This paper examines the design of longwall shields, focusing on those geotechnical aspects that are critical to either their success or failure during longwall extraction. The driver for the paper is the recognition that in many instances, longwall shields are assessed and designed along similar lines to that of roadway ground support systems, namely according to typical or normal geotechnical conditions. However the paper will contend that for longwall shields, this is inappropriate as unlike roadway ground support systems, longwall shields cannot be supplemented with additional or secondary support in localised areas of adverse geotechnical conditions. In other words, the longwall shield needs to be designed according to what are judged to be worst case or adverse geotechnical conditions, the outcome being that a well-designed shield will be significantly over-rated for the majority of its working life. However it will be argued that the additional capital cost of such shield design can be readily justified as a prudent risk-based outcome that is essential to minimising future business risks due to strata instability on the longwall face. Shield geometry is discussed including such relevant factors as leg angle, inclination of the top caving shield, canopy ratio, operating height range and tip to face distance, these all being well established longwall shield design considerations and most importantly, areas whereby inadequate design can render a longwall shield highly ineffective. The issue of tip to face distance is considered in detail, in particular the extent by which it is an important geotechnical design consideration. The critical importance of maximising set to yield ratio within practical operating limits is discussed. Overall the aim of the paper is to provide industry with a set of suggested guidelines for future use when designing longwall shields and hopefully to initiate discussion on a subject that unlike roadway ground support design is not well covered in its entirety in the published technical literature.