Gamma Exposure: Hedging You Cannot See
Gamma exposure is an estimate of how much hedging the dealer community must do for each unit of price movement in the underlying. It is inferred from option open interest rather than published, and it describes the likely character of movement — dampened or amplified — rather than its direction.
How it works
Options leave somebody holding a position they did not choose. A dealer who sells a contract takes on directional exposure, and offsets it by trading the underlying.
That hedge cannot be set once. Delta, how much the option moves per unit of the underlying, changes as price moves — so the hedge has to be adjusted continuously.
Gamma is the rate at which delta changes, and so the rate at which the hedge must change. Gamma exposure aggregates that across the dealer community into a single estimate of hedging per unit of price movement.
One sign means they sell rallies and buy dips. Net long gamma makes the hedge sell into strength and buy into weakness, which leans against the move and compresses ranges.
The other sign does the opposite and amplifies it. Net short gamma makes a rise require buying and a fall require selling, so the flow pushes with the move. Market makers hedge either way without preference.
The flip point, and what it rests on
The flip point is the level at which the estimated position changes sign. Behavior is expected to change character there, which is the whole reason the level is watched.
It is not support or resistance. Nothing says price should stop there — only that the texture of movement either side should differ, a claim about behavior rather than market structure.
But dealer positioning is estimated, not published. Nobody outside the dealers knows their book; every published figure is an inference from option open interest plus an assumption about who holds which side.
The usual assumption is that dealers sold what retail bought. Somebody wrote those contracts and the dealer community is the obvious candidate — reasonable, and still an assumption.
And it resets wholesale on every expiry. A large share of open interest ceases to exist at options expiry, so the estimate changes discontinuously on known dates and a reading carried across one is already stale.
In practice
It is a context reading, not a trade on its own. The question it answers well is why a market has been unusually calm or unusually jumpy — worth knowing, never an entry.
The option volume is where the estimate comes from. Contracts traded become open interest, open interest is what the calculation reads, and a heavy options day moves the figure most.
On a daily chart it explains calm, not direction. The reading suggests which regime you are in, dampened or amplified, and never which way the next bar goes.
And a gap through the flip point changes everything. An opening gap can carry price across the level before anything trades, so the session’s expected character inverts at once.
It never tells you where the stop belongs. A stop loss sits where your idea is wrong, and this reading has no opinion about your idea.
And acting on it repeatedly is expensive. Every regime change it prompts is another round trip through the spread, and a $10,000 account paying $10 a round trip spends $200 on twenty adjustments, so extra trades must be worth the liquidity they consume.
Using it alongside a method
Nothing here is an entry. The responsible use is narrower and duller: the reading suggests what character of movement to expect, and character is an input to sizing and target distance, not to selection.
In a dampened regime, expect movement to be given back. That argues for closer targets and smaller size on breakout ideas. It does not argue for taking a mean-reversion trade you would otherwise have skipped.
In an amplified regime, expect follow-through and wider swings. That argues for wider stops, which argues in turn for smaller size, and for giving a working trade more room. The trade itself still has to come from somewhere else.
The boundary is the discipline. If a reading ever becomes the reason you took a position, it has stopped being context and become a signal — the same trap as reading implied volatility as a forecast.
What gamma exposure is not
It is not a direction. It describes how movement propagates, not which way it goes.
It is not published data. Every figure is an inference plus an assumption about who sold what.
It is not a support level. The flip point marks a change of character, not a place price must stop.
And it is not gamma scalping. That is a trade run on your own book; this is a reading of somebody else’s.
How this connects
The raw inputs are ordinary contracts. Every call option and put option at a given strike price contributes, weighted by how much gamma it carries, and contracts at the money dominate the sum.
Theta is the other half of the dealer’s arithmetic. A dealer short gamma is collecting decay in exchange for hedging against the move, so the regime describes who is being paid to absorb volatility.
It is related to, but not the same as, the volatility risk premium. That premium is the gap between implied and later historical volatility; gamma positioning is one reason realized movement can undershoot or overshoot what options priced.
Vega is a separate exposure on the same books. A change in implied volatility moves dealer hedges too, which is why readings shift after large volatility events even without a move in price.
The heavily optioned index products carry the most meaningful readings. SPY and the Nasdaq contracts have deep option markets, while a thinly optioned stock produces an estimate built on very little.
A short-gamma regime can resemble volatility clustering, where large moves follow large moves, and in single stocks it is one route to a short squeeze.
Practically, it informs position sizing rather than selection. An amplified regime argues for a wider stop based on average true range and a smaller position, while a dampened one is friendlier to mean reversion.
When it fails
In a range it is usually the reason for the range — or it is credited with being one, which is not the same thing. A trading range is the ordinary state of most series.
The second failure is trusting the sign. If the assumption about who sold what is wrong, the sign can be inverted, and with it the whole expected behavior.
A third is a stale reading across an expiry. The figure is rebuilt on known dates, so a level quoted from before one describes a book already dismantled.
A fourth is treating the flip point as precise. It rests on an assumption, so a level stated to a fine tolerance shows more confidence than its inputs support.
A fifth is single-market thinking. A reading covers only the contracts it counts, and exposure held in related instruments or on other venues never appears in it.
And the sixth costs the most. Using it to justify a trade your method did not give you turns background information into a reason.
The original data
Three of the 24,971 videos collected for this site’s search study name gamma exposure in a title — a median of 31,371 views across three channels, with a maximum of 109,534. Two more use the abbreviation GEX, at a median of 21,124. Against 110 order-flow videos at 22,195, that is a small, engaged audience for a technical subject.
The question: Price just crossed the flip level. Fade or follow?
But a calm market is not evidence of dealer hedging. Most people read a quiet tape as a sign that something is pinning price in place, yet markets drift for years with no such cause. The S&P 500 spent 1,897 trading days, from January 1973 to July 1980, below a record close — before options on the index itself were even listed. Ranging is the default, so treat this as a statement about how a market is likely to behave rather than where it is going, and never size a position on it alone.
Related
Gamma is the single-option quantity this estimate aggregates. Market makers are the firms whose hedging it describes. And options are the instruments whose open interest it is built from.
I treat this the way I treat a weather forecast before a long drive. It changes how fast I go and how much room I leave, and it never decides where I am driving to. I do not trade the flip level directly, because it is a level nobody actually publishes, so it only ever adjusts a trade my method has already given me.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.