Competitive Search

The Web is a canonical example of a competitive search setting that includes document authors with ranking incentives: their goal is to promote their documents in rankings induced for queries. The incentives affect some of the corpus dynamics as the authors respond to rankings by applying strategic document manipulations. This well known reality has deep consequences that go well beyond the need to fight spam. As a case in point, researchers showed using game theoretic analysis that the probability ranking principle is not optimal in competitive retrieval settings; specifically, it leads to reduced topical diversity in the corpus. We provide a broad perspective on recent work on competitive retrieval settings, argue that this work is the tip of the iceberg, and pose a suite of novel research directions; for example, a general game theoretic framework for competitive search, methods of learning-to-rank that account for post-ranking effects, approaches to automatic document manipulation, addressing societal aspects and evaluation.