Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
In a biased contest, one of the players has an advantage in the winner determination process. We characterize a novel class of biased contest success functions pertaining to such contests and provide necessary and sufficient conditions for zero bias to be a critical point of arbitrary objectives satisfying certain symmetry restrictions. We, however, challenge the common wisdom that unbiased contests are always optimal when contestants are symmetric ex ante or even ex post. We show that contests with arbitrary favorites, i.e., biased contests of symmetric players, can be optimal in terms of various objectives such as expected aggregate effort, the probability to reveal the stronger player as the winner or expected effort of the winner.