plus exercise. Rolling is why the volume shows up: a large share of open positions is closed and reopened in a short window.
Hedges come off at the same time. Dealers who have sold options usually hold shares or futures against them, and when the options expire there is nothing left to hedge. That unwinding lands on the same few hours as the rolling.
Where the open positions sit
Most option positions are parked on third Fridays. In Cboe’s delayed SPY option chain, taken after the close on 25 September 2026, 31 future expiration dates carried 17,323,989 contracts of open interest. The largest single date was 18 December 2026, the next quarterly expiration, with 2,538,613 contracts, 14.7% of the total. The monthly date of 16 October 2026 was close behind at 14.5%.
Third Fridays held 67.9% of it. That counts the eleven listed third-Friday dates plus one Thursday: the June 2027 contracts are listed for 17 June, the day before that month’s third Friday. When the Friday is a market holiday, expiration moves back a day, and the chain shows it.
This concentration is the reason the day matters at all. The chain lists daily and weekly dates as well, yet the quarterly and monthly dates carry the bulk of the positions that have to be dealt with.
A worked example
18 September 2026, the most recent quarterly expiration. SPY traded 65,395,100 shares that day. The median volume of the 20 sessions before it was 40,157,500 shares.
- Volume ratio: 65,395,100 / 40,157,500 = 1.63 times the recent norm.
- The day before, 17 September, SPY traded 49,652,800 shares. The next session, 21 September, it traded 50,484,700.
- High $762.00, low $757.97, against a previous close of $762.60. The range was ($762.00 - $757.97) / $762.60 = 0.53%.
- SPY closed at $761.69, down 0.12% on the day.
So the extra volume was large and the price barely moved. Five sessions later, on 25 September, SPY closed at $771.35, up 1.27% from the expiration close.
The June quarter looked different. The third Friday, 19 June 2026, had no SPY session, so this study uses Thursday 18 June. Volume was 80,875,700 shares, 1.60 times its 20-session median, and five sessions later SPY had fallen 2.38%. Two quarters, two opposite weeks: the rest of this page is about which of them was typical.
The original data
How each day was classified. SPY daily bars from 29 January 1993 to 25 September 2026. A quarterly expiration day is the third Friday of March, June, September or December, or the session before it when that Friday has no session (twice: 20 March 2008 and 18 June 2026). Monthly expirations follow the same rule. Volume ratio is that day’s volume divided by the median volume of the 20 sessions before it. Range is high minus low over the previous close. Every expiration day, with its ratio, range and next-week change, is in the expiration-day file.
Volume: higher, and more so lately. Across 135 quarterly expirations, the median volume ratio was 1.25. On the 268 monthly expirations in the other eight months it was 1.08, on 1,297 other Fridays 1.03, and on all 8,049 non-expiration days, other Fridays included, 0.99. By decade the quarterly median ran 0.92 in the 1990s, 1.03 in the 2000s, 1.37 in the 2010s and 1.51 in the 2020s so far. The 1990s figure is below one: in that decade, the quarterly day was often a quieter session than usual for SPY.
The biggest bulges were rare. 39 quarterly days traded at least 1.5 times normal and 17 at least twice normal. The largest was 16 March 2001, at 5.24 times. Only 22 of the 135 beat the volume of every one of the 20 sessions before them, and only 9 were the heaviest session of their three-month quarter.
Range: no wider. The median high-to-low range on quarterly expirations was 0.91% of the previous close, against 1.07% on other Fridays and 1.05% on all non-expiration days. The median close-to-close move was 0.58%, against 0.55% and 0.53%. The extra shares changed hands without a bigger day.
The week after is where the record stands out. SPY closed lower five sessions after 79 of the 135 quarterly expirations, 58.5%. After monthly expirations the share was 39.2%, after other Fridays 43.4%, and after all non-expiration days 41.6%; the daily group file holds every day’s ratio and next-week change. Against the 41.8% of all 8,447 days with 20 sessions of history and a full five-session window, a two-sided binomial test gives p = 0.0001. The median next-week change after a quarterly expiration was -0.23%.
It is not steady across time. By decade, the share of lower weeks was 50.0% (14 of 28) in the 1990s, 72.5% (29 of 40) in the 2000s, 60.0% (24 of 40) in the 2010s, and 44.4% (12 of 27) in the 2020s so far. Most of the effect sits in the twenty years from 2000. In the 2020s the pattern broke: 44.4% is close to the 41.8% of an ordinary week, and a binomial test puts it well within chance of that rate (p = 0.85). Monthly expirations show nothing similar.
Why the week after might lean
No single cause is measured here. A common explanation is that buying tied to hedges against expiring options stops once the options are gone, so support that was there into expiration is missing the following week. This page does not test that mechanism. It only reports the price record.
A small p-value is not a trading edge. Over the whole span, 58.5% of 135 is unlikely to be chance alone. But it describes the past, it was measured on closing prices before any costs, and it did not hold in the 2020s. This page does not know why the lean appeared or why it faded, and it does not treat the old record as a reason to trade.
When it fails
Reading volume as direction
A heavy expiration day is not a signal. The extra volume is mostly positions being closed and rolled, and on 18 September 2026 it came with a 0.53% range and a 0.12% move. Nothing about the volume said which way the following week would go, and that week rose 1.27%.
Trading the old pattern
The 2020s broke the lean. Since 2020, SPY closed lower in only 12 of the 27 weeks after a quarterly expiration, so a bet on every one of those weeks falling would have been wrong more often than right. A calendar effect that depends on who is hedging what can change as the market’s structure changes.
Forgetting a position expires
The most expensive failure is administrative. An E-mini S&P 500 future left open at 9:30 a.m. ET on the third Friday stops trading and is settled in cash, which CME lists as its settlement method. On the same morning, Cboe sets the SPX options settlement from each stock’s opening price, whatever those turn out to be. A trader who meant to roll can find the position closed at a price nobody chose.
Assuming the date is always a Friday
Holidays move it. June 2026 expired on a Thursday, and the SPY chain already lists June 2027 for Thursday 17 June. Check the actual expiration date for each contract instead of assuming the pattern.
Related
Options expiry covers what happens to a single option contract on its last day, including exercise and assignment.
A futures contract is the instrument whose quarterly roll drives much of the extra volume, and that page explains how settlement and rolling work.
Open interest is the count behind the chain figures above, and max pain is the expiration-day theory built from it, tested on a real chain.
I mark the quarterly expiration dates on the calendar at the start of the year so a heavy-volume Friday never surprises me. If you hold futures or options that expire that day, decide whether to close or roll by Thursday, and do not read the extra volume as a signal about direction.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.