Unusual Options Activity: What the Flag Measures, and One Friday of Data
Unusual options activity is trading in an option contract that is large next to its normal level, usually meaning the day's contracts traded exceed the open interest carried in from the day before. Scanners flag it as a sign of a big, informed bet, but the flag alone cannot show who bought, who sold, or why.
Scanners that promise to show where “big money” is betting have made this phrase familiar to anyone who trades options. The idea is simple and the display is persuasive: a contract that normally trades a few hundred times suddenly trades tens of thousands. This page explains what that number actually measures, and then checks a full Friday of real exchange data to see what most flags turn out to be.
How it works
Every listed option series has two running counts. Volume is the number of contracts traded today. Open interest is the number of contracts still open, that is, created and not yet closed, exercised or expired.
The two are measured differently. The Options Industry Council explains, on a page read on 25 Sep 2026, that the OCC calculates open interest by consolidating the exchanges’ reports of each day’s opening and closing trades. So open interest moves once a day, after the fact, while volume builds through the session.
What a scanner flags
Unusual activity compares the two. When a series trades more contracts in one session than were open at the start of it, a scanner flags it. Some also require a minimum size, and others compare against the series’ own average volume instead.
The appeal is the inference. If far more contracts traded than existed, the reasoning goes, someone must be opening a large new position, and a large new position suggests someone knows something. The rest of this page tests how much of that holds.
What the four order types hide
Every option trade has a buyer and a seller, and each side is either opening or closing. The OIC sets out the arithmetic: if both sides open, open interest rises by the size of the trade; if both close, it falls; if one opens and the other closes, it does not change.
A traded count records none of that. A 20,000-contract print could be a fund buying puts to open, a fund selling puts to open, or two firms passing an existing position between them.
Direction is hidden too. A trade at the asking price hints at an eager buyer, but the other side may be a market maker who hedges at once in the shares. The flag measures size, not intent.
And the same contract can trade many times in a day. A contract bought at 10am and sold at 11am counts twice in volume and zero times in the next day’s open interest. On a busy day, one series can turn over its whole open interest several times without anyone holding a new position overnight.
A worked example
Take one real series from 25 Sep 2026: the QQQ put with a $725 strike expiring 30 Oct 2026. Cboe’s data showed 39,074 contracts traded against open interest of 2,018 at the start of the day. With QQQ at $745.40, the strike sat 2.7% below the price, with 35 days to run.
The ratio is 39,074 divided by 2,018, or 19.4 times. It passes the test this page uses by a wide margin, and a put this size is easy to read as a large bet on a fall.
What the data cannot say. It cannot show whether those puts were bought or sold, or whether they were opened as protection on a large holding of the fund rather than as a bet. A buyer of puts who owns QQQ shares is hedging, not predicting a fall.
A second case is starker. The QQQ call with a $975 strike expiring 19 Mar 2027 traded exactly 12,000 contracts against open interest of 385, a ratio of 31.2 times, with the strike 30.8% above the price and 175 days to run. A round number like 12,000 may mean one or a few large orders, but this data cannot confirm even that.
The honest next step is the next day’s open interest. If the count for the $725 put rises by thousands, positions were opened and are still held. If it barely moves, the day was mostly contracts passing back and forth.
The original data
The data: every listed option on ten of the most traded names, SPY, QQQ, IWM, AAPL, NVDA, TSLA, AMD, AMZN, META and MSFT, from Cboe’s delayed quotes after the close on Friday 25 Sep 2026. That is 64,358 series, and 36,620,072 contracts traded in them that day.
The test used here: at least 500 contracts traded, and more than the opening open interest. 1,987 series passed. Between them they held 31,565,881 contracts, 86.2% of everything traded in these ten names, so on this Friday the unusual was the norm. The per-name counts are published as a CSV of flagged options.
Most of the flags were short-dated. 309 of the flagged series expired that same afternoon, 1,286 within the next seven days, 285 in 8 to 30 days and 107 later than that.
By contracts, the tilt is stronger. 75.5% of the flagged contracts were in options expiring that day and 21.7% in the next week. Series with 8 to 30 days left held 1.7%, and those with more than 30 days left only 1.2%.
Why same-day options dominate. Across all ten names, 66.6% of the day’s contracts were in series expiring that day, the 0DTE options that now trade every weekday on the big index funds. Their open interest at the start of the day is small next to the trading they draw, so they trip the flag almost by design.
The share differs by name. Flagged series held 90.7% of the contracts traded in SPY and 90.1% in QQQ, but 63.7% in AMD, which had the lowest share of the ten. In between came TSLA at 88.5%, AAPL at 83.3%, NVDA at 81.1%, MSFT at 80.3%, META at 75.5%, IWM at 69.3% and AMZN at 69.2%. No name came in below 60%, so the pattern is not confined to the index funds.
Calls and puts were close to even: 958 flagged series were calls and 1,029 puts. On this day, at least, the flags did not lean one way.
Video coverage is small. In the 24,971-video corpus this site studies, 3 titles cover unusual options or options flow, from 3 channels, at a median of 1,604 views; the most watched reached 2,313. Each video is counted once.
When it fails
It flags the calendar, not conviction. On the Friday measured, three quarters of the flagged contracts expired within hours. A rule that fires on most of a day’s trading cannot separate one informed trade from the rest.
It cannot see the other side. For every contract bought, one was sold. A flag that reads as a big bet on a fall is equally a big bet by someone else that the fall will not come.
It cannot see hedges. Institutions buy puts against stock they hold and sell calls against it. Both look like directional bets on a scanner and are the opposite of one.
It arrives late. By the time a flag appears on a feed, the trade is done. Copying it means trading after the price may already have moved, and paying the bid-ask spread again on the way in.
And a single day is a single day. The figures above cover one Friday with a weekly expiry, which favors same-day contracts. A midweek day would likely show a different mix, and this page does not claim otherwise.
Related
The open interest page explains the count every flag is measured against and what it can say on its own. The 0DTE options page covers the same-day contracts that made up most of the flags in the data above. And the options page sets out how trading them differs from trading shares, including the clock that runs against every position.
Before reading anything into a flagged option, look at its expiry date first. If it expires this week, assume it is short-term trading until the next day’s open interest says otherwise. The ones worth a second look are further out, and even those need the next day’s open interest before they mean anything.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.