Binomial Option Pricing Model Explained

The binomial option pricing model, developed by Cox, Ross, and Rubinstein in 1979, values an option by imagining the underlying stock taking small steps up or down over time. Working backward from expiry, where the option's value is obvious, it computes today's fair price one step at a time.

This matters because the model is the clearest window into how all option pricing works: no-arbitrage replication and risk-neutral valuation, without any calculus. It also handles American options, which can be exercised early, something the famous Black-Scholes formula cannot do directly.

Key Concepts

• Up factor (u) and down factor (d): the multiples by which the stock price moves in one step, for example u = 1.2 and d = 0.9.
• Risk-neutral probability (p): the artificial probability that makes the stock's expected growth equal the risk-free rate; it is a pricing device, not a forecast.
• Replicating portfolio: a mix of stock and risk-free borrowing that exactly mimics the option's payoffs; its cost is the option's fair price.
• Backward induction: valuing the option at expiry first, then discounting expected values step by step back to today.
• Convergence: as the number of steps grows and u and d shrink appropriately, the binomial price converges to the Black-Scholes price.

Formulas

For one step with risk-free growth factor (1 + r) per step: risk-neutral probability p = [(1 + r) − d] / (u − d). Option values at the up and down nodes: C_u = max(uS − K, 0), C_d = max(dS − K, 0). Today's value: C_0 = [p × C_u + (1 − p) × C_d] / (1 + r). The model assumes no arbitrage, frictionless trading, and that the stock moves only to the two specified nodes each step.

Worked Example

Price a one-period European call with stock price S = ₹100, strike K = ₹105, u = 1.2, d = 0.9, and risk-free rate 5% per period. At expiry the stock is either 100 × 1.2 = ₹120 or 100 × 0.9 = ₹90. The call values there are C_u = max(120 − 105, 0) = ₹15 and C_d = max(90 − 105, 0) = ₹0. The risk-neutral probability is p = (1.05 − 0.9) / (1.2 − 0.9) = 0.15 / 0.3 = 0.5. Today's fair value is C_0 = (0.5 × 15 + 0.5 × 0) / 1.05 = 7.5 / 1.05 = ₹7.14. Notice the actual probability of an up move never entered the calculation: only u, d, r, and the strike matter. For a two-period tree you would repeat this at each node, and at any node an American option would be exercised early if its immediate exercise value exceeded the discounted expectation.

Common Misconceptions

• p is the real probability of the stock rising. It is the risk-neutral probability implied by no-arbitrage; real-world probabilities are irrelevant to the price, which is why the model works without forecasting.
• Volatility does not appear, so it is ignored. Volatility is encoded in the spread between u and d: wider factors mean a more volatile stock and a more valuable option.
• The model only works for one step. Recursion extends it to any number of steps, and with many small steps it reproduces Black-Scholes values closely.

Real-World Applications

Binomial trees price American stock options with early exercise, value employee stock options, teach derivatives in every finance course, and price simple exotic options by customizing the tree's payoffs at each node.

The binomial model prices options by rolling back risk-neutral expected values through a tree of up and down moves, revealing that option value comes from the range of outcomes and the risk-free rate, not from anyone's forecast of the stock.