Available routing strategies for traffic networks may be classified as feedback and iterative strategies. Recently, a new predictive feedback strategy has been developed incorporating the advantages of both kinds of strategies and attenuating their disadvantages. The new strategy runs a mathematical model once at each time step and bases its feedback routing decisions on the predicted traffic conditions. Preliminary investigation indicated that very satisfactory routing results could be achieved by use of this strategy. In this paper, the performance of the new strategy is evaluated in more detail by comparison with the feedback and iterative strategies.