Directed acyclic graph that represents a hierarchical Bayesian model of a set of N parallax measurements characterised by uncertainties σ_ϖ and true distances drawn from an exponentially decreasing density distribution of distances (see Eq. (18)). The scale length of the exponential decrease, L, is itself a model parameter that we can infer from the sample. Its prior is defined in this case for the sake of simplicity as a Gaussian distribution of mean μ_L and standard deviation σ_L.