The new multi-step grid simulation module reduces the valuation time of computationally demanding instruments by applying grid technology on an instrument-by-instrument basis. Pricing analytics are often assigned at the portfolio level rather than at the instrument level, which results in unnecessarily reduced accuracy or performance. With Algorithmics' multi-step grid simulation module users can control pricing model selection at the instrument level, resulting in increased overall performance in risk calculations at the portfolio level. Recent tests showed up to a 10- fold performance gain with minor degradation in accuracy.
"Algorithmics' multi-step grid simulation can be used for more computationally demanding instruments, such as CMOs, callable bonds, and Bermudan swaptions. It increases the overall performance of risk calculation for the portfolio," says Diane Reynolds, Director of Analytics at Algorithmics. "Because the grid approach can be configured for each instrument it can be used for just these computationally demanding instruments while others are priced using standard pricing methods."
The module creates a low-dimensional grid and computes the instrument values of each node. During simulation, the instrument values under different scenarios are quickly calculated by interpolating between the previously computed values of the grid points. Users can specify the number of dimensions in the grid, the number of points in the dimension, and spacing of these points for ultimate flexibility in accuracy and performance in both single and multi-step simulations.