The changeable, variable and fragile nature of snow creates unique sampling challenges. In this paper, we present Star: an efficient, field-usable method for use in point-sampling spatial studies. We validate the accuracy of the Star method using a comparative Monte Carlo simulation of 1024 detailed samples of elevation data. As spatial snow studies generally attempt to find spatial continuity in layers and other properties, we use variogram ranges to compare the ability of four sampling methods to accurately reveal such spatial correlation. The three methods compared to Star represent gridded, gridded-random and pure-random methods; Star can be described as a linear-random method. The simulation shows Star’s accuracy to be comparable to both gridded and gridded-random methods. Following this comparative process we introduce a new measure of appropriateness for sampling methods: the correct convergence on a variogram model, which we call correct spatial correlation detection. This directly measures how many sampled areas become correctly classified with either spatially correlated or non-correlated variance for a given variogram model fit. In this measure, Star performs equivalently to the other methods, and in correct convergence it performs as well as pure-random sampling.