Discrete dynamical models of   Centauri

We present a new framework for modelling discrete kinematic data. Current techniques typi-cally involve binning. Our approach works directly with the discrete data and uses maximum-likelihood methods to assess the probability of the dataset given model predictions. We avoid making hard cuts on the datasets by allowing for a contaminating population in our models. We apply our models to discrete proper-motion and line-of-sight-velocity data of Galactic globular cluster ω Centauri and find a mildly radial velocity anisotropy β = 0.10 ± 0.02, an inclination angle i = 50 ◦ ± 1◦, a V-band mass-to-light ratio Υ = 2.71 ± 0.05 M/L and a distance d = 4.59 ± 0.08 kpc. All parameters are in agreement with previous studies, demonstrating the feasibility of our methods. We find that the models return lower distances and higher mass-to-light ratios than expected when we include proper motion stars with high errors or for which there is some blending. We believe this not a fault of our models but is instead due to underestimates or missing systematic uncertainties in the provided errors.