Using the output of optimisation models to make real-time changes to railway timetables can be an effective way to reduce the propagation of delay. In this study, we develop a methodology for evaluating the fairness of such optimisation models with respect to competing train operators. Whilst both fairness and optimisation-based railway timetable rescheduling have both been widely studied, they have not previously been studied together. We propose definitions of fairness and efficiency for timetable rescheduling, and analyse the fairness of efficiency-maximising solutions for a case study with seven train operators. We also investigate the pairwise trade-offs between operators and show that the priority given to different train classes has an important impact on fairness.