8. A Photograph of a Market That No Longer Exists

8. A Photograph of a Market That No Longer Exists

Let's return to the morning table that opened lesson 5 — the daily open interest report, the foundation of all classical analytics. We already know how walls, magnets and GEX profiles are made from it. Now, having met 0DTE, let's ask the table one simple question: what does it actually see?

The invisible half

Open interest is recorded once a day, after the close. Let's follow a typical 0DTE position across that calendar.

Yesterday's evening snapshot did not catch it — it did not yet exist. Today's will not catch it either — it was closed at 3:50 pm, and whatever survived to the end of the day expired and went to zero. The contract lived for six hours, its hedging chased the futures around all day — and in the main source of data it left not a single record. Not a late record — none at all.

To be fair: today's expiration does have a past. The series expiring today was listed in advance and traded yesterday and before — and the positions opened in it ahead of time, the morning OI honestly shows. But that visible part is a thin layer: the bulk of 0DTE volume opens on the expiration day itself, lives for hours, and vanishes, like the position in the diagram. And in the evening, expiration zeroes out the whole series — so even its "past" will disappear from tomorrow's report.

Now multiply by the scale from the previous lesson: about half the volume of the world's largest options market flows through same-day series — and the lion's share of those trades is born after the morning snapshot. Daily OI is not late in reporting them — it is in principle incapable of seeing what opened and closed between its frames. The instrument didn't break; the world it was designed for did.

And what about the other half?

Positions longer than a day do make it into OI — but here too the calendar spoils the picture. The snapshot was taken yesterday evening. A wall computed from it is a wall of yesterday's market; by lunchtime, today's flows may have built up a node at a different strike or dismantled yesterday's. Levels that move within the day, a daily photograph by definition shows a day late.

Worse still: the weak spot of GEX analytics from lesson 5 — the assumption about who stands on which side of the trade — became even more vulnerable in the new world. When positions accumulated over weeks, an error in the guess about one trade drowned in the mass. When the book is rebuilt within hours, the sides have to be guessed afresh every day — and the data for that is as absent from OI as it ever was.

A spec sheet for new data

Let's spell out what it would take to truly see the 0DTE market. The list is short, and each item strikes at a specific blindness of daily OI.

Frequency: minutes, not days. The dealer book lives within the day — so position data has to update within the day too.

Sides, not guesses. To compute the dealer book without assumptions, we need to know whether each trade was a buy or a sell — the very thing OI has been silent about since lesson 2.

Participants, not faceless volume. Retail buying lottery puts and institutions selling premium create opposite dealer books. Telling them apart means knowing the book's sign for the first time instead of assuming it.

For a long time such a list looked like fantasy: exchanges published nothing of the kind. And then the exchange where the entire SPX complex trades began selling exactly this — minute-by-minute positioning data broken down by participant type and trade side. That feed — and what we built on top of it — is what the entire remainder of the course is about.