The task of associating earthquake arrivals to their underlying source significantly increases in complexity with the number of phases to correlate. Continued advancements in sensor resolution, along with the recent adoption of deep learning techniques to seismic phase picking, have greatly increased the volume of phases now detected across seismic networks. As traditional phase association methods struggle to associate during periods of intense seismicity, phase association algorithms, therefore, need to evolve in tandem to accurately correlate the latest pick catalogs. We adopt the random sample consensus technique widely used throughout the computer vision community, proven to perform parameter estimation in the presence of significant outliers, and apply this to the phase association problem. Our algorithm, Hyperbolic Event eXtractor (HEX), is stress-tested on a synthetic dataset with a receiver distribution located throughout the northern Chilean subduction zone and preliminary results indicate the accurate association of events up to ~15s apart. The synthetic results indicate that HEX allows seismologists to perform accurate event correlations on a finer scale than historic, grid-based association algorithms. ... mehrIt is anticipated that HEX will improve the resolution of seismic catalogs and enable the extraction of greater amounts of information from large pick datasets.