Much of the theory upon which the Sublevel Caving (SLC) method is based was developed in Scandinavia many years ago. It was based mainly on bin' theory and ellipsoid of draw' configurations derived from sandbox models. In many cases, classical SLC layouts influenced by these early theories obtained relatively poor results. Early dilution entry, low tonnage factors and low recoveries were commonly experienced. SLC soon gained the reputation of being a high dilution, low recovery and development intensive method. This led to SLC falling into disfavour generally, other than as a pillar reclamation method used at the end of the life of mines. More recently a number of mines, notably those in Australia, have adopted SLC as a primary extraction method and have achieved good recovery results. The reason for the better results is believed to be an outcome of modifications in some of the basic design parameters and operational procedures. This has led to questions being raised regarding the original bin theory' assumptions and classical SLC layout designs. It is clear that the fragmentation and flow characteristics occurring in a choke blast situation (compaction of waste, variable size and shape of fragmentation throughout the ring) conflicts significantly with the assumptions of regular loose flow of uniform, fine grained, material in a bin. Interactive versus independent draw conditions can also materially effect the outcome of a SLC operation. The good results obtained more recently can only be explained by using new' models of behaviour, which have to a large extent been borrowed from block caving. If these models are correct, they point to a number of design and operational changes that can be applied to improve the effectiveness of SLC. This paper discusses the factors that may not have been fully considered in early SLC design and modelling. It emphasises the importance of delayed dilution entry as a primary objective of sound SLC design and provides a basis for efficient SLC design and operation.