The occurrence of coarse gold particles leading to grade complexity and sampling challenges is a common feature of many gold deposits. Poorly designed sampling protocols applied to these deposit types can lead to an excessively high fundamental sampling error. Together with other errors, the fundamental sampling error contributes to the nugget effect. Proper sample collection, preparation and assay protocols are required to minimise this error, and hence reduce the total nugget effect. Reduction of the nugget effect is a key requirement for subsequent geostatistical resource estimates to be effective. A case study from the Nalunaq gold mine in Greenland is presented, which is characterised by notable quantities of coarse gold (>30 per cent +100 m gold) hosted in quartz veins. Diamond drilling provides a good measure of gross geological continuity at practical grid spacings, but is not considered to represent the mineralisation effectively. On-reef development with face and bulk sampling, together with detailed geological studies were undertaken during the resource delineation phase. This contribution describes the nature of the Nalunaq deposit, its gold particle size characteristics, how sampling and assay protocols were developed (including QA/QC issues), application of Gy's sampling theory and its calibration, and how the data was subsequently applied to grade estimation.