Open interest in gold futures, and how to read it without overreaching

Open interest counts commitments, not activity

Open interest is the number of futures contracts that have been opened and not yet closed, offset or delivered. It is a stock, measured at a point in time, whereas volume is a flow measured over a session. A market can trade enormous volume and finish with open interest unchanged, because the same positions were passed between hands all day without anybody adding fresh risk. It can also trade modest volume and see open interest jump, because the few trades that happened were all new commitments.

Because every contract has a buyer and a seller, open interest is counted once per contract rather than once per side. The exchange tallies it overnight from clearing records, which is why the figure you read today describes yesterday. That one session delay is unavoidable and matters less than people assume, since the useful signal is the direction of change across days rather than the level on any single morning.

One habit removes most confusion: say the word commitment in your head instead of the word interest. Open interest has nothing to do with how interested anyone is.

How a single trade moves the number

Every futures transaction has a buyer and a seller, and each of them is either establishing a new position or getting out of an existing one. That gives three outcomes, and they are worth memorising because almost every misreading of open interest comes from skipping them.

• A new buyer trades with a new seller. A contract that did not exist now exists, so open interest rises.
• A holder closing a long trades with a holder closing a short. The contract is extinguished, so open interest falls.
• A holder closing a long trades with a new buyer. The position simply changes owner, so open interest is unchanged even though volume was recorded.

The third case is the one that breaks naive analysis. A violent session with heavy volume and flat open interest means the risk was redistributed rather than increased. Somebody got out, somebody else got in, and the market as a whole is no more committed than it was at the open. That is a very different tape from one where the same volume arrived with open interest climbing all day.

The four combinations of price and participation

Pair the direction of price with the direction of open interest and you get four cells. Each describes a different kind of market, and none of them is a prediction.

Price up with open interest up means new longs are being established and the sellers meeting them are also new. Fresh money is funding the advance. Price up with open interest down is the opposite character: the move is being driven by shorts getting out rather than by anybody wanting to own gold, and once the trapped sellers are finished that fuel is gone. Price down with open interest up means new shorts are committing, which is a market being actively pressed. Price down with open interest down is ordinary liquidation, longs leaving without a committed seller replacing them, which often marks exhaustion rather than conviction.

The value of the grid is that it separates moves funded by new risk from moves funded by somebody closing old risk. A rally built on short covering and a rally built on fresh buying look identical on a candlestick chart and behave very differently when they meet resistance.

Why the front month dominates the total

Gold futures exist in a ladder of delivery months, and participation is wildly uneven across it. Nearly all speculative activity concentrates in whichever month is currently most liquid, while the distant months carry hedging and spread positions that barely move. This means the headline total for all months can drift for reasons that have nothing to do with anybody taking a view on gold.

Two distortions follow. First, spread positions count towards open interest even though they carry almost no directional risk, because a trader long one month and short another has two contracts open and no real exposure to the price. Second, during the period when liquidity migrates from the expiring month to the next one, the total gets temporarily inflated by exactly that spreading activity as positions sit in both months at once. A spike in total open interest during that window is plumbing, not conviction.

The fix is to read the active month on its own, and to check whether a change in the total came from directional positions or from the ladder rearranging itself. Looking only at the aggregate is how people announce a surge in participation that is really just a calendar.

Delivery, and the small physical tail

Open interest in a delivery month collapses before the month opens for delivery, and the mechanism is instructive. Most participants never intend to handle metal. They are there for price exposure, so they either close out or move the position into a later month. Only a small residue of holders stands for delivery, typically firms that genuinely want or genuinely have bars.

Those who want to swap a futures position for physical metal can do it directly, through a privately negotiated trade that exchanges the contract for the equivalent physical position. This is a normal, dull piece of market machinery, and it is one of the main bridges between the exchange and the over-the-counter market. It also explains why a large open interest figure is not a claim on the vault in any dramatic sense. The contract is a promise that can be settled in several ways, and the overwhelming majority of promises are settled by cancelling them.

Where it gets genuinely interesting is when delivery month open interest stays unusually high as the deadline approaches, because that means holders who could have left have chosen not to.

What rising open interest into a level actually tells you

This is the part a chart trader can use. When price grinds into an obvious level and open interest climbs through the approach, both sides are adding rather than waiting. Longs are buying the breakout attempt, shorts are defending the level, and the position stack above and below is growing. A resolution out of that condition tends to move faster, because whichever side is wrong has to cover, and that covering is itself new order flow.

When price reaches the same level and open interest is flat or falling, the approach is being made by people closing positions rather than taking them. Breaks from that condition more often stall, because the move was the unwinding and there is nothing left to force. This is the same reasoning as effort versus result, applied to commitment rather than to candle size.

The caveat is severe. Open interest is a daily figure, published after the fact, so you are describing yesterday's commitment while trading today's level. It frames expectations. It cannot confirm anything in real time.

Where the number misleads

Four traps account for most bad conclusions. The first is venue blindness: exchange open interest is one pool of gold risk, and the over-the-counter market alongside it is invisible, so a fall in futures commitment can simply mean exposure moved somewhere you cannot see. The second is the spread inflation already described. The third is intent: a dealer adding open interest to hedge a client trade carries no opinion at all, yet the number looks the same as a fund taking a view.

The fourth is the one worth labouring. Spot gold feeds do not publish traded volume. What appears as volume on a spot chart is a count of price updates, so it tracks activity and volatility well and reveals nothing about buying versus selling. Real traded volume and real open interest live on the futures market, not on the spot series most retail traders watch. Pretending otherwise is the root of a long list of bad indicators, as the cautions around volume divergence set out in detail.

Fitting it into a routine without over-weighting it

A practical approach is to treat open interest as a once a day note that colours how you read structure, never as an entry condition. Check the active month rather than the total. Ask one question: did yesterday's move come with more commitment or less. Write the answer down in two words. That is the whole exercise.

Then let the chart do the work. If your structural read says a level matters, and commitment has been building into it, you can reasonably expect a sharper resolution and should plan for worse slippage rather than better. If commitment has been bleeding away, expect the level to be tested without much follow through. Keep the read in the same place as the rest of your context, alongside whatever your market structure work already told you, and execute from the live chart. A number published overnight has no business overruling what price is doing in front of you.