This paper compares the performance of a new improved heuristic (STIL) proposed by Coleman and McKnew (1991) against those of the Wagner-Whitin (WW), cost-modified Wagner-Whitin (MWW) and some simple lot-sizing rules found to perform well by Zhao et al. (1995). It also investigates the impact of the lot-sizing rules on the selection of the parameters for freezing the master production schedule (MPS). The experimental settings include forecasting, master production scheduling and material requirements planning under a rolling time horizon. The result of the study shows that the STIL heuristic, the WW rule and the MWW rule do not really outperform the simple cost-modified Silver-Meal (MSM) and the Silver-Meal/Lot-for-Lot (SM/LFL) even though they require substantially higher computation time. The study also shows that the selection of the parameters for freezing the MPS is not significantly influenced by the selection of the lot-sizing rules.