This document discusses forecasting techniques for software projects. It introduces Little's Law which relates the number of customers in a system to the arrival rate and time spent in the system. It notes that traditional "agile" forecasting is heuristic while stochastic forecasting uses statistical models based on historic data. Building an accurate stochastic model requires accounting for factors like work in progress, cycle time, bugs, and scope changes. The model is used for Monte Carlo simulations to generate forecasts, with the 85th percentile often quoted as the forecast range. Reducing variance through stable processes improves forecast accuracy.