Delta Hedging: What It Actually Means to Be Delta-Neutral

Delta is the first Greek most people learn, and usually the one they understand least precisely — "delta tells you how much the option moves when the stock moves" is true but incomplete. What it actually enables is delta hedging: the technique that lets an options desk sell you an option without betting on which way the stock goes.

What delta actually is

Delta is the rate of change of the option's price with respect to the stock price — the derivative ∂C/∂S. For a call option:

Δ = N(d1)

where d1 is the same term from the Black-Scholes formula and N is the standard normal CDF. Because N(x) always returns a value between 0 and 1, call delta is always between 0 and 1: an at-the-money call has delta around 0.5, a deep in-the-money call approaches 1 (it starts moving almost one-for-one with the stock), and a deep out-of-the-money call approaches 0.

A delta of 0.5 means: for a $1 move in the stock, the option's price moves about $0.50, for a small move. It's a local approximation, not an exact prediction over a big move — that curvature is what gamma measures.

Why a market maker cares

Say a market maker sells you one call option with delta 0.60. If the stock rises $1, that option's value rises roughly $0.60 — and since the market maker is short the option, they lose roughly $0.60 per share of exposure. They didn't choose to make a bet on the stock going up or down; they just made a market and collected the bid-ask spread. The directional exposure is an unwanted side effect they now need to cancel.

They cancel it by buying 60 shares of stock for every option contract sold (1 contract = 100 shares of exposure, so delta 0.60 × 100 = 60 shares). Now:

• Stock rises $1 → option position loses ≈ $60, but 60 shares gained ≈ $60. Net ≈ $0.
• Stock falls $1 → option position gains ≈ $60, but 60 shares lost ≈ $60. Net ≈ $0.

That combined position — short the option, long the offsetting shares — is delta-neutral: to a first approximation, insulated from small moves in the stock in either direction. The market maker still profits from the bid-ask spread they charged, without needing a directional view.

Why it's not "set and forget"

Delta itself changes as the stock price moves — that's what gamma measures, the rate of change of delta. So a hedge that's neutral right now drifts out of neutral as soon as the stock moves at all, and has to be rebalanced — continuously in theory, at discrete intervals in practice (hourly, daily, whenever the desk's risk limits require it). This constant rebalancing is exactly what "gamma scalping" describes, and it's also where trading costs start to matter: rebalance too often and fees eat the spread you collected; rebalance too rarely and you're carrying real directional risk again.

Computing delta in Python

import math

def norm_cdf(x):
    return 0.5 * math.erfc(-x / math.sqrt(2))

def call_delta(S, K, T, r, sigma):
    d1 = (math.log(S / K) + (r + 0.5 * sigma**2) * T) / (sigma * math.sqrt(T))
    return norm_cdf(d1)

def put_delta(S, K, T, r, sigma):
    return call_delta(S, K, T, r, sigma) - 1  # from put-call parity

S, K, T, r, sigma = 100, 100, 1, 0.05, 0.20
delta = call_delta(S, K, T, r, sigma)
print(round(delta, 4))          # ≈ 0.6368
print(round(delta * 100))       # shares to buy per contract sold, ≈ 64

Put delta is always negative — a put gains value when the stock falls, so it moves opposite the stock — and it's a clean one-liner off call delta rather than a separate derivation, because put-call parity (C − P = S − K·e^(−rT)) holds for every value of S, so differentiating both sides with respect to S gives Δ_call − Δ_put = 1 directly.