Skip to main content

Gamma

Updated Jun 24, 2026 at 2:35pm

Research Draft Medium 1,295 words

Gamma is the second-order option Greek that measures how fast delta itself changes as the underlying price moves. If delta is the "speed" of an option's price (its directional exposure), gamma is the "acceleration" — the curvature, or convexity, of the option's value curve. Formally it is the second derivative of option value with respect to the underlying price, Γ = ∂Δ/∂S = ∂²V/∂S². Gamma is the Greek that makes options nonlinear: it is precisely the reason an option is not just a leveraged share of stock, and it sits at the center of both individual position management (gamma scalping, hedging) and a modern market-structure thesis (dealer gamma exposure). Its core tension is that gamma is bought and sold together with theta — convexity is never free.

How it's calculated

Under Black-Scholes, gamma is identical for a European call and put at the same strike (Wikipedia, Greeks (finance)):

Γ = e^(−qτ) · φ(d₁) / (S · σ · √τ)

where φ is the standard normal density, S the spot, σ implied volatility, τ time to expiry, q the dividend yield, and d₁ the usual Black-Scholes term. Practically, gamma is quoted as the change in delta per $1 move in the underlying. A long call with delta 0.40 and gamma 0.10 becomes a 0.50-delta call after a $1 rise; a short call with delta −0.25 and gamma 0.05 moves to −0.30 (tastylive examples).

Two behaviors fall straight out of the formula:

  • Moneyness: Gamma peaks approximately at-the-money (ATM) and decays toward zero deep in- or out-of-the-money, where delta is pinned near 1.00 or 0.00 and has little room to change.
  • Time: Here a common one-liner ("gamma rises into expiration") is half-wrong. For an ATM option gamma spikes as τ → 0 — the √τ in the denominator collapses — so a near-dated ATM option has explosive delta. For OTM/ITM options gamma falls toward zero into expiry. The "gamma rises near expiry" rule is only true at the money.

Long options (calls or puts) have positive gamma; short options have negative gamma.

How it's used in practice

Position read. Long gamma is a long-convexity bet: delta improves in your favor as you're right and shrinks as you're wrong — gains accelerate, losses decelerate. The price is paid in negative theta (time decay). Short gamma is the mirror: you collect theta but your delta moves against you on every swing, creating the unbounded-loss profile of a short straddle.

Gamma scalping. The classic long-gamma trade: hold a delta-hedged long-options position (e.g. long straddle, delta-flattened with stock), and as the underlying oscillates, the positive gamma keeps regenerating delta you can sell high / buy low. The scalps fund the theta bleed. It is profitable only when realized volatility exceeds the implied volatility you paid — gamma scalping is fundamentally a bet that realized > implied.

Gamma hedging. A pure delta hedge is accurate only for an instant; gamma is the error term that makes it drift. Desks neutralize gamma (using other options, since stock has zero gamma) so a delta hedge stays valid across a wider price range and survives gaps and vol spikes.

Dealer gamma exposure (GEX). Aggregating gamma across all open contracts produces a dollar-map of how much delta market makers must hedge per move. The standard story: when dealers are net long gamma they hedge counter-cyclically (sell rallies, buy dips), dampening volatility and pinning price; when net short gamma they hedge pro-cyclically (buy strength, sell weakness), amplifying moves. The "zero-gamma / gamma-flip" level is the modeled crossover between these regimes. This is now a mainstream lens for intraday SPX/SPY behavior, popularized by vendors like SpotGamma, Unusual Whales, and MenthorQ.

Adoption, debate & evidence

Gamma the Greek is uncontested textbook mathematics. Gamma as a market-structure signal is partly evidenced and partly folklore.

What academic work actually supports: hedging-driven flows do leave measurable footprints. Barbon & Buraschi's "Gamma Fragility" and Baltussen et al. tie net dealer gamma to intraday autocorrelation — positive gamma strengthens reversals, negative gamma strengthens momentum. One body of work estimates roughly 9.5%–13.4% of daily absolute returns on optioned stocks is attributable to option-hedge rebalancing, with rebalancing shifting the probability of extreme daily moves by an estimated 5%–40% depending on period (per a ScienceDirect study summarized in the literature). The Dim-Eraker-Vilkov 0DTE work and BSIC's review document a Granger-causal link from gamma exposure to short-horizon volatility. So the direction of the effect — short-gamma amplifies, long-gamma stabilizes — is real and replicated.

What is NOT established with the precision vendors imply: the sign and size of dealer inventory are estimated, not observed. GEX models hinge on an assumption about who is long vs short each strike (the common heuristic that customers buy puts / sell calls so dealers are the opposite), which is not universally true and breaks down in the 0DTE era where 60%+ of index option volume now turns over intraday. Vendor "hit-rate" stats (e.g. SpotGamma's claim that ~68–78% of sessions close inside its predicted range) are self-reported, lack peer benchmarking, and a wide range will contain most quiet days regardless of model quality. Treat specific gamma-flip price levels as one probabilistic input, not a mechanical trigger.

Strengths & limitations

Strengths. Gamma is the honest accounting of an option's nonlinearity — it is what you are truly buying or selling, more than direction. It explains pin risk near expiry, the violence of ATM 0DTE options, and a genuine slice of intraday index dynamics. For a hedger, ignoring gamma means a delta hedge that quietly fails exactly when it's needed (gaps, vol spikes).

Limitations. (1) Gamma is inseparable from theta and vega — there is no free convexity; long-gamma scalping loses if realized vol disappoints. (2) Black-Scholes gamma assumes a single, constant σ; real surfaces (skew) distort it. (3) The #1 misuse is treating dealer-GEX levels as deterministic — they are inferred from estimated positioning, ignore non-dealer participants, and degrade fast in the 0DTE regime. A "negative-gamma" tape raises the odds of trend persistence; it does not guarantee a move.

Sources

Flags / disputes: (1) Dealer-positioning sign is contested — vendors use heuristics (customers buy puts/sell calls) that aren't universally valid; sources even differ on the default. Stated as an assumption, not fact. (2) GEX vendor hit-rate stats are self-reported and not independently validated. (3) The "gamma rises into expiry" claim is true only for ATM options — corrected in-text.