At the end of lesson 8 we wrote out a spec in three points: minute-level frequency, known trade sides, a breakdown by participant. Now let's ask: who could even collect data like that?
There is only one answer. Every SPX options trade passes through a single point — the CBOE exchange, where this market trades. Only the exchange sees both sides of every trade, knows what type of account each order came from, and whether it opens a position or closes one. Analysts spent years reconstructing this picture from circumstantial evidence — the exchange simply has it, in full. On top of this exchange data stands a minute-by-minute positioning feed for SPX options, broken down by participant type and trade side; we receive it through a licensed data vendor, and for the rest of the course we will simply call it the minute feed. This feed is the foundation of every indicator we will discuss through the end of the course.
What's inside a minute
The feed is built simply — and that simplicity is its strength. Every minute, for every strike and every series, it reports how many contracts were bought and sold, opening and closing positions, by participants of each of four types:
Retail — retail clients: individual accounts, those very buyers of lottery tickets and one-day insurance.
Pros — professional clients: active traders and small funds who trade enough that the exchange stops counting them as retail. Retail and the pros regularly end up on opposite sides — we will be watching their tug-of-war in the indicators.
Firms — the proprietary books of brokerage houses, and the institutions that route through them.
Brokers — broker-dealers executing client flow.
Add up the buys and sells minute after minute and the exchange gets the net position of each cohort at each strike — not a guess about it, but an accounting fact: this much bought, this much sold, here is the balance.
One caveat about units, right away, because it shapes everything that follows. Raw contract counts are the exchange's property and never leave it: that is a licence condition. Positions arrive in our indicators already converted into delta exposure in ES-equivalent — risk rather than counts — and that is not a loss but a gain: counts are incomparable across strikes, risk is comparable. We will unpack this unit in detail in the next lesson.
The fifth participant
An attentive reader will notice: market makers are not on the list. They don't need to be — and this is the most beautiful part of the whole construction. Recall lesson 4: the dealer stands on the other side of nearly every client trade. If the four client cohorts together bought a thousand contracts at a strike, someone sold that thousand. The dealers' position is the sum of the client positions with the sign flipped — at every strike, every minute.
Let's pause and appreciate what this means for our story. In lesson 5, all of GEX analytics rested on the assumption "suppose the clients behaved typically" — the very slack we asked you to remember. The minute feed removes the assumption entirely: the dealer book is not guessed from habits — it is computed from known trade sides. The very book that "nobody publishes" has become observable. Minute by minute.
From positions to forces: four exposures
From here the pipeline we assembled in the first part of the course kicks in. Know the book, and you know its greeks. Multiply the dealers' positions by each option's greeks and sum across strikes — and you get four maps of dealer forces, minute by minute:
GEX (gamma exposure) — the gamma map familiar from lesson 5: where hedging brakes price, where it accelerates it. Only now it is not a daily estimate but a minute-by-minute fact.
DEX (delta exposure) — the delta map: how much futures equivalent the dealers hold against each strike.
VEX (vanna exposure) — the vanna map: how the dealer book will react to a shift in volatility. The very vanna introduced by name in lesson 3 — its hour on stage is drawing near.
CEX (charm exposure) — the charm map: what flow the mere passage of time will generate.
GEX and DEX you can read already; we'll take vanna and charm on in earnest in lesson 11 — first let's show the simpler things on charts.
What the feed does not see
In the tradition of this course — an honest paragraph about the limits, before we allow ourselves to admire.
Positions, not the tape. The feed shows the change in positions over a minute, not every trade. A position opened and closed within the same minute vanishes without leaving a trace — so any activity computed from the feed (repositioned in the flow indicators, lesson 10) is a lower bound on the real thing: everything we see certainly happened, but something happened on top of it.
Facts, not intentions. We know that retail bought a thousand puts at a strike. We do not know why: insurance, a bet, or one leg of a complex structure. The data shows positions — interpretation remains the analyst's job.
SPX, not everything under the sun. No such data exists for oil, gold, or single stocks — and not merely because the exchange doesn't sell it. Depth like this makes sense where the options tail truly wags the dog, and that is the index complex, with its trillions in notional and half its volume in 0DTE. Why a signal from SPX options can legitimately be read on the chart of the ES futures we covered back in lesson 2 — the arbitrage linkage of the complex; we'll say more about the limits of applicability in lesson 15.
So here is the full set of inputs in hand: minute-by-minute positions of four client cohorts, the dealer book derived from them, and four maps of forces on top. Let's start unpacking these riches with the simplest question any market begins with: who trades, and how much? Even that question — you will see — sounds entirely different with the new data.
- Only the exchange sees both sides of every trade, the account type, and the open/close flag — the minute positioning feed is built on that data.
- The feed delivers minute-by-minute positions of four cohorts: Retail, Pros, Firms, Brokers; the dealers are derived as their sum with the sign flipped.
- Positions arrive in the indicators not as contract counts but as delta exposure in ES-equivalent — the risk the dealers will have to hedge.
- The dealer book went from hypothesis to computable fact — minute by minute; on top of it we compute the GEX, DEX, VEX and CEX maps.
- The limits are honest: the feed sees position changes once a minute (activity is a lower bound), shows facts rather than intentions, and exists only for SPX.