Using Simultaneous Best Response to Find Symmetric Bayes-Nash Equilibria in Auctions

Finding Nash and Bayes-Nash equilibria in games is a hard problem both analytically and computationally. We restrict out attention to symmetric Bayes-Nash equilibria in auctions and propose a computational method that takes advantage of the symmetry of equilibria and structure of auction games. The method is iterated best-response where all players move simultaneously. We present experimental results for single unit firstand second-price auctions with discrete values and bids.