MyAlgoGuru Help
Education Glossary Open the app
Chapter 1 of 5

Open interest and the four build-ups

The 2-by-2 grid everybody quotes is genuinely useful and says less than it is usually made to say. Here is the grid, and then the honest version.

Volume tells you how much changed hands. Open interest tells you how much is still held.

They are different questions and the difference is the whole value of open interest. A million contracts could trade in a session with open interest unchanged — the same positions passed around. Or a hundred thousand could trade and open interest rise by eighty thousand, which means new positions were created and somebody now has exposure they did not have this morning.

That second case is information about commitment. Price tells you where the market settled; open interest tells you whether anybody backed it.

The grid

Combine the day's price change with the day's change in open interest and you get four cases, each with a conventional name.

price rising Short covering price up, positions closing a move driven by exits, which tend to run out Long build-up price up, positions opening new money behind the direction Long unwinding price down, positions closing holders leaving, not new sellers arriving Short build-up price down, positions opening new money behind the direction price falling OI falling OI rising The right-hand column is commitment. The left-hand column is unwinding, and it exhausts itself.
The grid's real content is the left-right split, not the four names. On the right, open interest is rising: positions are being created, so the move has new money behind it and a reason to continue. On the left, open interest is falling: the move is being produced by people closing out, and there is a finite supply of positions to close. A rally on falling open interest and a rally on rising open interest look identical on a price chart and mean different things about tomorrow.

That left-right distinction is worth more than the four labels. "Is this move creating exposure or retiring it?" is a question with a real answer, and the answer carries information about whether the move has fuel.

Now the honest version

The four names contain an assumption that is almost never stated, and once you see it you will read every open-interest commentary differently.

Every contract has a long and a short. Open interest is a count of contracts outstanding — it does not have a side. When open interest rises by fifty thousand, fifty thousand new long positions and fifty thousand new short positions were created. There is no such thing as open interest going up on the long side only.

So what does "long build-up" actually claim? It claims that the aggressor — the side that crossed the spread and pushed the price — was the buyer, inferred from the fact that price rose while positions were created. That inference is reasonable. It is not a measurement, and it is the entire basis of the label.

This matters in three practical ways:

The label is an interpretation; the numbers are the data. Price change and open-interest change are facts. "Long build-up" is a reading of them. Treat the reading as a hypothesis you hold loosely, especially when it conflicts with something you can measure directly.

The same data supports an opposite story. Price up with open interest up is conventionally bullish. It is also precisely what happens when sellers write calls into a rally — new positions created, price rising, and the new money is positioned against the move. Both happen. Open interest cannot tell you which, because it has no side.

Who is on each side matters more than the grid. The exchange's participant-category data — which separates client, proprietary, institutional and foreign activity — carries considerably more information than aggregate open interest, because it does distinguish sides. If you are going to build a view on positioning, that is the better input.

Reading it in practice

Three habits that make open interest useful rather than decorative.

Compare today's change against the recent average change, not against total open interest. Total open interest is a slow-moving stock; the day's change is the flow. A 3% rise in total open interest sounds small and may be the largest single-day build in two months.

Look at where it is being added, not just how much. In options, the strike matters enormously — fifty thousand contracts added at a strike 2% away means something quite different from fifty thousand added at the money. The chapter on reading the chain is about exactly this.

Check whether your open-interest number is live or a snapshot. This is the mundane one that catches people. Open interest is not always updated in real time, and a figure that is a snapshot from an earlier point is a statement about then, not now. Acting on a stale open-interest reading as though it were current is a common and entirely avoidable error — and it is worth knowing which your own screen is showing you before you build a rule on it.

Illustrative

Two sessions in the same instrument. Both close up 1.2%.

Session A: open interest falls 7%. Volume is heavy. The conventional read is short covering — the rally was produced by shorts buying back, and the supply of shorts to squeeze is now 7% smaller than it was. Whatever drove it has partly spent itself.

Session B: open interest rises 9%. Same price move. New positions were created on both sides and total exposure in the instrument is meaningfully larger than yesterday. Whatever this move is about, participants are still committing to it.

The useful conclusion is not "A is bearish and B is bullish". It is that A's move has a fuel source that is measurably depleting, and B's does not. That is a statement about the character of the move, and character is what internals are for.

The useless conclusion — and the common one — is to treat Session B as a buy signal. Open interest rose on both sides. Someone committed to each direction, in equal measure, and the grid cannot tell you which of them was right.

What to take from this chapter

  • Volume is what changed hands; open interest is what is still held. Open interest is the commitment measure.
  • The grid's real content is whether the move is creating exposure or retiring it. Moves driven by unwinding have a finite fuel supply.
  • Open interest has no side. Every new contract creates a long and a short, so "long build-up" is an inference about the aggressor, not a measurement.
  • The same reading — price up, OI up — is equally consistent with sellers writing into a rally. The data cannot distinguish them.
  • Participant-category data distinguishes sides; aggregate open interest does not. Prefer it for positioning views.
  • Compare the day's change against recent daily changes, not against the total. And in options, look at which strikes.
  • Know whether your open-interest figure is live or a snapshot before you build a rule on it.