In many decisions under uncertainty, individuals base their choices on what they learn from observing the choices of others as well as noisy private information, which can lead to herding behavior. However, herding may be discouraged if costs are incurred when an individual follows the action of too many others. For example, observing people on their way to a restaurant or a rival opening a new location provides information about the relative quality of their chosen alternative but also increases the likelihood that one who follows them will be penalized because the location has already reached capacity. Though such decisions have received little attention in the literature, they create a rich environment for exploring broader questions about herding behavior. In this book, I develop a model of herding with capacity constraints called the Restaurant Game and study play of this game in a lab experiment. The results provide new insights on the extent to which the common bias in herding behavior can be explained by Level-k thinking and other factors such as strategic uncertainty and cognitive ability.