We present an argument concerning the temporal dynamics of the policy positions of the main opposition parties in a legislative body; what we label competitor parties. In the article we ask how competitor parties react when losing an election, and what happens if they continue to lose election upon election. It is argued that competitor parties from the outset are trying to maximise both policy- and office-seeking preferences. We posit that this is best achieved by moving away from the median towards the party base. Yet in the event that electoral success continues to elude the competitor party, a process of self-reflection, or learning, will set in within the party because the policy position of the competitor party is clearly unacceptable to pivotal voter groups. In this event, the competitor party will move strongly towards the policy position of the incumbent party, thereby neutralising any advantage the incumbent party may previously have had in terms of policy position. By giving up on its policy preferences, the competitor party maximises total utility because then, if nothing else, it becomes more likely to win office. If the competitor party continues to lose, a new cycle of divergence and convergence will set in, generating a wave-like pattern. The argument is illuminated using two well-known examples from the United States and the United Kingdom and tested employing a unique large-N data set from Danish municipalities. We show that the temporal dynamics of competitor parties are as expected.