Modeling User Performance in Multi-Lane Moving-Target Acquisition

Modern video games often feature moving target acquisition (MTA) tasks, where users must press a button when a moving target reaches an acquisition line. User performance models in MTA are useful for quantitative skill analysis and computational game level design, but have so far been constructed only for cases where there is a single lane for a target to appear and follow. In this study, the first user performance model is presented and validated for an MTA task with multiple lanes. The model is built as an integration of the existing MTA model and the drift-diffusion model, a model of human decision-making process under time-pressure. In a user study, we showed that the model can fit lane recognition error rates and input timing distributions with significantly higher coefficients of determination (R2) and accuracy than a baseline model.