Open Interest Explained: What It Measures, and Why a Quarter of It Is Not Directional

Key takeaways

• Open interest counts contracts outstanding, each one once, not trades. It rises only when a new buyer and a new seller create a contract, and falls when both sides close.
• It is not volume. Volume resets daily and counts activity. Open interest carries forward and counts commitment.
• The part almost nobody accounts for: a large share of open interest is spread positions, where the same trader is long and short at once. These express a view on the shape of the curve, not on direction.
• Across the 35 markets COTInsight covers, on the report dated 1 September 2026, spread positions averaged 29.0% of open interest, with a median of 26.4%.
• The range is enormous: SOFR at 70.5% and natural gas at 54.6% against Dow Jones at 1.9%. Treating open interest as a conviction gauge without checking the spread share means comparing markets that are not comparable.
• A share-of-open-interest ratio can move several points without anyone trading, because the denominator moves. Always check whether open interest changed before reading a ratio as new positioning.

What Open Interest Actually Counts

A futures contract comes into existence when a buyer and a seller agree to one. It stops existing when both sides close out or the contract is delivered or expires. Open interest is the number of those contracts currently outstanding.

The crucial detail is that each contract is counted once, not twice, even though it always has two sides. If open interest in corn is 2,594,329, there are 2,594,329 contracts alive, each with a long holder and a short holder. The long side and the short side of the whole market are always exactly equal. This is why "there are more sellers than buyers" is never a valid explanation of anything in futures.

Four things can happen when two parties trade, and only two of them change open interest:

So open interest measures whether money is entering or leaving the market, while price measures what it costs to be in it.

Open Interest Is Not Volume

These get confused constantly, and they answer different questions.

Volume is how many contracts changed hands during the session. It resets to zero every day. A single contract traded back and forth fifty times adds fifty to volume and nothing to open interest.

Open interest is how many contracts are outstanding at the end of the session. It carries forward. It only changes when positions are created or extinguished.

The useful shorthand: volume is activity, open interest is commitment. A market can be extremely busy and see no change in open interest at all, which tells you the day's activity was traders passing existing positions between each other rather than new money arriving.

One practical note. Open interest is published with a lag, typically the following morning, because the exchange has to reconcile the day's clearing. Volume is available immediately. If you are looking at an open interest figure on a live chart, check whether it is the confirmed number or an estimate.

The Part That Gets Missed: Spread Positions

Here is the thing that changes how you should read every open interest number you have ever looked at.

A trader who is long the December contract and short the March contract in the same market holds a spread. They are not betting on the price going up or down. They are betting on the relationship between two delivery months, which is a view on storage costs, seasonality and the shape of the forward curve. If the whole market rallies, a perfectly balanced spread makes nothing.

Both legs of that position sit inside open interest. The CFTC reports spread positions explicitly in the Commitments of Traders report, in their own column, for each category of trader. Almost nobody looks at that column.

We took every market COTInsight covers and measured spread positions as a share of total open interest, on the report dated 1 September 2026. All reported spreading, across every trader category, over total open interest. Eighteen of the 35 are shown here, spanning the range:

Across all 35 covered markets the mean spread share was 29.0% and the median 26.4%.

What that means in practice. In the median major futures market, roughly a quarter of the open interest expresses no opinion about direction whatsoever. In natural gas it is more than half. In three-month SOFR, the largest contract on the entire board at nearly 24 million contracts, it is over 70%: that market's open interest is mostly a structure trade, and reading it as a directional commitment to the path of rates would be badly wrong.

And the spread between markets is the real lesson. Comparing open interest conviction in Dow Jones futures, where 1.9% is spreads, to natural gas, where 54.6% is, is comparing two different quantities that happen to share a name.

The same caution applies inside the positioning data. Our natural gas guide makes this concrete: roughly half of the Managed Money book in gas is calendar structure, so the headline net position describes a much smaller directional bet than it appears to.

The Classic Open Interest Rules, and an Honest Caveat

Most trading education presents a four-way table combining price direction with open interest direction:

The logic is sound as far as it goes. Rising price on rising open interest does mean new money is taking the long side rather than old positions being unwound.

We are not going to tell you it has been validated, because we have not tested it. It is widely repeated folklore with a plausible mechanism, and a plausible mechanism is not evidence. We publish forward statistics only where we have measured them with enough independent observations to survive a t-test, and this is not one of those cases. Treat the table as a way to describe what happened, which it does well, rather than as a way to predict what happens next.

There is also a real complication in applying it. Because spread positions sit inside open interest, a rise in open interest can be entirely calendar-spread activity around a roll, with no directional new money at all. The rule assumes the change is directional. In natural gas or SOFR, that assumption is usually false.

What you can do properly is check the composition. The COT report gives you the long leg, the short leg and the spreading leg separately for each trader category. Our grain guide walks through a live case: corn's record fund long was built as much on shorts capitulating, from 261,095 down to 62,002 contracts, as on new longs arriving, and the net number alone hides that completely. Open interest rising 11.5% at the same time is what confirmed new money was genuinely entering.

The Denominator Trap

This one costs people real money and is trivially avoidable.

A great many published figures are ratios with open interest on the bottom: the concentration ratios, the swap dealer share, spread share, any "percent of open interest" number. Open interest moves for reasons that have nothing to do with the numerator, especially around contract roll and expiry.

A live example from the same report. In COMEX silver, the Swap Dealer short as a share of open interest went from 36.15% to 41.12% in one week, a jump of nearly five points that would look like aggressive new selling:

The dealers sold nothing. They reduced their short slightly. The entire move in the ratio came from open interest falling 13%. Anyone who read the 41% as fresh dealer selling drew a conclusion the data does not support. The silver guide has the fuller version.

The rule: whenever a share-of-open-interest figure moves, look at the numerator and the denominator separately before you interpret it.

How to Use Open Interest With Positioning Data

Open interest on its own is a weak signal. Paired with the COT report it becomes genuinely useful, in three specific ways.

As a confirmation of an extreme. A z-score at an extreme while open interest is rising means the crowd is still committing. The same z-score on falling open interest means the position is being wound down. Level plus flow beats level alone.

As a discount on the headline. Before you treat a net position as a directional bet, check the spreading column. In a market where half the book is spreads, the directional exposure is roughly half what the net figure implies.

As a divergence check. Price making new highs while open interest falls means the move is being carried by shorts closing rather than longs opening, and the buying pressure has a natural end point. That is one of the setups our divergence article covers.

What open interest will not do is time anything on its own. It is context, and it is the specific context that stops you from misreading a positioning number.

Where COTInsight Fits

Inside the app, open interest sits beside the positioning data for around 350 weekly-reporting markets rather than in a separate tab: the z-score, the 156-week COT Index, the regime state, the open-interest trend and the divergence check are on one screen, so "is this extreme being funded by new money" is a glance rather than a research project.

Ultimate adds the historical outcome statistics per z-score bucket, the twenty-year archive, the weekly AI commentary, and the TradingView indicator so the same panels sit on your chart.

What it is not: no intraday data, no entry signals, no free tier. COT is weekly data with a three-day lag, built for swing and position timeframes.

Market pages referenced above: natural gas, corn, gold, S&P 500, silver. Pricing is here.

Frequently Asked Questions

What is open interest in simple terms?

The number of futures contracts currently outstanding, counted once each, not twice. It rises when a new buyer and a new seller create a contract and falls when both sides close. It measures how much money is committed to a market, as distinct from how much activity there is.

What is the difference between open interest and volume?

Volume counts contracts traded during a session and resets to zero daily. Open interest counts contracts outstanding and carries forward. A busy day with no change in open interest means existing positions changed hands rather than new ones being created.

Does rising open interest mean the price will go up?

No. Open interest has no direction of its own. Rising open interest means new contracts are being created, which happens in rallies and selloffs alike. It tells you money is entering the market, not which way it is leaning.

Why does open interest overstate directional positioning?

Because spread positions, where the same trader is long one delivery month and short another, sit inside it while expressing no view on direction. Across the 35 markets COTInsight covers on the report dated 1 September 2026, spreading averaged 29.0% of open interest, ranging from 70.5% in three-month SOFR to 1.9% in Dow Jones futures.

Where can I see spread positions?

In the CFTC Commitments of Traders report. The disaggregated and financial versions both publish a spreading column for each trader category. It is public and free, and it is the column most readers never open.

Why did a percent-of-open-interest figure jump without anyone trading?

Because open interest is the denominator and it moves independently, especially around contract roll. In COMEX silver in the week to 1 September 2026, the swap dealer short share rose 4.97 points while the dealer position itself fell 540 contracts, purely because open interest dropped 13.03%.