This document discusses software reliability models and the Rayleigh model in particular. It explains that reliability models can be static or dynamic, and the Rayleigh model is a dynamic model based on a Weibull distribution. The Rayleigh model uses parameters estimated from project data to project defect rates. Higher defect rates during development generally correlate with higher field defect rates. More defects found and removed earlier in the process yields better quality. Accuracy of models depends on valid input data and establishing predictive validity for different organizations.