GEX (gamma exposure)
advanced3 min
Definition
GEX (gamma exposure) estimates, strike by strike, how much delta dealers would need to trade for a 1% move in the underlying, if their options position matched an assumption. It is a model built on an assumed dealer position, not a measured inventory: venues do not publish who holds which side.
Why it matters
If dealers are long gamma, their hedging buys dips and sells rallies, which can dampen moves near large strikes. If they are short gamma, hedging chases the move. GEX tries to show where those effects might be largest. Its value depends entirely on whether the assumed dealer position is right, so OptionsMatch always shows the assumption next to the number.
How it works
Step 1: take open interest for calls and puts at each strike. Step 2: assume which side dealers hold. OptionsMatch's naive-OI assumption: customers buy calls and puts, so dealers are long the calls and short the puts. Step 3: compute each strike's gamma. Step 4: convert to USD per 1% move and sign by the assumption. Step 5: sum across strikes for net GEX, and look for where the cumulative sign changes (the gamma flip). Why F² and 0.01: gamma is the change in delta per 1 USD move. A 1% move is 0.01 × F USD, which changes delta by Γ × 0.01 × F coins per option, and those coins are worth F USD each, so the hedge trade is Γ × F² × 0.01 USD per option. Multiply by OI and sign by the assumption. Long dealer gamma means hedges trade against the move; short dealer gamma means hedges trade with it.
Formula
Per strike, in USD per 1% move: GEX = Γ × F² × 0.01 × (dealer-signed OI). Naive-OI sign: + call OI − put OI. Total GEX = Σ over strikes. Γ is Black-76 gamma per USD per coin, F the forward in USD, OI in coins.Worked example
EXAMPLEF = 100,000, 30 days, flat IV 34.99%, open interest (EXAMPLE): 95k 100 calls / 500 puts, 100k 300 / 300, 105k 400 / 50. Gamma per BTC (MODEL): 95k 0.00003397, 100k 0.00003972, 105k 0.00003616. GEX: 95k −1,358,789, 100k 0, 105k +1,265,608 USD per 1% (MODEL under the naive-OI assumption). Net ≈ −93,181. If dealers were actually short the calls and long the puts, every sign flips and net becomes +93,181: same data, opposite reading.
EXAMPLESheet inputs (EXAMPLE): BTC expiry forward F = 100,000, 30 days, r = 0, T = 30/365. Strike IVs from the C1 smile (MODEL): 90k 42.01%, 95k 38.04%, 100k 34.99%, 110k 41.55%. Model: Black-76, theta per day = yearly ÷ 365.25, vega per 1 vol point. Outputs are MODEL unless tagged otherwise.
Crypto angle
Crypto options open interest is concentrated on Deribit, with growing shares on other venues; GEX built from one venue misses the rest, so OptionsMatch labels it partial when venues are missing. Crypto flows include many retail call buyers and structured-product sellers, so the naive assumption can be wrong in either direction. Coin-settled OI must be converted to coins of underlying consistently. Hedging happens in perps and futures, and funding can change how dealers hedge.
Common mistakes
Treating GEX as a measured dealer book. Hiding the assumption. Mixing per-1% and per-USD units. Using stale OI. Building GEX from one venue and calling it the market. Reading a gamma flip as a price prediction. Comparing GEX across days without rescaling for F: GEX grows with F², so the same OI shows a larger number after a rally.
Not financial advice. Glossary definitions are education only. They are not venue legal terms. When a venue rulebook disagrees with this page, the venue wins for your account. Options can lose all premium paid; short options can lose far more.
Back to top