Brownian Bridge for Barrier Option Pricing Explained
Learn how a Brownian bridge estimates missed barrier crossings between Monte Carlo dates and why endpoints alone can bias continuously monitored option prices
Direct answer
A Brownian bridge conditions a continuous diffusion path on its simulated endpoints. In barrier-option Monte Carlo, it estimates the chance that the path crossed a barrier between two dates, reducing the bias created when discrete samples are treated as continuous monitoring
Endpoints do not reveal every crossing
Suppose a simulated price is below an up barrier at both the start and end of a one-day interval. The continuous path can still cross the barrier and return below it
Checking only daily or hourly endpoints therefore tends to miss knock-out or knock-in events when the contract is continuously monitored
More time steps reduce this error but can be expensive and still leave residual monitoring bias
The bridge fills in conditional path uncertainty
For Brownian motion, conditioning on two endpoint values gives an intermediate distribution whose mean follows the straight line between them
Its variance is zero at both endpoints and largest inside the interval. That conditional fluctuation is the Brownian bridge
After transforming a diffusion into a suitable log-price form, the same idea provides an interval-by-interval crossing probability under the stated model
A near-barrier interval needs more than one quote
Imagine an up-and-out level of 110, with simulated endpoint prices of 104 and 106 over a short interval
Neither endpoint triggers the barrier. A bridge adjustment still assigns a positive crossing probability because the unobserved path can rise above 110 and return
The probability depends on distance to the barrier, interval length, and modeled variance. It should not be replaced with a fixed percentage
Monte Carlo can use weighting or conditional draws
One method multiplies each path payoff by its interval survival probabilities. Another samples whether a crossing occurred using bridge probabilities
Weighting often lowers variance because it avoids an extra binary draw, while conditional sampling can be useful when the crossing time itself affects cash flows
Both methods must use the contract's actual observation convention. Some listed products monitor only at specified times rather than continuously
A bridge is model correction, not market certainty
The standard formula assumes a continuous diffusion with the chosen volatility, interest-rate, and dividend inputs over the interval
Jumps can cross a barrier in a way that a continuous bridge understates. Stochastic volatility, local volatility, discrete dividends, and price limits also change the model
Validate against a finer-grid calculation, stress the barrier distance and volatility, and keep the legal contract terms separate from a convenient simulation approximation
Common questions
Why is a Brownian bridge useful for barrier options?
It estimates whether a continuously monitored barrier was crossed between simulated dates, where endpoint-only simulation can miss the event
Does a Brownian bridge make a barrier simulation exact?
It is exact only for the specified continuous diffusion setup and monitoring convention. Jumps, changing volatility, and contract details can require another model
Should I simply add more Monte Carlo time steps?
Finer steps help, but the bridge can remove much of the between-step bias more efficiently. It is still wise to compare both approaches as a numerical check