Good statistics for measuring large-scale structure in the Universe must be able to distinguish between different models of structure formation. In this paper, two and three dimensional ``counts in cell" statistics and a new ``discrete genus statistic" are applied to toy versions of several popular theories of structure formation: random phase cold dark matter model, cosmic string models, and global texture scenario. All three statistics appear quite promising in terms of differentiating between the models.