Help us double down on what's working, instead of guessing. Takes 5 seconds, totally optional.
gex · 20 min read
Complete Guide to Trading Gamma Exposure (GEX)
Everything a GEX trader needs in one place: how dealer hedging moves price, positive vs negative gamma regimes, the math behind net GEX by strike, reading settled vs flow-GEX, the trading playbook for each regime, risk sizing, the kinks that trip up experienced traders, worked SPY and NVDA examples, and the FlashAlpha endpoints and MCP connector that put it all together.
Free: a supported individual equity, such as AAPL, and one listed GEX expiry after today; 5 requests/day. SPY/QQQ single-expiry API access: Basic and above. Full-chain GEX: Growth and above. 0DTE and supported CME futures: Growth and above. Historical replay: Alpha and above.
What Is Gamma Exposure - and Why Does It Move Price?
Options dealers (market makers) are in the business of providing liquidity, not taking directional bets. When they sell a call, they acquire a short delta position in that option. To stay delta-neutral - to remove directional risk from their book - they buy the underlying. When the underlying rises and the call's delta increases, they must buy more. When the underlying falls and delta decreases, they sell some back. This continuous buy-as-price-rises, sell-as-price-falls cycle is called delta hedging, and its intensity at any given price level is proportional to gamma.
Gamma exposure aggregates signed sensitivities across the included contracts. Under the stated assumptions it estimates dollar-delta sensitivity per 1% underlying move, rather than stock orders every dealer must execute. The same timestamp and expiry scope are needed when comparing observations.
The useful hypothesis is conditional: if the assumed dealer signs and hedge behavior are representative, gamma gives context for how hedging could respond to price changes. Open interest alone cannot identify the actual dealer book, its hedges or the next price move.
When net GEX is positive, dealers are collectively long gamma - they own more gamma from the puts and calls they have sold and then hedged, than they are exposed to as short gamma from options they have bought. Long-gamma dealers buy dips and sell rallies to maintain delta neutrality. The effect on price is stabilizing: every move away from the highest-OI strike encounters natural dealer resistance. Price tends to pin, mean-revert, and respect walls.
Practical signs of a positive-gamma environment:
Intraday range compresses relative to the implied move. Realized vol runs below IV.
Price bounces cleanly off round numbers and known OI concentrations.
Call and put concentrations may coincide with observed resistance and support; breaches remain possible.
Range-dependent premium-selling rules need separate validation, including the sessions when the range fails.
Negative Gamma: The Momentum Regime
When net GEX is negative, dealers are short gamma. They need to sell as price falls and buy as price rises to stay delta-neutral - the exact opposite of the stabilizing dynamic above. Their mechanical hedging amplifies price moves. A 0.5% drop triggers dealer selling, which pushes price down further, which triggers more selling. This feedback loop is the mechanical engine behind many of the most dramatic intraday moves in liquid names.
Practical signs of a negative-gamma environment:
Intraday realized vol exceeds recent averages. Gaps extend rather than reverse.
Breakouts from walls or levels hold and follow through.
Momentum and mean-reversion rules can behave differently; measure both with matched execution assumptions.
Option buyers (long calls or puts) benefit; premium sellers face unexpected moves.
The Gamma Flip: The Regime Dividing Line
The gamma flip is a hypothetical underlying price where aggregate GEX crosses zero after repricing the included options. Keep expiry scope, open interest, volatility assumptions and dealer-position signs fixed. It is not a cumulative sum of the current strike bars. A crossing can be absent, multiple or unstable; inspect gamma_flip_status and the net GEX at the actual price rather than assuming that every price above a flip is positive gamma.
A flip can change because spot, IV, time to expiry, the included expirations or estimated positions change. Settled OI stays fixed between official updates while Greeks can still move. Flow-GEX estimates intraday position changes; it is not a direct observation of true open interest.
The Math: Net GEX Per Strike and the Dealer-Position Sign Convention
The standard formula for net GEX at a single strike K is:
Net GEX Per Strike
$$ \text{GEX}_K = \underbrace{\Gamma_K^{\text{call}} \cdot \text{OI}_K^{\text{call}}}_{\text{call contribution}} - \underbrace{\Gamma_K^{\text{put}} \cdot \text{OI}_K^{\text{put}}}_{\text{put contribution}} $$
The minus sign on the put contribution encodes the standard GEX dealer-positioning convention: dealers are modeled as long calls and short puts - the consensus assumption that customers buy puts as downside hedges and write calls for income, leaving dealers on the opposite side. Gamma is always positive for any long option and negative for any short option, whether call or put. So a long call carries positive (long) gamma and a short put carries negative (short) gamma; the formula therefore adds call gamma and subtracts put gamma. When the long gamma dealers hold from calls outweighs the short gamma from puts, net GEX is positive and dealer hedging is stabilizing.
To express the signed per-strike contribution as dollar-delta notional per 1% underlying move, multiply by the common spot squared, the contract multiplier and 0.01:
where S is the same underlying spot price for every strike, not that option's strike. The 100 multiplier applies to an unadjusted US equity option. The example below is hypothetical, prepared October 16, 2026; it is not a current SPY quote.
Concrete example - two SPY strikes: Suppose SPY is at S = $595. At the $595 call, gamma is 0.018 and call OI is 22,000 contracts; put gamma is 0.016 and put OI is 14,000 contracts. At the $590 strike, call gamma is 0.009 with 8,000 call OI, put gamma is 0.021 with 31,000 put OI.
Summing just these two strikes: +$60.9M − $205.0M = −$144.1 million net GEX. If this pattern extends across the full chain, the net book is negative-gamma - the model suggests hedging that could reinforce a move, subject to actual inventory and execution. In practice the API returns this aggregated dollar figure as net_gex in the /v1/exposure/gex response.
Sum across all strikes to get net dealer dollar GEX:
Under the stated sign convention, positive NetGEX represents an estimated long-gamma book and negative NetGEX an estimated short-gamma book. Actual hedge activity can differ. Open interest does not reveal who owns each side. The calls-positive/puts-negative convention models dealers as long calls and short puts; that is an assumption, not observed dealer inventory. A long option has positive gamma whether it is a call or a put; a short option has negative gamma.
Why the Dealer-Position Sign Convention Is an Approximation
The standard model assumes one uniform dealer position across all open interest. In reality, some options are dealer-to-dealer, some large institutions buy puts as portfolio hedges (making the dealer short the put, and so short gamma on that position), and retail flow is not monolithic. In practice this means GEX is best treated as conditional sensitivity context rather than a precise forecast of the exact shares dealers will trade. The effective OI methodology partially addresses this by confidence-weighting buy vs sell flow classification intraday.
Reading the Data: Settled GEX, Flow-GEX and Dealer Risk
Settled OI Is a Daily Input; Greeks Can Still Change
Settled GEX uses official daily open interest together with the calculation's price, IV and time inputs. A daily OI field does not freeze every GEX output for the day. Record the response timestamp and expiry scope before comparing readings; the daily position data is not a direct observation of the current dealer book.
Flow-GEX Estimates Intraday Position Changes
Flow-GEX combines settled OI with estimates of intraday position changes from classified options flow. It can disagree with settled exposure because the position inputs differ. Classification, open/close inference and coverage introduce uncertainty; the flow surface is not the true dealer book and does not become universally correct at a particular clock time.
Compare the two around unusually active sessions, expiry concentration or events, using matched symbols, timestamps, expiry scopes and polarity conventions. Measure the differences rather than assuming a fixed dollar shift or a fixed time-of-day cutoff. See live versus settled GEX and flow versus exposure endpoints.
Dealer-Risk Labels Are Model Outputs
The dealer-risk endpoint compares modeled settled and live exposure. Inspect its documented fields and unavailable states. A label such as amplifying or dampening does not prove actual dealer orders, forecast the next move or select an entry.
Do Not Read a Portfolio Flip From Strike Bars
Each strike bar shows a contribution at the same current underlying price.
Call/put walls use their own concentration methodology. They are not interchangeable with the tallest positive/negative net-GEX bars.
A portfolio flip is a zero of aggregate GEX after repricing at hypothetical spots; it is not a zero along the strike axis.
Positive and negative contributions can coexist. Sum the included positions under one unit and sign convention.
A visually balanced profile does not establish a profitable pinning strategy.
For an inspectable profile, use the saved example below and the levels reference.
The GEX Trading Playbook
Positive Gamma Regime: Mean Reversion and the Pin
When modeled net GEX is positive, the following are range-oriented hypotheses to test after costs:
Signal
Trade Idea
Logic
Price rallies into call wall
Fade / short, target the flip or midpoint
Dealers sell into the call wall; resistance is mechanical
Price dips toward put wall
Fade / buy, target midpoint or call wall
Dealers buy at the put wall; support is mechanical
Price near magnet strike, pin score elevated
Sell premium - butterfly or iron condor centered on magnet
Pin mechanics pull price to the OI concentration as expiration nears
Execution check: define the spread payoff and exit rule before entering. A wall breach need not execute at the chosen stop price; include fees, slippage and event gaps. The setup table describes hypotheses to test, not validated win rates.
Negative Gamma Regime: Trend Following and Breakouts
When modeled net GEX is negative, the following are momentum-oriented hypotheses to test after costs:
Signal
Trade Idea
Logic
Price breaks below put wall
Follow the breakdown with puts or short
Dealers sell with the move; mechanical amplification supports the trend
Price breaks above call wall in negative gamma
Follow the breakout with calls or long
Dealers buy with the move above the call wall
Spot approaches gamma flip from below
Wait for clean reclaim before reversing to positive-gamma playbook
A failed flip reclaim is a short-covering opportunity, not a regime change
Gamma acceleration 4x+ with 2hr to close
Momentum debit spread in direction of trend
Late-session negative gamma cascades are fast; defined risk structures limit damage from whipsaws
Execution check: a debit spread limits loss to the debit plus costs if held under its contract terms; a long option also has premium-limited loss. The short leg caps upside and may introduce assignment considerations. Neither the structure nor a flip crossing establishes a profitable signal.
Using the Flip as a Conditional Reference
First verify net GEX at the actual price, the crossing's orientation and the returned status. A flip is not a floor, ceiling or guaranteed entry. Record the timestamp and expiry scope before comparing it with price. If it is absent or unstable, leave the level unavailable and test a rule that can handle that state.
Risk and Position Sizing
A gamma regime alone does not justify a fixed account-risk percentage. Evaluate contract payoff, executable spreads, fees, liquidity and gap exposure. For an explicitly defined loss budget, a mechanical upper bound is:
Defined-Loss Contract Count$$N = \left\lfloor\frac{\text{chosen loss budget}}{\text{maximum loss per contract including costs}}\right\rfloor$$
That arithmetic does not select the budget or establish a strategy's edge. Stops may execute beyond their trigger; a theoretical stop loss is not a guaranteed maximum loss. The worked counterexample below distinguishes a $20 early-exit profit from a $110 maximum condor loss before costs.
The /v1/pricing/kelly endpoint uses option-pricing inputs. It is not a general win-rate/average-win-loss sizing calculator. Estimate a strategy's return distribution independently before using any optimal-sizing model.
The Kinks and Common Mistakes
This is the section most GEX guides skip. The mechanics above work - but they break down in specific, predictable ways. Here are the most common kinks, in order of how often they surprise traders.
1. Stale Settled OI by Midday
Settled open interest is updated on a daily schedule. New trades can change positioning intraday, but neither a fixed turnover percentage nor an 11 AM cutoff is established here. Compare settled and flow estimates at matched timestamps and expiry scopes, and keep the estimation limitations visible.
2. Walls Shifting Intraday
Even flow-GEX shows walls that move. A large institutional order in 0DTE calls can shift the call wall $2-3 in a single print. This is most common in the first 30 minutes of trading (as overnight positioning is established) and in the last 60 minutes (as large players roll or close expiring positions). Strategies that enter near a wall should use a buffer - don't short the call wall to the penny; leave a $0.50-1.00 buffer and require price to show rejection before entering, rather than entering mechanically at the wall price.
3. GEX Doesn't Tell You Everything - Vanna and Charm
GEX captures the gamma component of dealer hedging. It says nothing about vanna (the change in delta with respect to implied volatility) or charm (the change in delta with respect to time). On days when the VIX is moving significantly, vanna flows can dwarf gamma flows - especially for index ETFs where a large vanna exposure from the options skew can add or subtract hundreds of millions of dollars of synthetic share buying or selling that GEX does not capture. Similarly, charm flows compound near expiration as options lose time value. For a full picture, combine GEX with the vanna and charm exposure data - particularly on high-vol days or into the final hour of expiration.
4. 0DTE Distortions
Near-the-money 0DTE options can concentrate gamma in a narrow price region; far-out-of-the-money options may carry very little gamma. Compare today's expiry with the full chain rather than assuming a fixed 0DTE share. Use actual listed dates and Growth and above access; the two scopes can legitimately disagree.
5. Low-Liquidity Names
GEX works best when open interest is high and distributed across multiple strikes. On low-liquidity single names (small-cap stocks, names with thin options markets), OI may be concentrated in just 2-3 strikes, the bid-ask spreads are wide, and dealer hedging may not be a significant driver of price action at all - the underlying's own order flow dominates. A rule of thumb: if a name has fewer than 5,000 contracts of total OI, treat GEX signals with significant skepticism. The larger the total OI relative to average daily volume, the more reliable the GEX signals.
6. Index vs Single-Name Differences
SPX is a cash index; SPY and QQQ are ETFs. Their contract, exercise and settlement conventions differ. Liquid products may offer broader options participation, but news and macro flows can still overwhelm hedging. A customer buying a put leaves its dealer counterparty short that put and short gamma; the reverse trade leaves the dealer long gamma. OI alone does not distinguish these cases.
7. False Flips
An apparent flip breach can fail as a trading hypothesis even if both calculations are correct. Different expiry scopes, IV assumptions, position estimates or timestamps may produce different crossings. A null or unstable flip is not evidence that the market has no gamma; it means the crossing was not established under that calculation. Check the returned status before using a numeric level.
8. A Strike-Bar Zero Is Not a Portfolio Gamma Flip
The portfolio gamma flip is a zero of total modeled GEX as a function of hypothetical underlying price. A zero or sign change in the per-strike bar chart answers a different question. Reprice all included options at each hypothetical spot; do not walk along the strike axis and call its cumulative sign change a flip.
GROWTH PLAN
Compare settled and estimated intraday positioning on Growth
The /v1/flow/gex and /v1/flow/dealer-risk endpoints on Growth and above let you compare settled inputs with estimated intraday position changes. They do not reveal true dealer inventory or guarantee that a wall or flip stays accurate. Raw simulator state has separate Alpha and above access.
Start with the eligible Free equity example. SPY full-chain requests require Growth and above; estimated flow does not remove model or execution risk.
Worked Examples: Inspect the Data and a Failure Case
A reproducible saved SPY replay
The downloadable sample requests SPY at 2026-03-05T15:30:00 in America/New_York (20:30 UTC), across all expirations. The saved GEX response has spot 679.46, net GEX −$14.5045B per 1% move and numeric flip 687.1717. The companion levels response nests call wall 685 and put wall 675 under levels.
Inspect the request, units, assumptions and hashes; download the runnable evaluation pack and notebook. Run its offline fixture mode before making authenticated requests. This is the current engine applied to archived inputs, not JSON captured during that March session. Original capture time is unrecorded, and the payload predates flip-certification fields: a numeric flip alone does not certify its stability.
Counterexample: a wall does not prevent a loss
Hypothetical calculation prepared October 16, 2026: sell a $2-wide iron condor for $0.90 credit. If it is later bought back for $0.70, gross profit is $20 per unadjusted contract, not $70. Capturing $70 would mean buying it back for $0.20. If it settles at maximum loss, the $2 width less the $0.90 credit loses $110, before fees. These are alternative outcomes; a position closed early cannot also be counted as held to expiry.
A positive-GEX estimate and unchanged walls do not rule out an event or directional order flow breaching the range. Record the timestamp, expiry scope, flip status, spread entry/exit and costs for losing observations as well as winning ones. The saved replay demonstrates parsing and provenance, not the profitability of this condor.
The GEX Tooling: FlashAlpha Endpoints and the MCP Connector
REST API Endpoints
FlashAlpha exposes four primary endpoints for GEX trading:
Endpoint
Tier
What You Get
/v1/exposure/gex/{ticker}
Free for an eligible single-equity future expiry; Growth and above for full chain
Settled GEX by strike, net GEX, and gamma flip - using settled OI and current model inputs. Specify the expiry scope and check symbol access. Use with /v1/exposure/levels to get call wall and put wall as derived key levels.
/v1/flow/gex/{ticker}
Growth
Flow-GEX: recomputes the full GEX surface against effective OI (settled + intraday simulator). Returns live_net_gex, live_gamma_flip, and live per-strike profiles. Shows how today's trading is shifting the dealer book in real time.
/v1/flow/dealer-risk/{ticker}
Growth
Composite settled-vs-live shift metric: returns flow_direction (amplifying / dampening), flow_gex_pct_shift (how much GEX has moved since open), plus settled and live GEX/DEX side by side. The one-line pre-trade regime check.
/v1/exposure/levels/{ticker}
Eligible equity levels on Free; symbol-class gates still apply
Computed key levels: gamma flip, call wall, put wall, zero-gamma level, and high-vol move targets. Clean reference levels without the full by-strike breakdown.
For an eligible Free evaluation, download first-gex.py: it discovers an AAPL expiration after today, makes one GEX request, and prints the response timestamp and flip status. The quick-start shows the complete code and error handling.
For a paid SPY full-chain workflow, use the runnable dashboard tutorial. Its GEX and levels requests share all-expiration scope, parse the nested levels object and keep unavailable or uncertified flips distinct from a numeric zero. Do not combine single-expiry GEX with full-chain walls and describe them as one surface.
The GEX MCP Persona Connector - Bring GEX Into Claude, Cursor, or Any AI Tool
FlashAlpha's most direct integration path for GEX traders is the GEX persona MCP connector - a Model Context Protocol server tuned specifically for dealer-gamma and key-level work. Add it to any MCP-compatible AI tool (Claude, Cursor, Windsurf, VS Code) and you can ask questions like "Is SPY in positive or negative gamma right now?" or "What are NVDA's key GEX levels?" and get live, data-backed answers without writing a single line of code.
Two endpoints, depending on your client:
Client
URL
Auth
claude.ai (web browser)
https://lab.flashalpha.com/mcp-oauth/gex
OAuth - sign in once with your FlashAlpha account
Claude Desktop, Claude Code CLI, Cursor, Windsurf, VS Code
https://lab.flashalpha.com/mcp/gex
API key - pass X-Api-Key header or tell Claude your key in the first message
The GEX persona connector exposes a curated subset for dealer-gamma and key-level tasks. Check the current tool list in the setup guide; connecting does not change your API entitlements or allowance. If you use Claude for broader options work, the generic server at https://lab.flashalpha.com/mcp exposes the full catalog in one connection.
To add the connector in Claude Desktop:
# Add via Claude Code CLI
claude mcp add flashalpha-gex --transport http https://lab.flashalpha.com/mcp/gex
# Then in Claude: "My FlashAlpha API key is YOUR_KEY. What regime is SPY in?"
Add the GEX persona connector to Claude or Cursor and query dealer positioning, walls, and regimes in plain English. Choose OAuth or API-key setup according to your client's supported connection method.
Dollar GEX per 1% move sums signed option gamma times open interest, the correct contract multiplier, spot squared and 0.01. The standard calls-positive/puts-negative convention estimates dealer signs; OI alone does not identify actual dealer inventory.
Settled GEX uses daily settled OI; its Greeks can still change with price, volatility and time. Flow-GEX estimates intraday position changes. Compare the same symbol, timestamp and expiry scope; neither calculation observes every dealer position.
The modeled sign can change when price, volatility, time, the included expirations or estimated positions change. Recheck net GEX and flip status instead of inferring the regime solely from being above or below an old flip.
Flow-GEX and dealer-risk endpoints require Growth and above. Free can evaluate GEX on a supported individual equity and one listed expiry after today, with 5 requests/day. ETF/index single-expiry requests require Basic and above; full-chain GEX and 0DTE require Growth and above. Historical replay requires Alpha and above.
It can be computed for supported optionable equities, ETFs and indexes, with different access gates and contract conventions. Liquidity, events and the assumed position signs affect usefulness; no universal OI threshold establishes reliability.
Open interest does not reveal who owns each side. The calls-positive/puts-negative convention models dealers as long calls and short puts; that is an assumption, not observed dealer inventory. A long option has positive gamma whether it is a call or a put; a short option has negative gamma.
A wall is an options concentration, not an enforced price boundary. Directional flow, events, changes in IV or positions, stale inputs and mismatched expiry scopes can invalidate a reversal hypothesis.
Follow the current MCP setup guide for your client. The GEX persona uses https://lab.flashalpha.com/mcp-oauth/gex for OAuth-capable clients or https://lab.flashalpha.com/mcp/gex for API-key clients. Connecting does not change endpoint entitlements or request allowances.
Related Reading
Flow-Signed GEX (polarity=flow) - sign dealer gamma by the measured position from today's classified tape, not calls-positive/puts-negative
Use GEX as a conditional estimate: verify its units, dealer signs, timestamp, expiry scope and flip status, then test the actual entry and exit rule with costs and losing cases. Start with the reproducible sample above and choose API access by the workflow you can evaluate.
