In this talk, I will use coral reefs as a model system to explore the promise of several alternative approaches to overcome this problem. I will begin with neutral ecological approaches that make highly restrictive symmetry assumptions, and largely fail to capture important features of community dynamics. I will then explore somewhat more flexible approaches that use statistical distributions to allow demographic traits to vary among species, but do not aim to attribute particular trait values to particular species. Such approaches can robustly characterize aggregate community properties, such as distributions of commonness and rarity and the relative importance of stochastic demographic fluctuations in community dynamics, but cannot be used for species-level inference. I will conclude by presenting work that uses contemporary, reduced-parameter statistical innovations to estimate species interactions and patterns of covariation in species’ responses to environmental fluctuations. This work provides remarkably statistically robust, species-level characterization of demographic characteristics and interactive effects on other species, providing important insights into how high-diversity assemblages are structured, and on the implications of that structure for ecosystem functioning.