“Favorites,” individuals whom audiences expect to be victorious in competition, have little to gain and much to lose. “Underdogs,” who are anticipated to fail, have much to gain and little to lose. Recent theory extends this logic and indicates that the underdog label may increase risk-taking. We challenge this work by drawing upon prospect theory and developing a model that argues that favorite and underdog risk-taking behavior reflects their anticipation of wins and losses and the loss and gain frames associated with violating expectations. Two key predictions define this model. First, favorites will be more risk-taking than underdogs because they anticipate winning, and losing will be framed as a loss of that anticipated win. Second, we argue that current theorizing has created a false dichotomy, ignoring individuals defined by their unclear performance expectations, whom we term “wildcards.” For wildcards, we argue that risk-taking will vary depending on the emotional intensity of the decision-making environment. We test and find support for our theorizing across three studies: two archival studies (mixed martial arts and competitive chess) and a pre-registered experiment.