Block caving is the lowest cost underground mining method but it has relatively high risks and a high up-front capital and development cost. It has application to large, often lower-grade orebodies where high production rates can provide good returns on the investment. However, it has a range of technical risks that are peculiar to block caving and the usual financial, political and environmental risks that may be exacerbated by the long lead-in times associated with the development of a block cave operation. With initial development costs that may be $100 M to $1000 M when the building of a treatment plant is also required, it is obviously important to minimise and manage the risks. In a greenfield project, commitment to a substantial proportion of the capital expenditure may be required to gain access to the deposit, and the decision to proceed at this stage would be based very largely on interpretations of drill hole data. In established mines the data may be supplemented by geological mapping and past experience, etc. The quality of this data is critical and yet it is frequently degraded through poor techniques and avoidable omissions. The final decision to proceed would be assessed on projected commodity prices, mining capital and operating costs, dilution and run of mine ore grades. The latter will vary over the life of the project and these can be estimated with drawdown or draw control modelling techniques. In this modelling, there are further risks in the assumed modelling parameters or that the modelling technique may not be applicable or appropriate. The paper addresses the most critical technical and operational risk areas and provides some practical precautions that can be taken to reduce these and improve the reliability of the geological and geotechnical models.