Fall if it Lifts your Teammate: A Novel Type of Candidate Manipulation

We present a new interpretation of the traditional computational social choice framework, where what are traditionally the candidates are construed as the agents. The particular implementation in mind is the proposed system for determining the medal winners for sports climbing in the 2020 Olympic games. We consider the issues of ties and of potential manipulation with respect to this interpretation. Simulation results suggest that for the proposed system ties are unlikely to be a problem, but that there is at least potential for manipulation, of a novel type. We formalise this conception of manipulation axiomatically. The strongest axioms lead to an impossibility along the lines of Arrow's impossibility, while a small weakening leads to a possibility. We also provide a hardness result concerning the determination of possible manipulation.