Within the mainstream phonological framework of Optimality Theory (OT), grammars are parameterized by how they prioritize or rank a given set of constraints. OT online learning consists of slight re-rankings triggered by exposure to a single piece of data at the time. This paper presents a new online model for the acquisition of phonotactics in OT. Convergence and correctness are analytically investigated and the proposed model is shown to be superior to existing OT online models.