Students of programming languages in massive on-line open courses (MOOCs) solve programming assignments in order to internalize the concepts. Programming assignments also constitute the assessment procedure for such courses. Depending on their motivation and learning styles, students pursue different strategies. We identify which approach to attempt these assignments results in better performance. We predict students’ success from their online behaviour; and identify different paths students chose in order to complete the MOOC. We also discuss how students resign from the course after having difficulties with assignments. Moreover, we also predict, when would a student give-up (or succeed) submitting solutions to a given assignment.