Pricing discrete barrier and hindsight options with the tridiagonal probability algorithm

This paper develops an algorithm to calculate the Brownian multivariate normal probability subject to any preset error tolerance criteria. The algorithm is founded upon the computational simplicity of the tridiagonal structure of the inverse of the Brownian correlation matrix. Compared with existing pricing technologies without the ‘barrier too close’ problem, our calculation method can produce a more accurate and efficient analytic evaluation of barrier options monitored at discrete instants with well- or ill-behaved barrier levels, or discrete hindsight options, for a reasonably large number of monitorings.