Somebody is short the options you buy. Whoever they are, if they run a hedged book they cannot sit still: every dollar the underlying moves changes the delta of what they are short, and they have to trade the stock to stay flat. That mechanical, non-discretionary trading is the reason certain strikes behave like magnets and certain price zones feel wilder than others.

Gamma exposure (GEX) is the standard attempt to map where that hedging pressure concentrates. This guide is about using it honestly: how it is computed, what part of it is measured and what part is assumed, how to read the flip point, and — the part most GEX content skips — the specific questions it is structurally incapable of answering.

One disclosure up front, because it changes how you should read every number below. We do not have dealer-positioning data, and neither does anyone selling it to you retail. Option open interest is published without holder identity. Every GEX chart in existence — ours included — is open interest run through a sign convention, which is a behavioural assumption about who is short what. The mechanism is textbook; the attribution is a guess. The rest of this page is written to keep those two things separate.

What gamma exposure actually is

Delta tells you how much an option's value moves per dollar of underlying. Gamma tells you how fast that delta itself changes. A hedged options book cares about gamma more than almost anything else, because gamma is what forces it to re-trade the underlying rather than set a hedge once and walk away.

Gamma exposure takes that idea to the market level: aggregate the gamma of every open contract, weight each one by its open interest, assign a sign, and you get an estimate of how much underlying stock a fully hedged short-option complex would have to buy or sell for a given move in price.

The output is usually expressed one of two ways:

Share GEX — shares of the underlying per 1% move.
Dollar GEX — notional dollars per 1% move.

Both are estimates of a hedging requirement, not of anyone's inventory.

How GEX is computed from open interest

The per-contract building block is:

GEX_contract = gamma x open_interest x 100 x S^2 x 0.01
                 |          |           |     |     |
                 |          |           |     |     +-- scale to a 1% move
                 |          |           |     +-------- convert delta-shares to
                 |          |           |               dollars, twice over
                 |          |           +-------------- contract multiplier
                 |          +-------------------------- contracts outstanding
                 +------------------------------------- Black-Scholes gamma

Then you sign it — calls positive, puts negative — and sum:

GEX_total = SUM(call gamma terms) - SUM(put gamma terms)

Sum by strike and you get a profile you can plot. Sum the whole surface and you get the single headline number vendors quote.

A worked pass with illustrative inputs, so the units stop being abstract. Take a $536 stock, one call strike at $550 with 40,000 contracts open and a per-contract gamma of 0.012:

0.012 x 40,000 x 100 x 536^2 x 0.01 = ~$137.9M of notional per 1% move

That is the size of the hedge a fully delta-hedged short position in that one strike would have to transact for a 1% move. Repeat across every strike and expiry, net the signs, and you have the profile below.

Reading the profile

Plotted by strike, GEX is a two-sided bar chart with spot somewhere in the middle:

         GEX contribution by strike (all expiries summed)
         negative  <------------- 0 -------------> positive

   560                          |###
   555                          |########
   550  call wall               |####################
   545                          |#########
   540                          |####
   -----------------------------+--------------------- spot 536
   535                       ###|
   530                   #######|
   525  put wall  ##############|
   520              ############|
   515                    ######|

Two features carry almost all of the information.

Concentrations. The tallest positive bar above spot is the call wall; the deepest negative bar below spot is the put wall. These are where a hedging response, if it exists, would be largest — the levels most likely to behave like stall zones or cushions.

The crossing. Where the cumulative profile changes sign is the gamma flip point, also called the hedge wall. It separates two qualitatively different markets:

   cumulative GEX
        ^
    (+) |                                     ________   compression:
        |                            ______---/          moves tend to fade
        |                  ______---/
      0 +---------------o-/------------------------------>  spot price
        |         ___--/ ^
        |    __--/       flip point (gamma flip / hedge wall)
    (-) | __-/
        |/                          expansion: moves tend to run

One computational detail worth knowing, because it is where cheap GEX tools cut a corner: the flip point is not the price where today's bar chart happens to cross zero. Gamma has to be re-evaluated at each candidate spot — move the stock and every contract's gamma changes — so the flip is a root-find over a re-priced surface, not an interpolation of a static chart. A tool that reports a flip point without re-pricing is reporting an artifact of its own grid.

Positive and negative gamma are two different markets

If the short side of the aggregate book is net long gamma, hedging leans against price. If it is net short gamma, hedging leans with it. That single sign change swaps the character of the tape:

	Positive aggregate gamma (above the flip)	Negative aggregate gamma (below the flip)
Hedge response to a rally	sell the underlying	buy the underlying
Hedge response to a selloff	buy the underlying	sell the underlying
Effect on realised volatility	dampened	amplified
Typical price behaviour	mean-reverting; pins near heavy strikes	trending; gaps extend rather than fill
Intraday ranges	compress into the close	widen into the close
For premium sellers	friendlier regime	hostile regime
For direction	tells you nothing	tells you nothing

That last row is the one traders keep re-learning the hard way. Being below the flip point is a statement about the distribution of the next move, not its sign. It is a volatility read, and reading it as a bearish signal is the single most common misuse of GEX.

What our own testing says about this

We backtested the wall framework on every optionable ticker back to 2012, rebuilding each level from the positioning available on the day and walking forward. Two findings are directly relevant to anyone using GEX for regime calls, and they are written up in full in Do Gamma Walls Actually Work?:

The regime split is real. Forward realised volatility ran meaningfully higher below the hedge wall than above it. It was statistically significant over the decade, positive on every name in the deep-tested set, and it held up in calm and stressed markets — which is where a volatility signal has to earn its keep.

But it is name-specific, and roughly one in three tickers invert it. For about a third of names the relationship runs backwards — volatility is higher above their hedge wall — and for many more it is essentially flat. Leveraged and inverse-volatility products are the worst offenders. The average edge is real; applying the average to an arbitrary ticker is not a strategy. You need that name's own record, which is why we grade every symbol against its own multi-year history rather than shipping one universal rule.

The sign convention is an assumption, not an observation

This is the section that separates a usable GEX read from a confident-sounding one.

Open interest tells you how many contracts are outstanding at a strike. It does not tell you who is long and who is short. Nothing public does. The +calls / −puts convention fills that gap with a behavioural story: retail and institutions tend to overwrite calls and buy puts for protection, so the counterparty ends up long calls and short puts.

Sometimes that story is right. Often, at a specific strike, it is plainly wrong — a large institutional call buyer, a put spread where the customer is short the lower strike, a corporate collar, index puts warehoused by a pension fund. The convention has no way to see any of that. Here is the honest inventory of what goes into the number:

Input	Where it comes from	Observed or assumed?
Strike, expiry, multiplier	contract terms	Observed
Open interest per contract	published nightly, post-clearing	Observed — and one session stale
Spot price	equity print	Observed
Per-contract gamma	Black-Scholes on a fitted IV surface	Modelled
Who is long, who is short	no feed carries this	Assumed (`+calls / −puts`)
Whether the short side rehedges at all	no feed carries this	Assumed (continuous rehedging)
Whether the hedge trades this underlying	no feed carries this	Assumed (single-stock delta)

Three of the seven inputs are assumptions, and they are the three that decide the sign. That is not a reason to throw GEX away — it is a reason to treat it as a map of where a hedging response would be concentrated if the convention holds, and never as a report of what anyone is doing.

We also tried to do better and failed, which is worth saying out loud. We built the popular shortcut of inferring positioning from order flow — reading the tape to decide which side the liquidity provider took — tested it honestly, and it inverted the signal. More history made it worse, not better. Getting the real answer needs true market-maker account flags, which the public tape does not carry. So we stayed with the plain convention and label it as a convention. Anyone selling you "flow-signed" gamma is selling you the thing that failed our test.

Why the level matters more than the number

Pull up two GEX providers on the same ticker on the same day and the headline numbers will disagree, sometimes by an order of magnitude. That is not a bug in one of them. It is because the magnitude depends on choices nobody standardises:

per 1% move or per $1 move
multiplied by S², by S, or not at all
all expiries, or only the front two
one contract multiplier of 100 applied once, twice, or folded into the units
whether a fitted IV surface or the exchange's closing IV feeds the gamma

Cross-ticker comparison is worse. A $4B GEX print on a mega-cap and a $4B print on a mid-cap describe totally different hedging burdens relative to how much stock actually trades. If you must compare magnitudes, normalise: divide by 20-day average dollar volume and read it as "how many days of volume a full rehedge would represent." A number below a few percent of ADV is unlikely to be the marginal driver of anything.

Locations do not suffer from this. The flip point is a price. The call wall is a strike. Both are invariant to unit conventions, both are directly testable against subsequent price action, and the location is what our decade of testing actually validated — the above/below regime split, not the height of any bar. Practical rule: trade the geography, not the magnitude. Use the number only to rank whether a level is worth caring about at all.

What GEX cannot tell you

Question	Can GEX answer it?
Which direction is the stock going?	No. It is a volatility and location read only.
Who is actually short these contracts?	No. Not observable in any public feed.
Are they hedging right now?	No. GEX is a stock of positions, not a flow.
Will the put wall hold?	Only as a tendency, and only for some names.
What happens on an earnings gap?	No — gaps happen when hedging cannot operate.
How big is the hedge relative to liquidity?	Only after you normalise by traded volume.

A few more limits worth internalising:

Open interest is stale by construction. It is published after the session clears, so any intraday GEX is built on yesterday's positions. On heavy 0DTE days that is a serious gap: contracts opened and closed within the session never appear in the snapshot you are reading.
Gamma decays into expiry, then vanishes. A profile computed Monday is a different animal by Friday afternoon, and the Monday after expiration the anchor is simply gone. Walls move every session; a level is only meaningful for the day it was computed for.
Hedging is discrete, not continuous. Real books rehedge in bands, on a schedule, or by trading correlated instruments — index futures rather than the single stock. The model assumes a continuity that nobody actually practises.
It works best where option hedging is plausibly the dominant flow. On very liquid underlyings the hedging estimate can be large relative to real volume. On a thin name the same computation produces a number that no participant is large enough to act on.
A break inverts the mechanism. When price pushes through a heavy strike, the same hedging logic that was cushioning starts accelerating. Support becoming a trapdoor is the expected behaviour, not a failure of the model.

A reading checklist

When you pull up a GEX profile, in this order:

Where is spot relative to the flip point? Above means the compression regime, below means expansion. This is the only thing on the chart with a decade of statistical support behind it.
Does this ticker's own history support the regime read, or invert it? About one name in three runs backwards. Check before you act.
Where are the nearest concentrations above and below? Those are your zones — the levels where a hedging response, if it materialises, would be largest.
How big is the hedging estimate against average volume? If it is a rounding error next to daily turnover, stop treating it as a driver.
How much of the gamma expires this week? Front-dated gamma dominates the profile and disappears at the bell on expiration Friday.
What is the implied volatility regime doing? GEX and IV disagreeing is information; GEX alone is a partial picture.
Is there an event inside the window? Earnings, max pain dynamics into expiration, and macro prints all override the hedging anchor.

Then size the trade to the fact that steps 1 through 7 produced a distributional view, not a directional one.

See it on a live chart

Every level described here is computed daily across the whole optionable universe and published per ticker — for example, NVDA's gamma walls show the put wall, hedge wall and call wall with a rolling history, so you can see the levels move session to session instead of taking the anatomy on faith.

Three places to take this next:

Look up your own tickers on the Ideas board — free account, no credit card, live wall levels with the per-name reliability read attached.
See the levels mapped to how you actually trade — the persona matrix frames the same regime read for day, swing and long-horizon holders.
Read the full methodology — how these levels are computed, backtested and graded, including the places they do not work.

Using Gamma Exposure (GEX) to Anticipate Market Maker Hedging