Dealer gamma, explained.
Two sessions can look identical on a chart and behave nothing alike. Often the difference isn’t the setup — it’s which side of dealer gamma the market is on.
The mechanism
Every option a dealer holds has to be hedged, and how they hedge depends on which side of it they are on.
A dealer who is short an option is short gamma. Their delta moves against them as spot moves, so staying flat forces them to buy as price rises and sell as price falls. They trade in the direction of the move, and that amplifies trend.
A dealer who is long an option does the reverse — selling into rallies, buying into dips. They trade against the move, and that damps trend.
Aggregate every dealer’s position and you get a net. That net has a sign, and the sign decides which of the two behaviours the market gets.
Hedging flow leans against every move.
Mean reversion. Suppressed realized volatility. Pinning toward the strikes where gamma is concentrated.
Hedging flow leans with every move.
Momentum. Amplified realized volatility. Moves that would normally be absorbed instead accelerate.
One hedge, twice.
Take a dealer who is short a block of calls with spot sitting near the strike. Their book is short gamma: as spot ticks up their delta gets shorter, so staying flat forces them to buy. Price rose, and the hedge bought. It ticks up again, and the hedge buys again. Nothing here is a prediction — it is a mechanical chase, and every unit of chase is extra demand landing on an already-rising tape.
Now flip the sign and change nothing else. A dealer who is long that same block gets longer delta as spot rises, so staying flat means selling into the rally and buying it back on the way down. Same price path, same rule, opposite sign — and the hedging is now supplying exactly the liquidity that absorbs the move instead of feeding it.
Neither dealer has a view. That is the whole point. The flow is a by-product of staying delta-neutral, it is automatic, and in aggregate it is large enough that its sign shows up in the tape.
The zero-gamma level is the spot price at which aggregate dealer gamma crosses zero — the boundary between those two behaviours.
The gamma profile
Plot aggregate dealer gamma against spot and the whole idea fits in one picture. The vertical axis is how much gamma the dealer complex is carrying; the horizontal axis is where spot could be. Everything above is the shape of that curve and the point where it changes sign.
Illustrative shape only. The curve is drawn, not computed — it is here to show what a gamma profile looks like, and carries no prices, no ticker and no date.
Read it left to right. On the left spot is low, the dealer complex is net short gamma, and hedging leans with every move — this is the half of the axis where ranges break. Somewhere in the middle the curve crosses zero. On the right dealers are net long gamma, hedging leans against the move, and the tape compresses toward the strikes carrying the most gamma.
The crossing is not a level someone drew on a chart. It is the spot price where a computed quantity changes sign, and it is the single most useful thing on the picture — because which side of it the market is on decides which of the two behaviours you are trading against.
The zero-gamma level
Three things about it get got wrong constantly, and all three matter more than the number itself.
It is a root, not a strike.
The level is the spot price that solves aggregate dealer gamma = 0. Spot is continuous, so the solution is a point on that axis and will usually not land on a listed strike. Reporting it as a strike is not a rounding convenience — it throws away the only precision the number has.
There can be more than one.
A curve that crosses zero exactly once is the common case, not a guaranteed one. Lumpy open interest can push the profile back and forth across the axis, and when it does there are several boundaries rather than one. The meaningful crossing is the one nearest spot — it is the only boundary price can reach without passing through another first.
It moves.
Gamma is a function of spot, time and implied volatility, so the level that solves for zero is a function of all three. It drifts as the session ages even if price does nothing, it shifts when IV moves, and it resets when open interest updates. Marking it once at the open and treating it as a fixed line for the day is the most common way to be wrong about it.
Walls
The call wall and put wall are the strikes carrying the largest dealer gamma concentration on each side of spot. They are the two places where hedging flow is densest — and the two places most confidently described wrongly.
In positive gamma they behave the way the folklore says. Price approaches the call wall, hedging leans against it, dealers sell into strength, and the rally decelerates; the strike reads as resistance. Price approaches the put wall, dealers buy into weakness, and the dip decelerates; the strike reads as support. In this regime — and only in this regime — talking about price being pulled toward a heavy strike is defensible, because the hedging really is pulling it there.
In negative gamma the same two strikes do the opposite. The gamma concentration has not moved. The sign of the hedge has. Approaching the call wall now means dealers buy into strength, which puts the largest single concentration of hedging flow on the board behind the move rather than in front of it. A break through a wall in short gamma is not a failed bounce. It is the wall doing exactly what its size implies, in the other direction.
Which is the whole lesson: a wall is a magnitude, not a direction. It tells you where the flow is concentrated. The regime tells you which way that flow points. Read them in that order and the walls stop contradicting themselves.
Max pain
Max pain is the strike at which the aggregate payout to option holders is smallest — the settlement price that would leave the most contracts worthless. It is an accounting fact about the open chain: price every strike’s payout at every candidate settlement, and take the minimum.
Two consequences follow, and both cut against how it usually gets used. First, it is a tendency, not a target. Whatever pull exists comes from the same hedging flow described above, and that flow is only concentrated enough to matter when a large share of the chain is close to expiry. Through most of a cycle max pain has essentially no grip on price. Second, it moves with the chain, so a max pain computed on yesterday’s open interest is a statement about yesterday’s book.
Context in the last day or two of an expiry. Trivia before that. It is not a destination and nothing on this site will present it as one.
Vanna and charm
Gamma is not the only term that forces a hedge. Two second-order sensitivities move dealer delta with price standing perfectly still.
Vanna is how delta responds to implied volatility. When IV falls, the delta of out-of-the-money options falls with it, and a dealer who was flat a moment ago is now over-hedged and has to trade to get back to neutral. That is why a volatility crush can produce a directional grind on an otherwise quiet tape — the flow is not an opinion about price, it is a re-hedge against a change in IV. It matters most around scheduled volatility events, where IV has a known reason to move sharply in a known direction.
Charm is how delta responds to the passage of time. Options decay toward 0 or 1 delta as expiry approaches, so a hedged book drifts out of hedge simply because the clock ran. Charm flow is mechanical and calendar-driven: it builds through an expiry cycle, concentrates into the final sessions, and is largest where open interest is largest. It is one of the very few flows in the market whose timing is known in advance.
Both are genuinely second-order, and this page is not going to pretend otherwise. They are smaller than gamma hedging in most conditions, and they are estimated from the same once-a-day open-interest snapshot everything else here rests on. They are useful for understanding why a session felt the way it did. They are not something to hang a trade on by themselves.
Macro events and expiry
A calendar tells you when something is scheduled. Everyone has that. What it cannot tell you is what the market is carrying into the print — and that is the part that decides how the number actually trades.
The frame that joins the two: a scheduled event is when dealer positioning gets tested. The event supplies the shock. The regime decides what the market does with it.
Into a print with dealers long gamma, hedging is damping. Participation thins as people stand aside, and the flow that remains leans against every push, so the tape compresses into the release — the coil that gets noticed and then misattributed to anticipation. Resolution comes after the number, when the damping is overwhelmed, rather than before it.
The same print with dealers short gamma has no such brake. Hedging leans with the move, so the first push gets extended instead of absorbed, and the range that would have contained it on another day simply is not there. Identical headline, identical surprise, two completely different sessions — and the difference was set before the number printed.
Expiry changes the arithmetic again. Gamma concentration decays as contracts roll off, so the same event lands differently in the week of a large expiry than in the week after it. A heavy chain means more hedging flow anchored to specific strikes; a cleared board means less of everything — less damping when gamma is positive, less amplification when it is negative. An event landing into a heavy expiry is being absorbed by a very different book than one landing a week later.
And charm runs underneath all of it. The decay-driven re-hedge above is calendar flow, not news flow: it builds into expiry whatever is on the schedule. Some of what looks like positioning ahead of an event is just the clock.
None of this forecasts the print. It is the reason two traders can read the same calendar, take the same view on the number, and get two different outcomes.
QuantX carries the schedule and the wire on the macro page, and the regime read on GEX Levels. The point of having both is to read the second before the first.
What the research measured
This is not a theory we invented. Baltussen, Da, Lammers and Martens published it in the Journal of Financial Economics in 2021. They split the trading day by the sign of net dealer gamma and measured intraday momentum — whether the morning’s return predicts the last half hour’s.
The split is the entire result.
| Dealer positioning | Coefficient | t-statistic | R² |
|---|---|---|---|
| Net short gamma | 6.63 | 4.78 | 3.58% |
| Net long gamma | 0.82 | 1.03 | 0.05% |
Read the t-statistic column first. 4.78 is a result. 1.03 is noise. The regime does not merely change the size of the effect — it changes whether the effect exists at all. Intraday momentum is strong and significant when dealers are short gamma, and statistically absent when they are long.
The sample runs 1996–2020: 2,930 negative-gamma days against 3,158 positive. The effect concentrates in the final 30 minutes of the session and partially reverses within three days — consistent with transitory hedging pressure rather than new information arriving.
CBOE’s own research on same-day-expiry options points the same direction. On average, 0DTE dampens realized volatility, by roughly −0.19pp annualized. But dealer gamma turns negative at some point on at least 25% of days, and inside those windows the estimated volatility impact reaches +3.3pp daily and +6.4pp on 30-minute horizons. The average is calm. The tails live entirely in the negative-gamma windows.
What this does not tell you
Four limits, stated plainly, because they bound everything above.
A positioning map, not a forecast.
GEX describes where hedging pressure changes character. It does not say where price has to go, and nothing on this page should be read as a prediction that it will.
The positioning convention is an assumption.
Assuming dealers are long calls and short puts is a heuristic. It holds reasonably at the index level, but it is still an assumption — and every number derived from it inherits it.
Open interest publishes once a day.
Any measure built on open interest is a prior-day snapshot with live spot applied to it. It is not a live read of dealer inventory, and no one’s is.
Same-day expiries are largely invisible to it.
0DTE options were roughly 34.8% of SPX/SPXW volume in CBOE’s sample, and they do not appear in yesterday’s open-interest snapshot at all.
So we will not tell you where price is going. We will tell you which regime you are in, where the boundary sits, and what the market has historically done from there. What you do with that is the trade.
Sources
Guido Baltussen, Zhi Da, Sten Lammers, Martin Martens. Journal of Financial Economics, 2021. PDF.
Cboe Global Markets, research publication. PDF.
Informational and educational only. QuantX is not a registered investment advisor and nothing here is investment advice. Trading involves substantial risk of loss.