Discrete generation and rolling front breeding strategies are compared in terms of gain and inbreeding over a period of 40 years using stochastic simulation. In the rolling front strategy, crosses are made between the best available trees in each year, and new progeny trials are established using the crosses done in the previous year, rather than waiting for all crosses in that generation to be completed. For a given amount of resources, the rolling front strategy resulted in 25-35% greater gains per year, mainly due to a shorter generation interval. Inbreeding was also higher in the rolling front, although gains per unit of inbreeding were consistently greater than with the discrete generation strategy. Despite the smaller size of trials and greater imbalance between trials in rolling front, the results suggest that breeding value estimation using mixed-model BLUP is robust enough to ensure accurate prediction of breeding values and maintain the advantage of the rolling front strategy.