Binomial option tree
Model the stock as going up or down at each step, then price the option backwards from expiry: every node is the discounted risk-neutral average of the two nodes after it.
01The formula
- how much the stock multiplies by in one step, up or down
- length of one step in years (T divided by the number of steps)
- the risk-neutral probability of an up move
- option values at the two nodes after this one
02See it move
03Where it is used
- Options desksPrice American options, which can be exercised early, by checking at every node whether exercising beats waiting.
- Interviews and teachingExplain risk-neutral pricing with a one-step tree before any calculus.
- Convertible and exotic desksUse trees and lattices for products with early exercise or call features.
04Why traders use it
It shows why option prices do not depend on the stock's real expected return: you can build the option out of shares and cash at every step, so only the risk-neutral probability q matters.
05Worked example
With 4 steps the tree prices the call at $9.971; Black–Scholes gives $10.451. Add steps and the two agree.
06How it is built
- At expiry the option is worth its payoff at each final node.
- One step back, hold Δ shares and some cash so the portfolio matches the option in both the up and the down state.
- That replicating portfolio costs , with as above, so that is the option's value.
- Repeat back to today. As steps grow, the tree converges to Black–Scholes.
07The common mistake
Using the real probability of an up move. The tree only prices correctly with the risk-neutral ; the real-world probability never enters.
08In Python and LaTeX
import numpy as np S,K,r,sigma,T,n = 100,100,0.05,0.2,1.0,4 dt=T/n; u=np.exp(sigma*np.sqrt(dt)); d=1/u; q=(np.exp(r*dt)-d)/(u-d) V=np.maximum(S*u**np.arange(n,-1,-1)*d**np.arange(0,n+1)-K,0) for _ in range(n): V=np.exp(-r*dt)*(q*V[:-1]+(1-q)*V[1:]) print(round(V[0],4))
u = e^{\sigma\sqrt{\Delta t}},\ d = \frac1u,\ q = \frac{e^{r\Delta t} - d}{u - d},\quad V = e^{-r\Delta t}\big[q V_u + (1-q) V_d\big]09Where it fits
Rather not code the maths yourself? Describe your trading idea and let AI build the bot.
Build a bot with AgenKitThe worked example and the model run the same code, so every number matches. Models simplify markets on purpose. This is education, not financial advice.