Bulk sampling programs are typically applied in coarse gold mineralised systems to verify resource/reserve grades. In recent years, there have been controversies surrounding some results of underground bulk sampling exercises compared to the predictions using diamond drill data. Just how realistic is it to predict 100 t lots from drilling that was designed to model blocks of several thousand tonnes? This study looks at the variation of relatively small units within larger planned mining units and compares some of the geostatistical techniques which purport to provide tools for such studies. Relative difficulty of application, computer timings and summaries of results will be used to illustrate the likely variation in bulk sampling units as opposed to mining units and large-scale planning units. A case study based on publicly available data from a gold project in North America is presented. A simplistic and unconditional geostatistical simulation model demonstrated that up to 50 per cent of a relatively high-grade mining block could show as uneconomic low-grade subblocks. A subsequent study will consider a similar demonstration using uniform conditioning. Erroneous impressions of early sampling results can lead to unsupported prejudgements of the final results. Bulk sampling programs require detailed planning based on ore characteristics, where protocols must fit the mineralisation type. Any program requires definition of aims and data quality objectives across grade, geological, geotechnical and geometallurgical parameters.CITATION:Clark, I and Dominy, S C, 2017. Underground bulk sampling, uniform conditioning and conditional simulation - unrealistic expectations?, in Proceedings Eighth World Conference on Sampling and Blending , pp 3-20 (The Australasian Institute of Mining and Metallurgy: Melbourne).