Volume vs open interest in options: how to read the two numbers

What each number actually counts

Volume is a count of contracts traded during the session. It starts at zero when the market opens, increments with every transaction, and resets overnight. It says nothing about whether those trades opened positions or closed them.

Open interest is a count of contracts that exist — positions opened and not yet closed, exercised, or expired. It is a net figure reconciled by the clearing house (the OCC for US listed options), not a running tally your broker computes as trades print. It changes once per day.

The distinction matters because a single trade can move open interest three different ways depending on what the two sides are doing. Volume records the same +1 in all three cases, which is exactly why volume alone tells you so little.

Open interest can also move with no trade at all. Exercise, assignment, and expiration retire contracts without printing any volume, which is why the two numbers are not two views of the same event.

• Buyer opening, seller opening → open interest rises by one contract. New exposure was created.
• Buyer closing, seller closing → open interest falls by one. Exposure was retired.
• One side opening, the other closing → open interest is unchanged. The position simply transferred to a new holder.

Open interest is stale while you are looking at it

Open interest figures are cleared and reconciled after the session ends, then published before the next one opens. That means the open interest sitting on your option chain at 11am is a snapshot of positions as of the previous close. It does not tick.

The consequence is direct: none of today's opening trades appear in open interest until tomorrow morning. Every intraday V/OI number you see is a live numerator over a settled denominator — an estimate of pressure, not a measurement of new positioning.

This does not make the ratio useless. It changes what the ratio is entitled to claim. Intraday, V/OI answers "how large is today's activity relative to the position base that already existed before today," which is a genuinely useful question. It does not answer "how much new positioning was created today," and treating it as if it does is the most common error in reading flow.

Data vendors also differ in when they refresh the figure, and some display a preliminary number before the reconciled one lands. It is worth knowing when your own source updates open interest, because a chain that refreshes at a different hour than you assume will quietly shift every ratio you compute.

The four combinations and what each suggests

Once you have two consecutive sessions, you can compare today's volume against the day-over-day change in open interest. That pairing is far more informative than either number alone, and it collapses into four readable cases.

Two conditions apply to all four. They are next-morning reads, because the open-interest side of the comparison has not been published yet during the session. And none of them identify direction: a rise in open interest tells you contracts were created, not which side of them is the one with a view.

• High volume, open interest up sharply — new positioning. This is the only one of the four where new exposure was demonstrably created and survived the session.
• High volume, open interest flat or down — churn, closing, or transfer. A 10,000-contract day that leaves open interest unchanged built nothing; it moved existing contracts between holders or unwound them.
• Low volume, open interest up — quiet position building. Little traded, but most of what did trade was opening. Easy to miss precisely because the headline volume is unremarkable. It still says nothing about who is long the new contracts.
• Low volume, open interest down — interest draining out. Positions retired ahead of expiry, contracts exercised or assigned, or a strike going stale. Note that the exercise and assignment cases reduce open interest without any trade appearing in volume at all.

Why V/OI is the pressure read

V/OI is simply volume divided by open interest. Its value is that it is scale-free: it lets a contract with 300 open interest be compared against one with 200,000 without the larger name dominating purely because it is larger.

A hypothetical to make it concrete. A contract with 500 open interest that trades 5,000 times has a V/OI of 10 — today's activity is ten times the standing base. The identical 5,000 contracts against 250,000 open interest is a V/OI of 0.02, and tells you essentially nothing. Same volume, opposite conclusion.

The bands below are the labelling convention ConvexRadar uses in its interface — not laws of the market, and not the scoring formula itself, which treats V/OI as a continuous input rather than a set of steps. Treat the bands as triage for deciding what deserves a second look.

• Under 1x — today's activity is smaller than the position base that already exists. Usually routine.
• 1x to 2x — activity comparable to the standing base. Worth noting, rarely worth anything on its own.
• 2x to 5x — meaningfully out of line with the existing base.
• Above 5x — far beyond the base. Check the denominator before you read anything into it.
• Two failure modes to guard against: a tiny denominator inflates the ratio into nonsense (500 contracts against 3 open interest is roughly 167x and means nothing), and the ratio is blind to dollars. Always read it beside the premium behind the prints — a cheap far-dated contract can produce a spectacular ratio on a trivial amount of capital, while a large, expensive near-dated position may barely register.

The overnight open-interest check is the only confirmation

Because intraday V/OI cannot separate opening from closing, the honest confirmation step is to look at tomorrow's open interest. That is the only place where "someone built a position" and "someone traded a lot" come apart.

Hypothetically: a contract trades 6,000 on 800 open interest. If the next morning open interest reads roughly 5,900, most of that volume was opening and was held overnight. If it reads 900, the day was churn. The scanner row looked identical in both cases at 3pm.

ConvexRadar's Whale Map open-interest surge pins are built on exactly this comparison. The scanner retains a compact daily chain snapshot per trading date, then compares each strike's open interest against the previous retained snapshot. A pin is drawn only when the increase clears both an absolute floor of 500 contracts and a relative test of at least 50% above the prior figure — so a modest jump on an already-enormous strike does not qualify, and neither does a large percentage move off a handful of contracts.

Two caveats belong with the method. The read arrives a day late, which is the trade you make for knowing the positioning survived the session instead of guessing from a ratio computed against yesterday's base. And the comparison is against the previous retained snapshot rather than a guaranteed consecutive session, so a gap in the retained history widens the window a single pin covers.

Common misreadings

Most bad flow analysis comes from a small set of repeated mistakes. They are worth naming individually because each one feels reasonable in the moment.

• "Volume means buying." Every traded contract has a buyer and a seller. Volume is activity, never direction.
• "High V/OI means an institution is opening a position." It might be closing one, rolling one, or trading one leg of a spread whose other leg implies the opposite view.
• "The high open-interest strike is the target." Open interest accumulates across a contract's entire life, so a large figure there may be months old. Every open contract also has a short side, and much of that side is market-maker inventory hedged in the underlying rather than a directional bet.
• "Big ratio, therefore something is happening." On thin open interest the ratio is arithmetic noise. Apply a floor first — ConvexRadar flags contracts under 10 open interest explicitly for this reason.
• Reading ticker-level totals instead of the contract. Calls and puts, hedges and speculation, near-dated and far-dated all collapse into one meaningless line.
• Misreading expiration week. Open interest falls broadly as contracts expire, get closed out, or are exercised, and an expiring series goes to zero open interest whether or not anyone traded it. That is the calendar, not sentiment.
• Assuming the open-interest change identifies who initiated. The net figure is anonymous by construction.

What volume and open interest cannot tell you

The central limitation is that chain data does not identify the initiator or the intent. When a contract trades, the print does not record whether a fund opened a bullish bet, a market maker hedged unrelated exposure, or someone closed a short they had carried for months. Volume rises identically in all three. You cannot recover buyer-versus-seller intent from volume and open interest, and any tool that asserts it with confidence is inferring, not observing.

ConvexRadar's own limits, stated plainly: there is no licensed full options tape here and no dark-pool feed. Print typing — the block, sweep, split, and print labels — is derived from chain characteristics, specifically contract volume, open interest, V/OI, and estimated premium value, and it is labelled as a proxy in the interface. It is a reasonable derivation, not a look at the raw tape.

Open interest is also a net figure, not a census of who holds what. Multi-leg trades distort any single-leg reading, since a spread lands as volume and open interest on two different strikes. Exercise and assignment reduce open interest without any trade printing at all.

ConvexRadar is research software, not financial advice. Volume and open interest are one research input among several. Nothing on this page is a prediction, a signal, or a recommendation, and no combination of these two numbers guarantees direction or follow-through.

Using it in practice

A workable routine treats V/OI as the first filter and never the last word. The sequence below is roughly how the scanner is designed to be used, but it works with any data source.

The final step is the one most people skip. Logging the row and revisiting the open interest the following morning is the only way to find out, in your own universe of tickers, how often standout intraday pressure actually resolves into positioning that was held. Without that step you never learn whether the rows you liked were opening or churn.

• Start at the contract, not the ticker. Strike, expiration, and side are the information; the ticker total throws all of it away.
• Apply floors before ratios: a minimum absolute volume, a non-trivial open interest, and a bid/ask spread you could realistically trade.
• Read V/OI next to premium value. Pressure tells you something happened; premium tells you what it cost to make it happen, and the two frequently disagree.
• Add implied-volatility context. The same ratio into an earnings window with elevated IV is a materially different setup than in a quiet tape, and a known catalyst carries IV-crush risk that the ratio says nothing about.
• Check the catalyst window — earnings, filings, analyst changes, news — for a reason the positioning would exist at all.
• Write the row down, then check that contract's open interest the next morning. Confirm or discard.

Where ConvexRadar shows this

Every contract row in the scanner carries volume, open interest, and the computed V/OI side by side, along with premium value, IV rank, delta, gamma per premium, days to expiration, bid/ask spread, and catalyst context such as upcoming earnings. V/OI feeds the row's score on a continuous curve rather than in steps, and it also gates the signal-quality pill: the top "Prime" label requires V/OI at or above 2x alongside conditions on score, IV rank, gamma per premium, volume and premium value; "Clean" requires at least 1.5x; "Watch" requires at least 1x. Below 1x the row carries an explicit "weak fresh-flow read" risk flag rather than being ranked as if the ratio were fine.

The unusual volume scanner is the view built specifically around volume standing out against open interest, with side bias and repeat pressure separated from one-off prints. The broader options flow scanner ranks the same contract rows against premium, IV, and catalyst context. The Whale Map is where the overnight test lives — open-interest surges drawn from ConvexRadar's own retained daily chain history, pinned on a price chart alongside disclosed Senate EFD filings, which carry their own filing lag of up to 45 days and are reported as amount ranges rather than exact sizes.

If you want the surrounding concepts — the Greeks, implied volatility and IV rank, risk management, catalysts, and the scanner workflow itself — the options education center covers them. None of it is advice, and the data limits are printed where they apply.